SAS 9.

4 Graph Template
®

Language: Reference,
Fifth Edition

SAS® Documentation
May 8, 2024
The correct bibliographic citation for this manual is as follows: SAS Institute Inc. 2016. SAS® 9.4 Graph Template Language: Reference, Fifth
Edition. Cary, NC: SAS Institute Inc.
SAS® 9.4 Graph Template Language: Reference, Fifth Edition
Copyright © 2016, SAS Institute Inc., Cary, NC, USA

All Rights Reserved. Produced in the United States of America.

For a hard copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any
means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS Institute Inc.
For a web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the time you acquire
this publication.
The scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher is illegal
and punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronic piracy of
copyrighted materials. Your support of others' rights is appreciated.
U.S. Government License Rights; Restricted Rights: The Software and its documentation is commercial computer software developed at
private expense and is provided with RESTRICTED RIGHTS to the United States Government. Use, duplication, or disclosure of the Software
by the United States Government is subject to the license terms of this Agreement pursuant to, as applicable, FAR 12.212, DFAR
227.7202-1(a), DFAR 227.7202-3(a), and DFAR 227.7202-4, and, to the extent required under U.S. federal law, the minimum restricted rights
as set out in FAR 52.227-19 (DEC 2007). If FAR 52.227-19 is applicable, this provision serves as notice under clause (c) thereof and no other
notice is required to be affixed to the Software or documentation. The Government’s rights in Software and documentation shall be only
those set forth in this Agreement.
SAS Institute Inc., SAS Campus Drive, Cary, NC 27513-2414
May 2024
SAS® and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and
other countries. ® indicates USA registration.
Other brand and product names are trademarks of their respective companies.
9.4-P9:grstatgraph
Contents

Syntax Conventions Used in this Document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

What’s New in SAS 9.4 Graph Template Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii

PART 1 Fundamentals 1

Chapter 1 / Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Graph Template Language (GTL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Components of a Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Graphical Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Legends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Flexible Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
About the Examples in This Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Examples and Resources on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

PART 2 Graph Block 23

Chapter 2 / BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

PART 3 Layout Statements 45

Chapter 3 / Summary of Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Single-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Multi-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Data-driven Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Legend Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Chapter 4 / Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

PART 4 Plot Statements 199

Chapter 5 / Key Concepts for Using Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Minimum Requirements to Generate a Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
ODS Graphics Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Location and Position of Curve Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
iv Contents

Chapter 6 / Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

PART 5 Plot Axes 1037

Chapter 7 / Axis Features in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039
How Axis Features Are Determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1040
Controlling Axis Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049

Chapter 8 / Axis Options in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055

PART 6 Legend Statements 1279

Chapter 9 / Legend Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281

PART 7 Text Statements 1335

Chapter 10 / Managing Text Items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1337

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1337
Using Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1338
Using Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1343

Chapter 11 / Text Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347

PART 8 Custom Marker Definition Statements 1375

Chapter 12 / Custom Marker Definition Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1377

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1377

PART 9 Draw Statements 1393

Chapter 13 / Key Concepts for Using Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395
Types of Elements That Can Be Drawn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1398
About the Drawing Space and Drawing Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1398
How the Drawn Elements Are Anchored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1401
About Drawing Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1402

Chapter 14 / Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405

PART 10 GTL Annotation Facility 1473

Chapter 15 / About the GTL Annotation Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475

Data-Set-Driven Annotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475
Annotation Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1476
 Contents v

ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481

SGRENDER Statement SGANNO Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481

Chapter 16 / The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483

PART 11 Attribute Maps 1491

Chapter 17 / Key Concepts for Using Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1493

About Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1493
Defining a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1494
Defining a Range Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1498
Referencing an Attribute Map in Your Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . 1499

Chapter 18 / Discrete Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1505

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1505

Chapter 19 / Range Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1521

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1521

PART 12 Run-Time Programming Features 1533

Chapter 20 / Dynamic Variables and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1535

Template Types on PROC TEMPLATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1535
DYNAMIC, MVAR, and NMVAR Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536
Dynamic Variables Compared to Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1537

Chapter 21 / Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1539

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1539
GTL Expressions Compared to WHERE Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1540
An Expression in Statement Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1540

Chapter 22 / Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543
SAS Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1544
Functions Defined Only in GTL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547
GTL Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1550

Chapter 23 / Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1555

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1555
Conditional Logic Determines Statement Rendering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1556
GTL Does Not Provide ELSE IF Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1557

PART 13 Appendixes 1559

Appendix 1 / ODS GRAPHICS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1561

Appendix 2 / Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595

Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595

Appendix 3 / Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1599

General Syntax for Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1599
Attributes Available for the Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1600
vi Contents

Available Line Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607

Color-Naming Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1609

Appendix 4 / Format Support in ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1621

Assigning Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1621
Using Locale-Sensitive Decimal Separators with Numeric Formats . . . . . . . . . . . . . 1622
SAS Formats That Are Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1622
User-Defined Formats That Are Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1624

Appendix 5 / Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1627

Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1627

Appendix 6 / Memory Management for ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1633

SAS Options Affecting Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1633
Managing a Java Out of Memory Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1634
 vii

Syntax Conventions Used in this

Document

Syntax Conventions for the SAS

Language

Overview of Syntax Conventions for the SAS

Language
SAS uses standard conventions in the documentation of syntax for SAS language
elements. These conventions enable you to easily identify the components of SAS
syntax. The conventions can be divided into these parts:
n syntax components

n style conventions

n special characters

n references to SAS libraries and external files

Syntax Components
The components of the syntax for most language elements include a keyword and
arguments. For some language elements, only a keyword is necessary. For other
language elements, the keyword is followed by an equal sign (=). The syntax for
arguments has multiple forms in order to demonstrate the syntax of multiple
arguments, with and without punctuation.
viii Syntax Conventions Used in this Document

keyword
specifies the name of the SAS language element that you use when you write
your program. Keyword is a literal that is usually the first word in the syntax. In a
CALL routine, the first two words are keywords.

In these examples of SAS syntax, the keywords are bold:

CHAR (string, position)

CALL RANBIN (seed, n, p, x);
ALTER (alter-password)
BEST w.
REMOVE <data-set-name>

In this example, the first two words of the CALL routine are the keywords:

CALL RANBIN(seed, n, p, x)

The syntax of some SAS statements consists of a single keyword without

arguments:

DO;
... SAS code ...
END;

Some system options require that one of two keyword values be specified:

DUPLEX | NODUPLEX

Some procedure statements have multiple keywords throughout the statement

syntax:

CREATE <UNIQUE> INDEX index-name ON table-name (column-1 <,

column-2, …>)
argument
specifies a numeric or character constant, variable, or expression. Arguments
follow the keyword or an equal sign after the keyword. The arguments are used
by SAS to process the language element. Arguments can be required or
optional. In the syntax, optional arguments are enclosed in angle brackets ( <
> ).

In this example, string and position follow the keyword CHAR. These arguments
are required arguments for the CHAR function:

CHAR (string, position)

Each argument has a value. In this example of SAS code, the argument string has
a value of 'summer', and the argument position has a value of 4:
x=char('summer', 4);

In this example, string and substring are required arguments, whereas modifiers
and startpos are optional.

FIND(string, substring <, modifiers> <, startpos>

 Syntax Conventions for the SAS Language ix

argument(s)
specifies that one argument is required and that multiple arguments are
allowed. Separate arguments with a space. Punctuation, such as a comma ( , ) is
not required between arguments.

The MISSING statement is an example of this form of multiple arguments:

MISSING character(s);
<LITERAL_ARGUMENT> argument-1 <<LITERAL_ARGUMENT> argument-2 ... >
specifies that one argument is required and that a literal argument can be
associated with the argument. You can specify multiple literals and argument
pairs. No punctuation is required between the literal and argument pairs. The
ellipsis (...) indicates that additional literals and arguments are allowed.

The BY statement is an example of this argument:

BY <DESCENDING> variable-1 <<DESCENDING> variable-2 …>;

argument-1 <options> <argument-2 <options> ...>
specifies that one argument is required and that one or more options can be
associated with the argument. You can specify multiple arguments and
associated options. No punctuation is required between the argument and the
option. The ellipsis (...) indicates that additional arguments with an associated
option are allowed.

The FORMAT procedure PICTURE statement is an example of this form of

multiple arguments:

PICTURE name <(format-options)>

<value-range-set-1 <(picture-1-options)>
<value-range-set-2 <(picture-2-options)> …>>;
argument-1=value-1 <argument-2=value-2 ...>
specifies that the argument must be assigned a value and that you can specify
multiple arguments. The ellipsis (...) indicates that additional arguments are
allowed. No punctuation is required between arguments.

The LABEL statement is an example of this form of multiple arguments:

LABEL variable-1=label-1 <variable-2=label-2 …>;

argument-1 <, argument-2, ...>
specifies that one argument is required and that you can specify multiple
arguments that are separated by a comma or other punctuation. The ellipsis (...)
indicates a continuation of the arguments, separated by a comma. Both forms
are used in the SAS documentation.

Here are examples of this form of multiple arguments:

AUTHPROVIDERDOMAIN (provider-1:domain-1 <, provider-2:domain-2, …>

INTO :macro-variable-specification-1 <, :macro-variable-specification-2, …>

Note: In most cases, example code in SAS documentation is written in lowercase

with a monospace font. You can use uppercase, lowercase, or mixed case in the
code that you write.
x Syntax Conventions Used in this Document

Style Conventions
The style conventions that are used in documenting SAS syntax include uppercase
bold, uppercase, and italic:
UPPERCASE BOLD
identifies SAS keywords such as the names of functions or statements. In this
example, the keyword ERROR is written in uppercase bold:

ERROR <message>;
UPPERCASE
identifies arguments that are literals.

In this example of the CMPMODEL= system option, the literals include BOTH,
CATALOG, and XML:

CMPMODEL=BOTH | CATALOG | XML |

italic
identifies arguments or values that you supply. Items in italic represent user-
supplied values that are either one of the following:
n nonliteral arguments. In this example of the LINK statement, the argument
label is a user-supplied value and therefore appears in italic:

LINK label;
n nonliteral values that are assigned to an argument.

In this example of the FORMAT statement, the argument DEFAULT is

assigned the variable default-format:

FORMAT variable(s) <format > <DEFAULT = default-format>;

Special Characters
The syntax of SAS language elements can contain the following special characters:
=
an equal sign identifies a value for a literal in some language elements such as
system options.

In this example of the MAPS system option, the equal sign sets the value of
MAPS:

MAPS=location-of-maps
<>
angle brackets identify optional arguments. A required argument is not enclosed
in angle brackets.
 Syntax Conventions for the SAS Language xi

In this example of the CAT function, at least one item is required:

CAT (item-1 <, item-2, …>)

|
a vertical bar indicates that you can choose one value from a group of values.
Values that are separated by the vertical bar are mutually exclusive.

In this example of the CMPMODEL= system option, you can choose only one of
the arguments:

CMPMODEL=BOTH | CATALOG | XML

...
an ellipsis indicates that the argument can be repeated. If an argument and the
ellipsis are enclosed in angle brackets, then the argument is optional. The
repeated argument must contain punctuation if it appears before or after the
argument.

In this example of the CAT function, multiple item arguments are allowed, and
they must be separated by a comma:

CAT (item-1 <, item-2, …>)

'value' or "value"
indicates that an argument that is enclosed in single or double quotation marks
must have a value that is also enclosed in single or double quotation marks.

In this example of the FOOTNOTE statement, the argument text is enclosed in

quotation marks:

FOOTNOTE <n> <ods-format-options 'text' | "text">;

;
a semicolon indicates the end of a statement or CALL routine.

In this example, each statement ends with a semicolon:

data namegame;
length color name $8;
color = 'black';
name = 'jack';
game = trim(color) || name;
run;

References to SAS Libraries and External Files

Many SAS statements and other language elements refer to SAS libraries and
external files. You can choose whether to make the reference through a logical
name (a libref or fileref) or use the physical filename enclosed in quotation marks.

If you use a logical name, you typically have a choice of using a SAS statement
(LIBNAME or FILENAME) or the operating environment's control language to make
the reference. Several methods of referring to SAS libraries and external files are
available, and some of these methods depend on your operating environment.
xii Syntax Conventions Used in this Document

In the examples that use external files, SAS documentation uses the italicized
phrase file-specification. In the examples that use SAS libraries, SAS
documentation uses the italicized phrase SAS-library enclosed in quotation marks:
infile file-specification obs = 100;
libname libref 'SAS-library';

Value Type Notation Used for GTL

Statement Options
The value type notation that is used in the syntax descriptions and in some
examples in this document are as follows:
boolean
specifies a literal value that resolves to true or false. The following table lists
literal values that resolve to true or false.

Boolean Values

Values That Resolve To True Values That Resolve To False

ON OFF

_ON_ _OFF_

TRUE FALSE

YES NO

_YES_ _NO_

1 0

color
specifies a string that identifies a color. A color can be one of the following:
n any of the color names that are supported by SAS. See “Color-Naming
Schemes” on page 1609.
n one of the colors that exists in the SAS session when the style template is
used, such as DMSBLACK or DMSCYAN. (Use these color specifications
only if you are running SAS in the windowing environment.)
n an English description of an Hue/Light/Saturation (HLS) value. Such
descriptions use a combination of words to describe the lightness, the
 Value Type Notation Used for GTL Statement Options xiii

saturation, and the hue (in that order). You can use the Color Naming System
to form a color by doing one of the following:
o combining a chromatic hue with a lightness, a saturation, or both
o combining the achromatic hue gray with a lightness
o combining the achromatic hue black or white without qualifiers.
o combining words to form a wide variety of colors, such as light vivid
green, dark vivid orange, or light yellow.
n specify hues that are intermediate between two neighboring colors. To do so,
combine one of the following adjectives with one of its neighboring colors:
brownish, greenish, purplish, or yellowish (for example, bluish purple or
reddish orange).
column
specifies a column variable that contains either double-precision values or
string values, or a dynamic variable that refers to such a column.

See also: integer-column, numeric-column, and string-column.

dimension
specifies a nonnegative number. The number can be followed by one of the
following optional units of measure:

Units for Dimension

Unit Description

CM centimeters

IN inches

MM millimeters

PCT or % percentage

PT point size (72 points = 1 inch)

PX pixels

expression
specifies a selective, relational, or logical program structure that calculates
values when those values are not stored in the data. The expression must be
specified as an EVAL() argument. The following shows the structure of an
EVAL() argument:

x = EVAL(expression)

The expression returns a number and can be formed with consonants, data
columns, dynamic variables, functions, or other expressions. The following
example uses the data column Time and the SGE functions MEAN and ACF:
xiv Syntax Conventions Used in this Document

EVAL(MEAN(Time) + ACF(Time, NLags=10))

For more information about expressions, see Chapter 21, “Expressions,” on page
1539.
fill-pattern
specifies a fill pattern as a two-character code that consists of a line-direction
prefix (R for right, L for left, or X for cross hatch) and a line identification
number, 1–5. For more information about fill patterns, see “Fill Pattern Options”
on page 1601.
format
specifies a SAS format or a user-defined format.
integer, integer-column
specifies a member of the set of positive whole numbers, negative whole
numbers, and zero.

An integer column specifies a column that contains integer values, or a dynamic

variable that refers to such a column.
line-pattern-name, line-pattern-number
specifies a string value of a line pattern, a numeric value of a line pattern, a
dynamic variable that contains such a string or number, or a style reference to a
line pattern. Line patterns are chosen for discriminability. Because of different
densities, equal weighting is impossible for lines of the same thickness. Instead,
line patterns are ordered to provide a continuum of weights, which is useful
when displaying confidence bands.

For details about line attributes, see “Line Options” on page 1602.
marker-name
specifies a string value of a maker symbol, a dynamic variable that contains a
marker symbol, or a style reference to a marker symbol.

For details about marker attributes, see “Marker Options” on page 1604.
number, numeric-column
specifies a value, a dynamic variable that contains a double-precision value, an
expression that resolves to a double-precision value, or a style reference to a
double-precision value.

A numeric-column specifies a column that contains double-precision values, or a

dynamic variable that refers to such a column.
string, string-column
specifies a quoted character string.

A string-column specifies a column that contains string values, or a dynamic

variable that refers to such a column.

Note: For quoted character string options in GTL, a space enclosed in quotation
marks (" " or ' ') and empty quotation marks ("" or '') are not equivalent. A space
enclosed in quotation marks specifies a blank space or a missing string value.
Empty quotation marks have the same effect as not setting the option. To
 Value Type Notation Used for GTL Statement Options xv

specify a blank space or missing value in a quoted string option, use a space
enclosed in quotation marks (" " or ' ').

style-reference
specifies a reference to an attribute that is defined in a style element.

In the ODS Graphics templates that SAS provides, options for plot features are
specified with a style reference in the form style-element-name:attribute-name,
rather than a specific value. For example, the symbol, color, and size of markers
for a basic scatter plot is specified in a SCATTERPLOT statement as follows:

scatterplot x=X y=Y /

markersymbol=GraphDataDefault:markersymbol
markercolor=GraphDataDefault:contrastcolor
markersize=GraphDataDefault:markersize

The above style references guarantee a common appearance for markers used in
all basic scatter plots. For non-grouped data, the marker appearance is
controlled by the GraphDataDefault style element in the style template that
you specify.

In order to create your own style template, or to modify a style template to use
with ODS Graphics, you need to understand the relationship between style
elements and graph features. For more information, see the usage guide.
xvi Syntax Conventions Used in this Document
 xvii

What’s New in SAS 9.4 Graph

Template Language

Overview
New and enhanced statements for the Graph Template Language (GTL) extend the
versatility of the language and introduce new plot types. The following changes are
included in this release:
n new plot statements

n new features for general use

n enhancements to SAS 9.3 statements

Possible Changes Required for Your

Existing SAS GTL Programs
There are changes in GTL for SAS 9.4 and subsequent maintenance releases that
might require you to modify your existing SAS GTL programs. Review this section
carefully to determine whether you need to modify your existing SAS GTL
programs for SAS 9.4.

Programs That Use the BARCHART or

BARCHARTPARM Statement
In SAS 9.4, the BARCHART statement STAT=PCT option now displays percentages
in the range 0–100 in order to be consistent with other GTL statements. In prior
xviii What’s New in SAS 9.4 Graph Template Language

releases, STAT=PCT displays proportional values in the range 0–1. To restore the
proportional values in SAS 9.4 and later releases, change STAT=PCT to
STAT=PROPORTION.

SAS 9.4M3 adds option COLORSTAT= to the BARCHART statement. It is enabled

by the COLORRESPONSE= option. The COLORSTAT= option specifies the statistic
to be calculated for the data range of the bar-color gradient. The default is SUM.
For existing SAS programs that use the BARCHART statement, if STAT= is used
with COLORRESPONSE= in the BARCHART statement and STAT= specifies a
statistic other than SUM, then the bar-chart colors and color statistic might change
from those of previous releases. In that case, to restore the original colors and color
statistic, set COLORSTAT= in the BARCHART statement to the same statistic that
is specified in STAT=.

SAS 9.4M5 changes the behavior of the BARCHART and BARCHARTPARM

statement option FILLTYPE=GRADIENT. In earlier releases, GRADIENT fills each
bar with a color and a transparency gradient that starts at the bar top with the
initial transparency that is assigned to that bar, and ends with full transparency at
the bar baseline. In SAS 9.4M5, GRADIENT is replaced by ALPHAGRADIENT, which
has the same behavior as GRADIENT in earlier releases. GRADIENT is changed to
an alias of COLORGRADIENT, which fills each bar with an opaque color gradient
that starts with the fill color that is assigned to that bar and ends with the color
that is specified in a new option, FILLENDCOLOR=. For existing SAS programs that
use the BARCHART or BARCHARTPARM statement FILLTYPE=GRADIENT option,
the bar gradients will change. To restore the original bar gradients in that case,
specify FILLTYPE=ALPHAGRADIENT.

See BARCHART in “Plot Statement Enhancements” on page xxvii.

Programs That Use the DENSITYPLOT Statement

In SAS 9.4, the WEIGHT= option of the KERNEL( ) distribution option in the
DENSITYPLOT statement is changed to WEIGHTFUNCTION=. This change enables
the addition of the WEIGHT= option in the DENSITYPLOT statement. In SAS 9.4,
the WEIGHT= option is not valid in the KERNEL() distribution option and results in
a syntax error. If your existing SAS programs specify the WEIGHT= option in the
KERNEL( ) distribution option, then you must change it to WEIGHTFUNCTION=.
See DENSITYPLOT in “Plot Statement Enhancements” on page xxvii.

Programs That Use the HEATMAPPARM Statement

SAS 9.4M3 adds options DISCRETEX= and DISCRETEY= to the HEATMAPPARM
statement. These options specify whether the X or Y axis is discrete. The default is
FALSE. For existing SAS programs that use the HEATMAPPARM statement, if the X
axis is discrete, the DISCRETEX=TRUE option must be specified in the
HEATMAPPARM statement. Likewise, if the Y axis is discrete, the
 Possible Changes Required for Your Existing SAS GTL Programs xix

DISCRETEY=TRUE option must be specified in the HEATMAPPARM statement.

Otherwise, the heat map might not be drawn.

Programs That Use the HISTOGRAM or

HISTOGRAMPARM Statement
SAS 9.4M5 changes the behavior of the FILLTYPE=GRADIENT option. In earlier
releases, GRADIENT fills each bar with a color and a transparency gradient that
starts at the bar top with the initial transparency that is assigned to that bar, and
ends with full transparency at the bar baseline. SAS 9.4M5 replaces option
GRADIENT with option ALPHAGRADIENT, which has the same behavior as
GRADIENT in earlier releases. GRADIENT is changed to an alias of
COLORGRADIENT, which fills each bar with an opaque color gradient that starts
with the fill color that is assigned to that bar and ends with the color that is
specified in a new option, FILLENDCOLOR=. For existing SAS programs that use
the HISTOGRAM or HISTOGRAMPARM statement FILLTYPE=GRADIENT option,
the bar gradients will change. To restore the original bar gradients in that case,
specify FILLTYPE=ALPHAGRADIENT.

Programs That Use the SERIESPLOT Statement

SAS 9.4M4 changes the behavior of the BREAK=TRUE option to break the series
line at missing values of only the Y variable. In earlier SAS releases, the
BREAK=TRUE option breaks the series line at missing values of the X or Y variable.
If you want your existing SAS programs that specify BREAK=TRUE in a
SERIESPLOT statement to break the series line at missing X or Y variable values in
SAS 9.4M4 and later releases, do one of the following:
n In your data, set the Y variable value to missing when the X variable value is
missing.
n In the SERIESPLOT statement in your template, use the IFN or IFC function to
set the Y variable value to missing when the X variable value is missing, as
shown in the following examples.
seriesplot x=x-var-name y=eval(ifn(x-var-name ne ., y-var-name, .));
seriesplot x=x-var-name y=eval(ifc(x-var-name, y-var-name, ''));
xx What’s New in SAS 9.4 Graph Template Language

Programs That Use the WATERFALLCHART

Statement
SAS 9.4M3 adds option COLORSTAT= to the WATERFALLCHART statement. It is
enabled by the COLORRESPONSE= option. The COLORSTAT= option specifies the
statistic to be calculated for the data range of the color gradient. The default is
SUM. For existing SAS programs that use the WATERFALLCHART statement, if
STAT= is used with COLORRESPONSE= in the WATERFALLCHART statement and
STAT= specifies a statistic other than SUM, then the chart colors and color statistic
might change from those of previous releases. In that case, to restore the original
colors and color statistic, set COLORSTAT= in the WATERFALLCHART statement
to the same statistic that is specified in STAT=. See WATERFALLCHART in “Plot
Statement Enhancements” on page xxvii.

Programs That Use the DISPLAY=ALL Option in

Plot Statements
SAS 9.4M5 adds support for fill patterns to band plots, box plots, bubble plots,
high-low plots, ellipse plots, and polygon plots. SAS 9.4M5 also adds fill patterns to
the ODS styles. If your SAS program uses the DISPLAY=ALL option in any of these
plot statements, your graph might now display fill patterns in addition to fill color,
outline, and so on. If you do not want to use fill patterns in your graph, use the
DISPLAY= option in your plot statement to display only the features that you want.

New Plot Statements

The following plot statements are new:
n ANNOTATE draws annotations from annotation instructions that are stored in a
SAS data set. It can draw all of the annotations in the data set or a subset only.
n AXISTABLE draws textual values (character or numeric) on the graph at
locations that are aligned with the X or Y axis.
n LINECHART creates a summarized line chart that is computed from input data.

SAS 9.4M1 adds the following statement:

n POLYGONPLOT draws one or more polygons from data that is stored in a SAS
data set.
 New Function xxi

SAS 9.4M2 adds the following statement:

n TEXTPLOT displays text values at specific X and Y locations in the graph. For
information about subsequent enhancements to this statement, see “Plot
Statement Enhancements” on page xxvii.

SAS 9.4M3 adds the following statement:

n HEATMAP creates a plot of color-coded rectangles for the response variable of
a pair of X and Y variables after it bins the data in two dimensions.

New Legend Statement

SAS 9.4M3 adds the LEGENDTEXTITEMS statement. This statement creates the
definition for data-driven text items that can be included in a discrete legend. The
items and optional labels are stored in the plot data set.

New Marker Definition Statements

SAS 9.4M1 adds the following statements:
n SYMBOLCHAR defines a marker symbol from a Unicode character value.

n SYMBOLIMAGE defines a marker symbol from a GIF, JPG, or PNG image that is
stored in a file.

You can use these statements to define custom marker symbols for your graphs.
For information about subsequent enhancements to these statements, see “Marker
Definition Statement Enhancements” on page lvi.

New Function
SAS 9.4M1 adds the TYPEOF function. This function returns the type (numeric or
character) of a specified column at run time. You can use the TYPEOF function to
take specific actions in your template at run time based on the input data type.
xxii What’s New in SAS 9.4 Graph Template Language

New Features for General Use

Color-Priority Graph Data Attribute Rotation

GTL now supports a color-priority rotation pattern for cycling graph data
attributes. In the default rotation pattern, graph data attributes rotate as they are
defined in the GraphData1–GraphDataN style elements. The new color-priority
rotation pattern first cycles through all of the colors while holding the marker
symbol, line pattern, or fill pattern constant. When all of the colors have been used,
it increments to the next marker symbol, line pattern, or fill pattern, and then
repeats the colors.

Color-priority rotation is enabled in one of the following ways:

n the currently active style sets the Graph:attrPriority attribute to COLOR

n the ATTRPRIORITY=COLOR option is specified in the ODS GRAPHICS

statement, which overrides the Graph:attrPriority style attribute
n the ATTRPRIORITY=COLOR option is specified in the BEGINGRAPH statement
of the template, which overrides the ODS GRAPHICS ATTRPRIORITY= option
for all plots in that template

When none of these conditions are true, the default rotation pattern is used. For
more information, see “BEGINGRAPH Statement Enhancements” on page xxiv.

Template-Level Graph Data Attribute Overrides

GTL provides new options that enable you to override GraphData1–GraphDataN
style attributes for all of the plots within a template. They provide an easy way to
modify group attributes without having to create a custom ODS style. These
options are used in the BEGINGRAPH statement of a template. By using these
options, you can override attributes in the GraphData1–GraphDataN style elements
by specifying one or more of the following lists for all of the plots within a
template:
n a list of colors that replace the graph data colors (fill) in the current style

n a list of contrast colors that replace the graph data contrast colors in the current
style
n a list of marker symbols that replace the graph data marker symbols in the
current style
 New Features for General Use xxiii

n a list of line patterns that replace the graph data line patterns in the current
style

In addition, you can specify a priority for the group value attribute cycling. For more
information, see “BEGINGRAPH Statement Enhancements” on page xxiv.

Outer Padding
GTL provides a new OUTERPAD= option that enables you to control the padding
around the outside of layouts, legends, titles, footnotes, and text entries. You can
use this option to modify the outer padding when the default padding is not
sufficient. You can specify a single padding value for all four sides, or you can
specify different values for each side.

Unicode Values in User-Defined Formats

SAS 9.4M3 adds support for Unicode values in user-defined formats to ODS
Graphics. The Unicode value must be escaped with the (*ESC*) escape sequence, as
shown in the following example.
"(*ESC*){unicode beta}"

ODS Graphics does not support the use of a user-defined ODS escape character to
escape Unicode values in user-defined formats.

General Enhancements Supported by Many of the

Plots
The following new features that are supported by many of the plot statements are
worth highlighting. The individual plot statements that support these features are
identified in “Plot Statement Enhancements” on page xxvii.
n The algorithm that is used to place data labels has been improved to more
effectively position the data labels in the vicinity of their data markers while
avoiding label collisions.
n In certain cases, you can rotate or split data labels, curve labels, and discrete-
axis tick values in order to fit them in the available space.
n Subpixel rendering is now supported, which produces smoother curves and
more precise bar spacing in plots.
n You can now suppress tips and URLs from individual plots.
xxiv What’s New in SAS 9.4 Graph Template Language

Enhancements to SAS 9.4 Statements

BEGINGRAPH Statement Enhancements

n ATTRPRIORITY= specifies a priority for cycling of the attributes for group
attributes.
n DATACOLORS= specifies the list of fill colors that replaces the graph data
colors from the GraphData1–GraphDataN style elements.
n DATACONTRASTCOLORS= specifies the list of contrast colors that replaces
the graph data contrast colors from the GraphData1–GraphDataN style
elements.
n DATALINEPATTERNS= specifies the list of line patterns that replaces the graph
data line patterns from the GraphData1–GraphDataN style elements.
n DATASKIN= enhances the visual appearance of all plots in the template that
support data skins.
n DATASYMBOLS= specifies the list of marker symbols that replaces the graph
data marker symbols from the marker symbols that are defined in the
GraphData1–GraphDataN style elements.
n SUBPIXEL= specifies whether subpixel rendering is used for drawing smooth
curved lines or for spacing bars more precisely.
n SAS 9.4M3 has the following enhancements:

o AXISBREAKSYMBOL= specifies a symbol to use on the axis lines to indicate

a break in the axis.
o AXISBREAKTYPE= specifies whether the axis break is indicated in the full
display or only on the axis line.
o AXISLINEEXTENT= specifies the extent of the axis line for all axes.
o DISCRETEAXISOFFSETPAD= specifies whether additional padding is added
to the minimum and maximum axis offsets for discrete axes.
o OPAQUE= specifies whether the graph background is opaque or transparent.
n SAS 9.4M5 adds option DATAFILLPATTERNS=, which specifies a list of fill
patterns to replace the graph data fill patterns that are provided by the
GraphData1–GraphDataN style elements, if any.
 Enhancements to SAS 9.4 Statements xxv

Layout Statement Enhancements

LAYOUT DATALATTICE and LAYOUT DATAPANEL:
n For cell insets, match-merging is now supported for merging the inset and
analysis data. The DATASCHEME= suboption of the INSETOPTS= option
specifies whether one-to-one merging or match merging was used to merge the
data.
n HEADERLABELDISPLAY= now supports NONE, which suppresses the row and
column headings, or the cell headings.
n SORTORDER= specifies the order of the cells along the columns and rows.

n LAYOUT DATALATTICE only:

o COLUMNDATARANGE=, ROWDATARANGE=, COLUMN2DATARANGE=,

and ROW2DATARANGE= now support AUTO, which selects the range
automatically, based on the column weight or row weight, and the axis type.
o COLUMNWEIGHT= specifies how to assign weights to the columns widths.
o ROWWEIGHT= specifies how to assign weights to the row heights.
n SAS 9.4M1 has the following enhancements:

o HEADERPACK= specifies whether the values listed in the cell headers are
displayed as a delimited list in a single header cell in order to save space.
o HEADERSEPARATOR= specifies a separator to place between each value in
the cell header.
o HEADERSPLITCOUNT= specifies the number of class variables in the cell
header after which the cell header wraps to a separate line.
n SAS 9.4M2 has the following enhancements:

o For the INSETOPTS= option:

n CONTENTDISPLAY= specifies whether the variable information
displayed in the inset includes the column label and value, or the column
value only.
n SEPARATOR= specifies a new separator for the column label and value.
n SAS 9.4M3 adds option HEADERBORDER=, specifies whether a border is drawn
around the header cells.

LAYOUT GLOBALLEGEND:
n SAS 9.4M1 changes the default padding between the global legend and the plot
area (including the axes) to 10 pixels. Use the OUTERPAD= option to adjust the
padding, if necessary.

LAYOUT OVERLAY:
n INNERMARGIN statement:

o ALIGN= now supports LEFT and RIGHT in addition to TOP and BOTTOM.
xxvi What’s New in SAS 9.4 Graph Template Language

o BACKGROUNDCOLOR= specifies the color of the inner margin background.

o OPAQUE= specifies whether the inner margin's background is opaque.
o SAS 9.4M1 has the following enhancements:
n SEPARATOR= specifies whether a separating line is drawn between the
inner margin and the rest of the layout content.
n SEPARATORATTRS= specifies the attributes of the inner margin
separating line.
o SAS 9.4M2 has the following enhancements:
n GUTTER= specifies the gap between stacked items in the inner margin.
n PAD= specifies the amount of extra space that is added inside the inner
margin border.

LAYOUT OVERLAYEQUATED:
n The HEATMAPPARM statement is now supported.

n EQUATETYPE= now supports SQUAREDATA, which specifies that both the X

and Y axes have the same range, but can have different tick values.

LAYOUT PROTOTYPE:
n The HEATMAPPARM statement is now supported.

n SAS 9.4M1 enables the INNERMARGIN statement to be used in a LAYOUT

PROTOTYPE block.

Legend Statement Enhancements

AXISLEGEND:
n SAS 9.4M2 changes the BORDER= option default to style reference
GraphLegendBackground:FrameBorder.

DISCRETELEGEND:
n SAS 9.4M1 has the following enhancements:

o The default padding between the legend and the plot area (including the
axes) is 10 pixels, depending on the context. You can use the OUTERPAD=
option to adjust the padding if necessary.
o ITEMSIZE= specifies the size of specific types of items in a discrete legend.
n SAS 9.4M2 changes the BORDER= option default to style reference
GraphLegendBackground:FrameBorder.
n SAS 9.4M3 has the following enhancements:

o FILLITEMOUTLINE= specifies whether the fill swatches are outlined only

when enabled by the contributing statements or are always outlined.
o The ITEMSIZE= option now supports the following suboptions:
 Enhancements to SAS 9.4 Statements xxvii

n FILLASPECTRATIO= specifies the aspect ratio for the fill swatches.

n FILLHEIGHT= specifies the height of the fill swatches.
n HEIGHTSCALE= specifies a scaling factor that is to be applied to the fill
swatch height.
o SORTBY= specifies whether text legend items are sorted by label or by text.
o SAS 9.4M5 adds REVERSEAUTO to option SORTORDER=. REVERSEAUTO
presents groups of legend entries in the reverse order in which they are listed
in the legend statement.

CONTINUOUSLEGEND:
n EXTRACTSCALE= specifies whether to extract a scale factor from the tick
values and use it to reduce the tick value width.
n EXTRACTSCALETYPE= specifies whether to extract a named scale or a
scientific-notation scale.
n SAS 9.4M2 changes the BORDER= option default to style reference
GraphLegendBackground:FrameBorder.
n SAS 9.4M3 adds option INTEGER=, which specifies whether only integer tick
values are used in the continuous legend.

LEGENDITEM:
n FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
n SAS 9.4M3 adds option FILLDISPLAY=, which specifies whether the fill swatch
for this legend item displays fill only or displays fill and outline.

MERGEDLEGEND:
n SAS 9.4M1 has the following enhancements:

o ADDITIONALNAMES= specifies the name of one or more sources for legend

items that are to be added to the merged legend after merging takes place.
o ITEMSIZE= specifies the size of specific types of items in a merged legend.
n SAS 9.4M2 changes the BORDER= option default to style reference
GraphLegendBackground:FrameBorder.
n SAS 9.4M5 adds REVERSEAUTO to option SORTORDER=. REVERSEAUTO
presents groups of legend entries in the reverse order in which they are listed in
the legend statement.

Plot Statement Enhancements

AXISTABLE:
n SAS 9.4M2 has the following enhancements:

o CLASSDISPLAY= specifies how the class values are displayed.

xxviii What’s New in SAS 9.4 Graph Template Language

o CLASSORDER= specifies the order in which the class values are displayed.
o CLUSTERWIDTH= specifies the width of the group clusters as a fraction of
the midpoint spacing.
o DISPLAY= now supports VALUES, which displays the table values.
o DROPONMISSING= specifies whether the entire axis table is dropped when
all of the VALUE= column values are missing.
o GUTTER= specifies the gap between rows when a class variable is used.
o INCLUDEMISSINGCLASS= specifies whether missing class values are
represented in the table.
o LABEL= specifies the text for the table label.
o LABELHALIGN= specifies the horizontal alignment of the column labels
relative to the column width in a Y-axis table.
o LABELJUSTIFY= specifies the justification of the column label, when
displayed.
o PAD= specifies the amount of extra space that is added inside the table
border.
o SHOWMISSING= specifies whether missing values are represented in the
table.
o TITLE= specifies the text for the table title.
o TITLEATTRS=specifies the color and font attributes of the table title.
o TITLEHALIGN= specifies the horizontal alignment of the axis table header
label relative to the table width for a Y-axis table.
o TITLEJUSTIFY= specifies the justification of the title string.
o VALUEFORMAT= specifies a SAS format or a user-defined format for the
table values.
o VALUEHALIGN= specifies the horizontal alignment of the column values
relative to the column width in a Y-axis table.
o VALUEJUSTIFY= specifies the justification of the values in the axis table.
o The following options are replaced and considered deprecated:
n HEADERLABEL= is replaced with TITLE=.
n HEADERLABELATTRS= is replaced with TITLEATTRS=.
The deprecated options are still honored, but the new options are the
preferred options.
n SAS 9.4M3 adds option TITLEHALIGN=, which specifies the horizontal
alignment of the axis table header label relative to the table width for Y-axis
tables and X-axis tables.

BANDPLOT:
n TIP= now supports NONE, which suppressed data tips from the plot.

n SAS 9.4M2 adds option ANTIALIAS=, which specifies whether anti-aliasing is

turned off for a band plot.
 Enhancements to SAS 9.4 Statements xxix

n SAS 9.4M5 has the following enhancements:

o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.

o DISPLAY= now supports FILLPATTERN, which displays the band fill pattern.

BARCHART:
n X= is changed to CATEGORY= in order to be consistent with the other GTL plot
statements. The X= option is still valid for backward compatibility. However,
you should change the X= option to CATEGORY= in your BARCHART
statements.

Note: If you specify X as a data tip role when you change X= to CATEGORY=,
then you must also change X to CATEGORY in your data tip options.

n Y= is changed to RESPONSE= in order to be consistent with the other GTL plot

statements. The Y= option is still valid for backward compatibility. However,
you should change the Y= option to RESPONSE= in your BARCHART
statements.

Note: If you specify Y as a data tip role when you change Y= to RESPONSE=,
then you must also change Y to RESPONSE in your data tip options.

n BARLABELFITPOLICY= specifies a policy for avoiding collisions among the bar

labels when bar labels are displayed.
n BASELINEATTRS= specifies the appearance of the baseline.

n The baseline is now drawn by default. To suppress the baseline, use the
BASELINEATTRS= option to set the line thickness to 0.
n DATALABELFITPOLICY= specifies a policy for avoiding collisions among the
bar labels when bar labels are displayed.
n STAT= now supports PROPORTION, which displays proportions in the range 0–
1.
n STAT=PCT now displays percentages in the range 0–100 in order to be
consistent with other GTL statements.

Note: In prior SAS releases, the BARCHART statement STAT=PCT option

displays proportional values in the range 0–1. To restore the original STAT=PCT
results in SAS 9.4 and later releases, specify STAT=PROPORTION in your
BARCHART statements instead.

n TIP= now supports NONE, which suppressed data tips and URLs from the plot.

n SAS 9.4M2 has the following enhancements:

o GROUPORDER= supports REVERSEDATA, which orders the groups within a

category in the reverse group-column data order.
o FILLTYPE= specifies whether the fill color is solid or is a gradient that
transitions from fully opaque to fully transparent.
xxx What’s New in SAS 9.4 Graph Template Language

o SEGMENTLABEL= specifies whether a label is displayed inside each bar

segment.
o SEGMENTLABELATTRS= specifies the text properties of the bar segment
label text.
o SEGMENTLABELFITPOLICY= specifies a policy for fitting the bar segment
labels within the bar segments.
o SEGMENTLABELFORMAT= specifies the text format used to display the bar
segment labels.
n SAS 9.4M3 has the following enhancements:

o The BARCHART statement now supports a linear category axis or a time

category axis.
o COLORBYFREQ= specifies whether the bar colors are based on the
frequency of the category variable.
o COLORMODEL= specifies a color ramp that is to be used with the
COLORRESPONSE= option or the COLORBYFREQ= option.
o COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar colors.
o COLORSTAT= specifies the statistic to use for computing the response
colors.
o DISPLAYZEROLENGTHBAR= specifies whether zero-length bars are drawn.
o GROUP100= displays the computed response values (FREQ, SUM, or
MEAN), normalized to 100%.
o INTERVALBARWIDTH= specifies the width of the bars in an interval bar
chart as a ratio of the interval width.
n SAS 9.4M5 has the following enhancements:

o BARLABELFITPOLICY=INSIDEPREFERRED attempts to place the bar labels

in a horizontal bar chart inside the bars.
o DISPLAYBASELINE= specifies whether the baseline is displayed.
o FILLENDCOLOR=color specifies the end color of the bar fill color gradient.
o FILLTYPE=ALPHAGRADIENT replaces FILLTYPE=GRADIENT from the
previous releases. This option fills each bar with a color and a transparency
gradient that starts at the bar top with the initial transparency that is
assigned to that bar, and ends with full transparency at the bar baseline.
o FILLTYPE=COLORGRADIENT fills each bar with a color gradient that starts
with the fill color that is assigned to that bar and ends with the color that is
specified in the FILLENDCOLOR= option.
o FILLTYPE=GRADIENT is changed to an alias of
FILLTYPE=COLORGRADIENT.
o CONNECTBREAK= specifies whether the connect line is broken for values
that have no observations.

BARCHARTPARM:
 Enhancements to SAS 9.4 Statements xxxi

n X= is changed to CATEGORY= in order to be consistent with the other GTL plot

statements. The X= option is still valid for backward compatibility. However,
you should change the X= option to CATEGORY= in your BARCHARTPARM
statements.
n Y= is changed to RESPONSE= in order to be consistent with the other GTL plot
statements. The Y= option is still valid for backward compatibility. However,
you should change the Y= option to RESPONSE= in your BARCHARTPARM
statements.
n BASELINEATTRS= specifies the appearance of the baseline.

n The baseline is now drawn by default. To suppress the baseline, use the
BASELINEATTRS= option to set the line thickness to 0.
n DATALABELFITPOLICY= specifies a policy for avoiding collisions among the
bar labels when bar labels are displayed.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap.

n TIP= now supports NONE, which suppressed data tips and URLs from the plot.

n SAS 9.4M2 has the following enhancements:

o ERRORBARCAPSHAPE= defaults to style reference GraphError:CapStyle.

o DATALABELTYPE= specifies whether the data labels display the RESPONSE
values or the values of the column specified by the DATALABEL= option.
o FILLTYPE= specifies whether the fill color is solid or is a gradient that
transitions from fully opaque to fully transparent.
o GROUPORDER= supports REVERSEDATA, which orders the groups within a
category in the reverse group-column data order.
o SEGMENTLABELATTRS= specifies the text properties of the bar segment
label text.
o SEGMENTLABELFITPOLICY= specifies a policy for fitting the bar segment
labels within the bar segments.
o SEGMENTLABELFORMAT= specifies the text format used to display the bar
segment labels.
o SEGMENTLABELTYPE= specifies whether a label is displayed inside each
bar segment.
n SAS 9.4M3 has the following enhancements:

o The BARCHARTPARM statement now supports a linear category axis or a

time category axis.
o COLORMODEL= specifies a color ramp that is to be used with the
COLORRESPONSE= option.
o COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar colors.
xxxii What’s New in SAS 9.4 Graph Template Language

o DISPLAYZEROLENGTHBAR= specifies whether zero-length bars are drawn.

o GROUP100= displays the computed values normalized to 100%.
o INTERVALBARWIDTH= specifies the width of the bars as a ratio of the
interval width.
n SAS 9.4M5 has the following enhancements:

o DATALABELFITPOLICY=INSIDEPREFERRED attempts to place the data

labels in a horizontal bar chart inside the bars.
o DISPLAYBASELINE= specifies whether the baseline is displayed.
o FILLENDCOLOR=color specifies the end color of the bar fill color gradient.
o FILLTYPE=ALPHAGRADIENT replaces FILLTYPE=GRADIENT from the
previous releases. This option fills each bar with a color and a transparency
gradient that starts at the bar top with the initial transparency that is
assigned to that bar, and ends with full transparency at the bar baseline.
o FILLTYPE=COLORGRADIENT fills each bar with a color gradient that starts
with the fill color that is assigned to that bar and ends with the color that is
specified in the FILLENDCOLOR= option.
o FILLTYPE=GRADIENT is changed to an alias of
FILLTYPE=COLORGRADIENT.
o CONNECTBREAK= specifies whether the connect line is broken for values
that have no observations.
o ERRORBARCAPSCALE= specifies a positive number to use as a multiplier
for determining the default width of the error-bar caps.

BLOCKPLOT:
n BLOCKLABEL= specifies alternative text to display for the internal block text
values.
n VALUEFITPOLICY= now supports the SPLIT, SPLITALWAYS, and NONE
policies.
n VALUESPLITCHAR= specifies one or more characters on which the values can
be split if needed.
n VALUESPLITCHARDROP= specifies whether the split characters are included in
the displayed values.

BOXPLOT:
n CAPSHAPE= now supports NONE, which specifies that no shape is displayed at
the ends of the box whiskers.
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
 Enhancements to SAS 9.4 Statements xxxiii

n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are

inside the data label blocks.
n DATASKIN= enhances the visual appearance of the filled boxes.

n DISPLAYSTATS= specifies the statistics that are displayed for each box.

n TIP= now supports NONE, which suppressed data tips from the plot.

n WEIGHT= specifies a column that contains a statistics calculation a priori

weight for each observation of the input data object.
n SAS 9.4M1 has the following enhancements:

o WHISKERPERCENTILE= specifies the whisker length in percentile units.

o DISPLAYSTATS= now displays the DATAMAX, DATAMIN, and SUMWGT
statistics.
o LEGENDLABEL= now defaults to the Y= column label or name.
n SAS 9.4M2 adds option GROUPORDER=REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
n SAS 9.4M3 modifies the BOXPLOT statement to contribute its mean markers to
a discrete legend when the BOXPLOT DISPLAY= option includes MEAN, and
TYPE=MARKER is specified in the DISCRETELEGEND statement.
n SAS 9.4M5 has the following enhancements:

o DISPLAYSTATS= can be used with both vertical and horizontal box plots.
o CONNECTBREAK= specifies whether the connect line is broken for values
that have no observations.
o CAPSCALE= specifies a positive number to use as a multiplier for
determining the default width of the whisker caps.
o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.
o DISPLAY= now supports FILLPATTERN, which displays the box fill pattern.

BOXPLOTPARM:
n CAPSHAPE= now supports NONE, which specifies that no shape is displayed at
the ends of the box whiskers.
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n DATASKIN= enhances the visual appearance of the filled boxes.

n TIP= now supports NONE, which suppresses data tips from the plot.

n SAS 9.4M1 has the following enhancements:

xxxiv What’s New in SAS 9.4 Graph Template Language

o DISPLAYSTATS= now displays the DATAMAX, DATAMIN, and SUMWGT

statistics.
o LEGENDLABEL= now defaults to the Y= column label or name.
o URL= specifies an HTML page that is displayed when a box or an outlier
marker is selected.
n SAS 9.4M2 adds option GROUPORDER=REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
n SAS 9.4M3 modifies the BOXPLOTPARM statement to contribute its mean
markers to a discrete legend when the BOXPLOTPARM DISPLAY= option
includes MEAN, and TYPE=MARKER is specified in the DISCRETELEGEND
statement.
n SAS 9.4M5 has the following enhancements:

o DISPLAYSTATS= can be used with both vertical and horizontal box plots.
o CONNECTBREAK= specifies whether the connect line is broken for values
that have no observations.
o CAPSCALE= specifies a positive number to use as a multiplier for
determining the default width of the whisker caps.
o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.
o DISPLAY= now supports FILLPATTERN, which displays the box fill pattern.

BUBBLEPLOT:
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M3 adds option DRAWORDER=, which specifies whether the bubbles
are drawn according to bubble size or according to data order.
n SAS 9.4M5 has the following enhancements:

o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.

o DISPLAY= now supports FILLPATTERN, which displays the bubble fill
pattern.

CONTOURPLOTPARM:
n LEVELS= specifies a list of contour level values.

DENDROGRAM:
n TIP= now supports NONE, which suppresses data tips and URLs from the plot.
 Enhancements to SAS 9.4 Statements xxxv

DENSITYPLOT:
n CURVELABELSPLIT= specifies whether to split the curve label at the specified
split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the curve
label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the curve label block.
n WEIGHT= specifies a column that contains a density-curve calculation a priori
weight for each observation of the input data object.
n The KERNEL() distribution option WEIGHT= is changed to
WEIGHTFUNCTION=. This change enables the addition of the WEIGHT= option
in the DENSITYPLOT statement.

Note: The WEIGHT= option is not valid in the KERNEL() distribution option in
SAS 9.4. If you used the WEIGHT= option in the KERNEL() distribution option in
prior SAS releases, then you must change it to WEIGHTFUNCTION= in SAS 9.4
and later releases.

n SAS 9.4M2 has the following enhancements:

o GROUP= creates a separate density curve for each unique group value of the
specified column.
o INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.

DROPLINE:
n DATASKIN= enhances the appearance of the drop line.

n SAS 9.4M1 adds option DROPTO=BOTH, which draws one or more drop lines to
both the X and Y axes.

ELLIPSE:
n SAS 9.4M2 has the following enhancements:

o GROUP= creates a separate ellipse for each unique group value of the
specified column.
o INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.
n SAS 9.4M5 has the following enhancements:

o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.

o DISPLAY= now supports FILLPATTERN, which displays the band fill pattern.

ELLIPSEPARM:

SAS 9.4M5 has the following enhancements:

n FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.
xxxvi What’s New in SAS 9.4 Graph Template Language

n DISPLAY= now supports FILLPATTERN, which displays the ellipse fill pattern.

FRINGEPLOT:
n GROUP= creates a distinct set of lines for each unique group value in the
specified column.
n INCLUDEMISSINGGROUP= specifies whether missing values in the group
column are included in the plot.
n INDEX= specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
n TIP= now supports NONE, which suppresses data tips from the plot.

HEATMAPPARM:
n URL= specifies an HTML page to display when a rectangle is selected.

n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M2 adds option INCLUDEMISSINGCOLOR=, which specifies whether

missing values of the color-group variable or the color-response variable are
included in the plot.
n SAS 9.4M3 has the following enhancements:

o DISCRETEX= specifies whether the X axis is discrete when X= specifies a

numeric column.
o DISCRETEY= specifies whether the Y axis is discrete when Y= specifies a
numeric column.

HIGHLOWPLOT:
n CLIPCAP= specifies whether a special clip cap is displayed to indicate where
clipping occurred.
n CLIPCAPSHAPE= specifies the shape of the arrowhead on the clipped end of a
line.
n DATASKIN= enhances the visual appearance of the high-low chart filled bars or
lines.
n ENDCAPDISPLAYPOLICY= specifies the policy for displaying end caps when
end caps are present.
n TIP= now supports NONE, which suppresses data tips from the plot.

n SAS 9.4M2 adds option GROUPORDER=REVERSEDATA, which orders the

groups within a category in the reverse group-column data order.
n SAS 9.4M3 has the following enhancements:

o COLORMODEL= specifies a color ramp that is used with the

COLORRESPONSE= option.
o COLORRESPONSE= specifies the column or range attribute variable that is
used to map the bar or line colors.

HISTOGRAM:
n DATALABELTYPE= specifies the statistic to display at the end of each bar.
 Enhancements to SAS 9.4 Statements xxxvii

n DATASKIN= enhances the visual appearance of the filled bars.

n WEIGHT= specifies a column that contains a bin-width calculation a priori

weight for each observation of the input data object.
n SAS 9.4M1 has the following enhancements:

o The number of bins is limited to approximately 10,000. When the limit is

exceeded, SAS automatically adjusts the NBINS= or BINWIDTH= value to
set the number of bins to about 10,000.
o DISPLAY= now supports FILLPATTERN.
o FILLPATTERNATTRS= specifies the appearance of the pattern-filled bar
area.
n SAS 9.4M2 has the following enhancements:

o FILLTYPE= specifies whether the fill color is solid or is a gradient that

transitions from fully opaque to fully transparent.
o GROUP= creates a separate bar segment or bar for each unique group value
of the specified column.
o INCLUDEMISSINGGROUP= specifies whether missing values of the group
variable are included in the plot.
o The OUTLINEATTRS= option defaults are now consistent with that of
BARCHART.
n SAS 9.4M5 has the following enhancements:

o DATALABELATTRS= specifies the color and font attributes of the bar labels.
o FILLENDCOLOR=color specifies the end color of the bar fill color gradient.
o FILLTYPE=ALPHAGRADIENT replaces FILLTYPE=GRADIENT from the
previous releases. This option fills each bar with a color and a transparency
gradient that starts at the bar top with the initial transparency that is
assigned to that bar, and ends with full transparency at the bar baseline.
o FILLTYPE=COLORGRADIENT fills each bar with a color gradient that starts
with the fill color that is assigned to that bar and ends with the color that is
specified in the FILLENDCOLOR= option.
o FILLTYPE=GRADIENT is changed to an alias of
FILLTYPE=COLORGRADIENT.

HISTOGRAMPARM:
n DATALABELFITPOLICY= specifies a policy for avoiding collisions among the bin
labels when bin labels are displayed.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATASKIN= enhances the visual appearance of the filled bars.

n TIP= now supports NONE, which suppresses data tips from the plot.

n SAS 9.4M1 has the following enhancements:

xxxviii What’s New in SAS 9.4 Graph Template Language

o DISPLAY= now supports FILLPATTERN.

o FILLPATTERNATTRS= specifies the appearance of the pattern-filled bar
area.
n SAS 9.4M2 adds option FILLTYPE=, which specifies whether the fill color is
solid or is a gradient that transitions from fully opaque to fully transparent.
n SAS 9.4M5 has the following enhancements:

o FILLENDCOLOR=color specifies the end color of the bar fill color gradient.
o FILLTYPE=ALPHAGRADIENT replaces FILLTYPE=GRADIENT from the
previous releases. This option fills each bar with a color and a transparency
gradient that starts at the bar top with the initial transparency that is
assigned to that bar, and ends with full transparency at the bar baseline.
o FILLTYPE=COLORGRADIENT fills each bar with a color gradient that starts
with the fill color that is assigned to that bar and ends with the color that is
specified in the FILLENDCOLOR= option.
o FILLTYPE=GRADIENT is changed to an alias of
FILLTYPE=COLORGRADIENT.

LINECHART:
n SAS 9.4M2 adds option GROUPORDER=REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
n SAS 9.4M3 has the following enhancements:

o COLORMODEL= specifies a color ramp that is used with the

COLORRESPONSE= option.
o COLORRESPONSE= specifies the column or range attribute variable that is
used to map the line, marker, and fill colors.

LINEPARM:
n CURVELABELSPLIT= specifies whether to split the line label at the specified
split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the line
label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the line label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the line label block.

LOESSPLOT:
n CURVELABELSPLIT= specifies whether to split the curve label at the specified
split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the curve
label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
 Enhancements to SAS 9.4 Statements xxxix

n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are

inside the curve label block.

MODELBAND:
n SAS 9.4M2 has the following enhancements:

o ANTIALIAS= specifies whether anti-aliasing is turned off for a model band

plot.
o A confidence band that depicts confidence limits for individual predicted
values (CLI) for a weighted spline plot or regression plot is now displayed as
a high-low chart instead of a band.

NEEDLEPLOT:
n BASELINEATTRS= specifies the appearance of the baseline. This option enables
you to suppress the baseline by setting the line thickness to 0.
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n DATASKIN= enhances the appearance of the needle plot lines.

n MARKERATTRS= now supports transparency.

n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M2 adds option GROUPORDER=REVERSEDATA, which orders the

groups within a category in the reverse group-column data order.
n SAS 9.4M5 adds option DISPLAYBASELINE=, which specifies whether the
baseline is displayed.

PBSPLINEPLOT:
n CURVELABELSPLIT= specifies whether to split the curve label at the specified
split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the curve
label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the curve label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the curve label block.
n SAS 9.4M2 changes the range of the PBSPLINEPLOT DEGREE= option from 0–
174 to 0–10 .

PIECHART:
n TIP= now supports NONE, which suppresses data tips and URLs from the plot.
xl What’s New in SAS 9.4 Graph Template Language

n SAS 9.4M1 adds option CENTERFIRSTSLICE=, which specifies whether the first
pie slice is centered on the starting angle or starts on it.
n SAS 9.4M3 adds sub option TRANSPARENCY= to the PIECHART statement
FILLATTRS= option. The new TRANSPARENCY= sub option sets the
transparency of the other slice unless transparency is specified in the
OTHERSLICEOPTS= option.

POLYGONPLOT:
n SAS 9.4M2 has the following enhancements:

o ANTIALIAS= specifies whether anti-aliasing is turned off for a polygon plot.

o BACKLIGHT= specifies a back-light effect for the polygon label text.
n SAS 9.4M5 has the following enhancements:

o FILLPATTERNATTRS= specifies the appearance of the pattern-filled areas.

o DISPLAY= now supports FILLPATTERN, which displays the band fill pattern.

REFERENCELINE:
n CURVELABELSPLIT= specifies whether to split the reference line label at the
specified split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the
reference line label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the reference line label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the reference line label block.
n DATASKIN= enhances the visual appearance of the reference line.

n SAS 9.4M5 has the following enhancements:

o DISCRETELINETHICKNESS= specifies the line thickness as a fraction of the

midpoint spacing.
o USEDISCRETETHICKNESS= specifies that the line thickness is based on a
fraction of the midpoint spacing that is set by the
DISCRETELINETHICKNESS= option.

REGRESSIONPLOT:
n CURVELABELSPLIT= specifies whether to split the regression line label at the
specified split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the
regression line label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the regression line label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the regression line label block.
n SAS 9.4M2 changes the range of the DEGREE= regression option from 0–174 to
1–10.
 Enhancements to SAS 9.4 Statements xli

SCATTERPLOT:
n CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap. SAS
9.4M2 changes the default of this option to style reference
GraphError:CapStyle.
n FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
n JITTER= specifies whether to jitter data markers.

n JITTEROPTS= specifies options for managing jittering.

n LABELSTRIP= specifies whether leading and trailing blanks are stripped from
marker characters or fixed-position data labels before they are displayed in the
plot.
n MARKERATTRS= now supports transparency.

n MARKERCHARACTERPOSITION= specifies the justification of the marker

characters.
n MARKERFILLATTRS= specifies the appearance of the filled markers.

n MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.

n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M1 adds option OUTLINEDMARKERCHARACTERS=, which specifies

whether backlighting or a drop shadow is applied to the characters that are used
as marker symbols in order to enhance their appearance in the graph.
n SAS 9.4M2 has the following enhancements:

o CONTRIBUTEOFFSETS= specifies whether the plot's space requirements

contribute to the calculation of the axis offsets.
o The GROUPORDER= option supports REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
o The following options are replaced and considered deprecated:
n MARKERCOLORGRADIENT= is replaced with COLORRESPONSE=.
n MARKERSIZERESPONSE= is replaced with SIZERESPONSE=.
n MARKERSIZEMAX= is replaced with SIZEMAX=.
n MARKERSIZEMIN= is replaced with SIZEMIN=.
xlii What’s New in SAS 9.4 Graph Template Language

The new options are functionally the same as the deprecated options and
are more consistent with the other plot statements. The deprecated options
are still honored, but the new options are the preferred options.
o The OUTLINEDMARKERCHARACTERS= option is deprecated. It is still
honored, but the TEXTPLOT statement is now the preferred method for
creating scatter plots using text markers.
n SAS 9.4M3 adds option SUBPIXEL=, which specifies whether subpixel rendering
is used for image output when the scatter plot is rendered.
n SAS 9.4M5 adds option ERRORBARCAPSCALE=, which specifies a positive
number to use as a multiplier for determining the default width of the error-bar
caps.

SCATTERPLOTMATRIX:
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n DATASKIN= enhances the visual appearance of the plot markers.

n ELLIPSE= now supports the CLIP= option, which specifies whether clipped
confidence ellipses are included in the plot.
n LABELSTRIP= specifies whether leading and trailing blanks are stripped from
marker characters or fixed-position data labels before they are displayed in the
plot.
n MARKERATTRS= now supports transparency.

n MARKERCHARACTERPOSITION= specifies the justification of the marker

characters.
n MATRIXTYPE= specifies whether to display the full matrix, or just the upper or
lower triangle of the matrix.
n TIP= now supports NONE, which suppresses data tips from the plot.

n SAS 9.4M2 replaces option MARKERCOLORGRADIENT= with the

COLORRESPONSE= option. Option MARKERCOLORGRADIENT= is considered
deprecated. The COLORRESPONSE= option is more consistent with the other
plot statements. The MARKERCOLORGRADIENT= option is still honored, but
the COLORRESPONSE= option is the preferred option.
n SAS 9.4M3 adds option SUBPIXEL=, which specifies whether subpixel rendering
is used for image output when the scatter plots are rendered.

SERIESPLOT:
n CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
 Enhancements to SAS 9.4 Statements xliii

n DATALABELSPLIT= specifies whether to split the data labels at specified split

characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n CURVELABELSPLIT= specifies whether to split the series line label at the
specified split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the
series line label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the series line label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the series line label block.
n DATASKIN= enhances the visual appearance of the series plot lines.

n FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both

fills and outlines.
n MARKERATTRS= now supports transparency.

n MARKERFILLATTRS= specifies the appearance of the filled markers.

n MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.

n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M2 has the following enhancements:

o The GROUPORDER= option supports REVERSEDATA, which orders the

groups within a category in the reverse group-column data order.
o LINECOLORGROUP= specifies a column that determines the line colors for a
grouped plot independently of the GROUP= column.
o LINEPATTERNGROUP= specifies a column that determines the line patterns
for a grouped plot independently of the GROUP= column.
o MARKERCOLORGROUP= specifies a column that determines the marker
colors for a grouped plot independently of the GROUP= column.
o MARKERSYMBOLGROUP= specifies a column that determines the marker
symbols for a grouped plot independently of the GROUP= column.
n SAS 9.4M3 has the following enhancements:

o ARROWHEADPOSITION= specifies a position for arrowheads.

o ARROWHEADSCALE= specifies an arrowhead scale factor based on the
thickness of the arrow line.
o ARROWHEADSHAPE= specifies a shape for arrowheads.
xliv What’s New in SAS 9.4 Graph Template Language

o COLORMODEL= specifies a color ramp that is to be used with the

COLORRESPONSE= option.
o COLORRESPONSE= specifies the column or range attribute variable that is
used to map the line and marker colors.
o LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
o LINETHICKNESSMAXRESPONSE= specifies the response value that
corresponds to the maximum line thickness.
o LINETHICKNESSMIN= specifies the minimum line thickness when a
response variable is used to determine the line thickness.
o LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.
o SPLINEPOINTS= specifies a multiplier to apply to the time interval that is in
effect for the INTERVAL= axis option.
o SPLINETYPE= specifies the type of spline interpolation that is used to draw
the series line.
o The LINECOLORGROUP=, LINEPATTERNGROUP=,
MARKERCOLORGROUP=, and MARKERSYMBOLGROUP= options now
support a discrete attribute map variable.
n SAS 9.4M4 has the following enhancement and change:

o LINEJOIN= specifies how to treat the vertices of the series line.

o BREAK= breaks the series line at missing values of the Y variable only. In
earlier SAS releases, the BREAK= option breaks the series line at missing
values of the X or Y variable.

Note: If you want to break the series line at missing X or Y variable values in
SAS 9.4M4 and later releases, set the Y variable value to missing when the X
variable value is missing. See “Programs That Use the SERIESPLOT
Statement” on page xix.

n SAS 9.4M5 has the following enhancements:

o ERRORBARATTRS= specifies the attributes of the error bars that are

associated with the data points.
o ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap.
o XERRORLOWER= specifies values for the lower endpoints on the X error
bars.
o XERRORUPPER= specifies values for the upper endpoints on the X error
bars.
o YERRORLOWER= specifies values for the lower endpoints on the Y error
bars.
o YERRORUPPER= specifies values for the upper endpoints on the Y error
bars.
 Enhancements to SAS 9.4 Statements xlv

o ERRORBARCAPSCALE= specifies a positive number to use as a multiplier

for determining the default width of the error-bar caps.

STEPPLOT:
n CLUSTERAXIS= specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n DATASKIN= enhances the appearance of the step plot lines.

n ERRORBARCAPSHAPE= specifies whether the error bars have a serif cap.

n CURVELABELSPLIT= specifies whether to split the step line label at the

specified split characters.
n CURVELABELSPLITCHAR= specifies one or more characters on which the step
line label can be split if needed.
n CURVELABELSPLITCHARDROP= specifies whether the split characters are
included in the step line label text.
n CURVELABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the step line label block.
n FILLEDOUTLINEDMARKERS= specifies whether markers are drawn with both
fills and outlines.
n MARKERATTRS= now supports transparency.

n MARKERFILLATTRS= specifies the appearance of the filled markers.

n MARKEROUTLINEATTRS= specifies the appearance of the marker outlines.

n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M2 has the following enhancements:

o ERRORBARCAPSHAPE= defaults to style reference GraphError:CapStyle.

o The GROUPORDER= option supports REVERSEDATA, which orders the
groups within a category in the reverse group-column data order.
n SAS 9.4M3 has the following enhancements:

o ARROWHEADPOSITION= specifies a position for arrowheads.

o ARROWHEADSCALE= specifies an arrowhead scale factor based on the
thickness of the arrow line.
o ARROWHEADSHAPE= specifies a shape for arrowheads.
o LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
xlvi What’s New in SAS 9.4 Graph Template Language

o LINETHICKNESSMAXRESPONSE= specifies the response value that

corresponds to the maximum line thickness.
o LINETHICKNESSMIN= specifies the minimum line thickness when a
response variable is used to determine the line thickness.
o LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.
n SAS 9.4M5 adds option ERRORBARCAPSCALE=, which specifies a positive
number to use as a multiplier for determining the default width of the error-bar
caps.

SURFACEPLOTPARM:
n SAS 9.4M2 replaces option SURFACECOLORGRADIENT= with the
COLORRESPONSE= option. The SURFACECOLORGRADIENT= option is
considered deprecated. The COLORRESPONSE= option is more consistent with
the other plot statements. The SURFACECOLORGRADIENT= option is still
honored, but the COLORRESPONSE= option is the preferred option.

TEXTPLOT:
n SAS 9.4M3 has the following experimental options:

o OUTFILE= specifies a file for storing information about the text bounding-
box for each text value in the column specified in the OUTID= option.
o OUTID= specifies a column that contains text values to write to the file
specified in the OUTFILE= option.
n SAS 9.4M5 has the following enhancements:

o POSITION= now accepts the name of a keyword column.

o POSITIONOFFSETX= specifies the positive or negative X offset for an
individual text item.
o POSITIONOFFSETY= specifies the positive or negative Y offset for an
individual text item.

VECTORPLOT:
n DATALABELSPLIT= specifies whether to split the data labels at specified split
characters.
n DATALABELSPLITCHAR= specifies one or more characters on which the data
labels can be split if needed.
n DATALABELSPLITCHARDROP= specifies whether the split characters are
included in the data labels.
n DATALABELSPLITJUSTIFY= specifies the justification of the strings that are
inside the data label blocks.
n DATASKIN= enhances the appearance of the vector plot lines.

n TIP= now supports NONE, which suppresses data tips from the plot.

n SAS 9.4M3 has the following enhancements:

o COLORMODEL= specifies a color ramp that is used with the

COLORRESPONSE= option.
 Enhancements to SAS 9.4 Statements xlvii

o COLORRESPONSE= specifies the column or range attribute variable that is

used to map the line colors.
o LINETHICKNESSMAX= specifies the maximum line thickness when a
response variable is used to determine the line thickness.
o LINETHICKNESSMAXRESPONSE= specifies the response value that
corresponds to the maximum line thickness.
o LINETHICKNESSMIN= specifies the minimum line thickness when a
response variable is used to determine the line thickness.
o LINETHICKNESSRESPONSE= specifies a response column or range attribute
variable that is used to map a line thickness to each group value.

WATERFALLCHART:
n BARLABELFITPOLICY= specifies a policy for avoiding collisions among the bar
labels when bar labels are displayed.
n BASELINEATTRS= specifies the appearance of the baseline. This option enables
you to suppress the baseline by setting the line thickness to 0.
n TIP= now supports NONE, which suppresses data tips and URLs from the plot.

n SAS 9.4M3 adds option COLORSTAT=, which specifies the statistic to use for
computing the response colors.
n SAS 9.4M5 adds option CONNECTBREAK=, which specifies whether the
connect line is broken for values that have no observations.

Text Statement Enhancements

ENTRY:
n SAS 9.4M1 removes the 512-character limit on the length of the entry text.

ENTRYFOOTNOTE:
n HALIGNCENTER= specifies whether the footnote is centered automatically by
the system or is always centered in the graph area.
n SAS 9.4M1 removes the 512-character limit on the length of the footnote text.

ENTRYTITLE:
n HALIGNCENTER= specifies whether the title is centered automatically by the
system or is always centered in the graph area.
n SAS 9.4M1 removes the 512-character limit on the length of the title text.
xlviii What’s New in SAS 9.4 Graph Template Language

Draw Statement Enhancements

SAS 9.4M1 adds option URL= to the BEGINPOLYGON, BEGINPOLYLINE,
DRAWTEXT, DRAWLINE, DRAWARROW, DRAWRECTANGLE, DRAWIMAGE, and
DRAWOVAL statements. The new URL= option specifies an HTML page that is
displayed when the output of the draw statement is selected.

Axis Statement Enhancements

LAYOUT OVERLAY:
n LABELFITPOLICY= specifies a policy for fitting axis labels in the available
space.
n LABELPOSITION= specifies the position of the axis label.

n LABELSPLITCHAR= specifies one or more characters on which the axis labels

can be split if needed.
n LABELSPLITCHARDROP= specifies whether the split characters are included in
the axis labels.
n LABELSPLITJUSTIFY= specifies the justification of the strings that are inside
the axis label blocks.
n TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick
values that are displayed on the Y and Y2 axes.
n TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values
that are displayed on the X and X2 axes.
n SAS 9.4M3 adds option LINEEXTENT=, which specifies the extent of the axis
line.
n SAS 9.4M5 adds option LABELFITPOLICY=STACKED, which displays the Y or
Y2 axis label vertically as stacked letters.
n DISCRETEOPTS= supports the following new features for discrete axes:

o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATELAWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, SPLITTHIN, and SPLITROTATE.
o TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
 Enhancements to SAS 9.4 Statements xlix

o TICKVALUESPLITCHAR= specifies a list of characters on which the tick

values can be split if needed.
o TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
o TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
o SAS 9.4M5 has the following enhancements:
n TICKVALUEROTATION=DIAGONAL2 rotates the tick values to a –45-
degree diagonal position.
n TICKVALUEFITPOLICY=STACKEDALWAYS always displays the tick
values vertically as stacked letters.
n TICKVALUEFITPOLICY=STACKEDALWAYSTHIN is the same as
STACKEDALWAYS, except that thinning is performed when the tick
values do not fit the available space.
n LINEAROPTS= supports the following new features for linear axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables
you to specify the type of scale that you want to extract.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M1 adds option INCLUDERANGES=, which specifies the ranges for a
broken axis.
o SAS 9.4M2 has the following enhancements:
n The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
n The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor
grid lines.
n The MINORTICKCOUNT= option for linear axes defaults to one minor
tick and two intervals.
l What’s New in SAS 9.4 Graph Template Language

o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates

the tick values to a –45-degree diagonal position.
n LOGOPTS= supports the following new features for log axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o TICKVALUELIST= specifies the tick values for a log axis as a space-
separated list.
o TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
o VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
o SAS 9.4M2 has the following enhancements:
n The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
n The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor
grid lines.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
n TIMEOPTS= supports the following new features for time axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKINTERVAL= specifies the time interval between minor ticks.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M1 adds option INCLUDERANGES=, which specifies the ranges for a
broken axis.
o SAS 9.4M2 has the following enhancements:
n The MINORGRID= option defaults to style reference
GraphMinorGridLines:DisplayOpts.
n The MINORGRIDATTRS= option defaults to style element
GraphMinorGridLines in order to visually contrast the major and minor
grid lines.
o SAS 9.4M3 adds option INTERVALMULTIPLIER=, which specifies a multiplier
to apply to the time interval that is in effect for the axis.
 Enhancements to SAS 9.4 Statements li

o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates

the tick values to a –45-degree diagonal position.

LAYOUT OVERLAY3D:
n LINEAROPTS= supports the following new features for linear axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables
you to specify the type of scale that you want to extract.
o TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
o SAS 9.4M2 changes the MINORGRIDATTRS= option default to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
n TIMEOPTS= supports the following new features for time axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKINTERVAL= specifies the time interval between minor ticks.
o SAS 9.4M2 changes the MINORGRIDATTRS= option default to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
o SAS 9.4M3 adds option INTERVALMULTIPLIER=, which specifies a multiplier
to apply to the time interval that is in effect for the axis.

LAYOUT OVERLAYEQUATED:
n MINORGRID= specifies whether grid lines are displayed at the minor tick values.

n MINORGRIDATTRS= specifies the attributes of the minor grid lines.

n MINORTICKCOUNT= specifies the number of minor ticks that are displayed on

the axis.
n MINORTICKS= specifies whether the minor tick marks are displayed on the axis.

n TICKVALUEFITPOLICY= now supports policies NONE and ROTATEALWAYS.

n TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables

you to specify the type of scale that you want to extract.
lii What’s New in SAS 9.4 Graph Template Language

n VIEWMAX= specifies the maximum data value to include in the display (the
value might be adjusted by the threshold calculation).
n VIEWMIN= specifies the minimum data value to include in the display (the
value might be adjusted by the threshold calculation).
n SAS 9.4M3 adds option LINEEXTENT=, which specifies the extent of the axis
lines.

LAYOUT LATTICE:
n LABELFITPOLICY= specifies a policy for fitting axis labels in the available
space.
n LABELPOSITION= specifies the position of the axis label.

n LABELSPLITCHAR= specifies one or more characters on which the axis labels

can be split if needed.
n LABELSPLITCHARDROP= specifies whether the split characters are included in
the axis labels.
n LABELSPLITJUSTIFY= specifies the justification of the strings that are inside
the axis label blocks.
n TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick
values that are displayed on the Y and Y2 axes.
n TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values
that are displayed on the X and X2 axes.
n SAS 9.4M5 adds option LABELFITPOLICY=STACKED, which displays the Y or
Y2 axis label vertically as stacked letters.
n DISCRETEOPTS= supports the following new features for discrete axes:

o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATEALWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, and SPLITTHIN.
o TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o TICKVALUESPLITCHAR= specifies a list of characters on which the tick
values can be split if needed.
o TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
o TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
o SAS 9.4M5 has the following enhancements:
 Enhancements to SAS 9.4 Statements liii

n TICKVALUEROTATION=DIAGONAL2 rotates the tick values to a –45-

degree diagonal position.
n TICKVALUEFITPOLICY=STACKEDALWAYS always displays the tick
values vertically as stacked letters.
n TICKVALUEFITPOLICY=STACKEDALWAYSTHIN is the same as
STACKEDALWAYS, except that thinning is performed when the tick
values do not fit the available space.
n LINEAROPTS= supports the following new features for linear axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables
you to specify the type of scale that you want to extract.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates
the tick values to a –45-degree diagonal position.
n LOGOPTS= supports the following new features for log axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o TICKVALUELIST= specifies the tick values for a log axis as a space-
separated list.
o TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
o VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
liv What’s New in SAS 9.4 Graph Template Language

n TIMEOPTS= supports the following new features for time axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKINTERVAL= specifies the time interval between minor ticks.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M3 has the following enhancements:
n TICKVALUEFORMAT= specifies how to format the values for major tick
marks.
n INTERVALMULTIPLIER= specifies a multiplier to apply to the time
interval that is in effect for the axis.
o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates
the tick values to a –45-degree diagonal position.

LAYOUT DATALATTICE and LAYOUT DATAPANEL:

n LABELFITPOLICY= specifies a policy for fitting axis labels in the available
space.
n LABELPOSITION= specifies the position of the axis label.

n LABELSPLITCHAR= specifies one or more characters on which the axis labels

can be split if needed.
n LABELSPLITCHARDROP= specifies whether the split characters are included in
the displayed axis labels.
n LABELSPLITJUSTIFY= specifies the justification of the strings that are inside
the axis label blocks.
n TICKVALUEHALIGN= specifies the horizontal alignment for all of the tick
values that are displayed on the Y and Y2 axes.
n TICKVALUEVALIGN= specifies the vertical alignment for all of the tick values
that are displayed on the X and X2 axes.
n SAS 9.4M5 adds option LABELFITPOLICY=STACKED, which displays the Y or
Y2 axis label vertically as stacked letters.
n DISCRETEOPTS= supports the following new features for discrete axes:

o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= supports new fit policies: ROTATE,
ROTATEALWAYS, ROTATEALWAYSDROP, SPLIT, SPLITALWAYS,
SPLITALWAYSTHIN, and SPLITTHIN.
o TICKVALUELIST= specifies the list of tick values that are displayed on the
axis.
 Enhancements to SAS 9.4 Statements lv

o TICKVALUEROTATION= specifies how the tick values are rotated on the X

and X2 axes.
o TICKVALUESPLITCHAR= specifies a list of characters on which the tick
values can be split if needed.
o TICKVALUESPLITJUSTIFY= specifies justification of the strings that are
inside the tick value block.
o TICKVALUESPLITCHARDROP= specifies whether the split characters are
included in the displayed tick values.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
o SAS 9.4M5 has the following enhancements:
n TICKVALUEROTATION=DIAGONAL2 rotates the tick values to a –45-
degree diagonal position.
n TICKVALUEFITPOLICY=STACKEDALWAYS always displays the tick
values vertically as stacked letters.
n TICKVALUEFITPOLICY=STACKEDALWAYSTHIN is the same as
STACKEDALWAYS, except that thinning is performed when the tick
values do not fit the available space.
n LINEAROPTS= supports the following new features for linear axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o MINORTICKS= specifies whether the minor tick marks are displayed on the
axis.
o TICKDISPLAYLIST= specifies the text that is displayed for the tick values
that are defined in the TICKVALUELIST= option.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEFORMAT= now supports EXTRACTSCALETYPE=, which enables
you to specify the type of scale that you want to extract.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M2 changes the MINORGRIDATTRS= option default to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates
the tick values to a –45-degree diagonal position.
n LOGOPTS= supports the following new features for log axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
lvi What’s New in SAS 9.4 Graph Template Language

o MINORGRIDATTRS= specifies the attributes of the minor grid lines.

o MINORTICKCOUNT= specifies the number of minor ticks that are displayed
on the axis.
o TICKVALUELIST= specifies the tick values for a log axis as a space-
separated list.
o TICKVALUEPRIORITY= specifies whether the TICKVALUELIST=
specification can extend the axis data range.
o VALUETYPES= specifies how the VIEWMIN=, VIEWMAX=, and
TICKVALUELIST= option values are interpreted.
o SAS 9.4M2 changes the MINORGRIDATTRS= option default to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
o SAS 9.4M3 adds option TICKVALUEFORMAT=, which specifies how to
format the values for major tick marks.
n TIMEOPTS= supports the following new features for time axes:

o MINORGRID= specifies whether grid lines are displayed at the minor tick
values.
o MINORGRIDATTRS= specifies the attributes of the minor grid lines.
o MINORTICKINTERVAL= specifies the time interval between minor ticks.
o TICKVALUEFITPOLICY= now supports policies NONE and
ROTATEALWAYS.
o TICKVALUEROTATION= specifies how the tick values are rotated on the X
and X2 axes.
o SAS 9.4M2 changes the MINORGRIDATTRS= option default to style element
GraphMinorGridLines in order to visually contrast the major and minor grid
lines.
o SAS 9.4M3 adds option INTERVALMULTIPLIER=, specifies a multiplier to
apply to the time interval that is in effect for the axis.
o SAS 9.4M5 adds option TICKVALUEROTATION=DIAGONAL2, which rotates
the tick values to a –45-degree diagonal position.

Marker Definition Statement Enhancements

SYMBOLCHAR:

SAS 9.4M3 has the following enhancements:

n HOFFSET= and VOFFSET= now move the marker character within the marker
character bounding box. The bounding box position remains centered on the
data point.
n Offsets applied to the marker character with HOFFSET= and VOFFSET= are
also applied to the marker character that is displayed in the legend.
 Documentation Enhancements lvii

Discrete Attribute Map Enhancements

The following enhancements have been made to discrete attribute maps:

VALUE statement:
n TEXTATTRS= specifies the text attributes to use when an attribute map is
applied to text in a plot.

SAS 9.4M1 has the following enhancements:

n You can now specify the attribute mapping information for a discrete attribute
map in a SAS data set. You now have an alternative to coding your mapping
information in a DISCRETEATTRMAP block in your template. For more
information, see SAS ODS Graphics: Procedures Guide.
n DISCRETEATTRMAP statement:

o DISCRETELEGENDENTRYPOLICY= specifies whether the items that the

associated plot contributes to a discrete legend are items that appear only in
the data or items that are defined only in the attribute map.
n DISCRETEATTRVAR statement:

o ATTRMAP= now accepts a name that is specified in the ID column in a

discrete attribute map data set. This feature enables you to create a discrete
attribute map variable for a discrete attribute map that is defined in a SAS
data set. To resolve the attribute map name in that case, you must specify
the name of the attribute map data set in the DATTRMAP= option in the
SGRENDER statement that renders the graph. For information about
discrete attribute map data sets and the SGRENDER statement
DATTRMAP= option, see SAS ODS Graphics: Procedures Guide.
n VALUE statement:

o FILLATTRS= now supports the TRANSPARENCY= fill option.

o LINEATTRS= now supports the THICKNESS= line option.
o MARKERATTRS= now supports the SIZE=, TRANSPARENCY=, and
WEIGHT= marker options.

Documentation Enhancements
In SAS 9.4M2, SAS Graph Template Language: User’s Guide is reorganized to make it
easier to find information about how to use the Graph Template Language.
lviii What’s New in SAS 9.4 Graph Template Language
 1

PART 1

Fundamentals

Chapter 1
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2
 3

1
Overview

Graph Template Language (GTL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

GTL and the Output Delivery System (ODS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
A Quick Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Template Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Run-Time Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Components of a Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Graphical Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Legends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Flexible Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Expressions and Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Dynamic Variables and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
ODS GRAPHICS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
ODS Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
About the Examples in This Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Examples and Resources on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Graph Template Language (GTL)

GTL and the Output Delivery System (ODS)

The SAS Graph Template Language (GTL) is an extension to the Output Delivery
System (ODS) that enables you to create sophisticated graphics. For example,
 4 Chapter 1 / Overview

using GTL, you can generate Model-Fit plots, Distribution Plots, Comparative plots,
Prediction Plots, and more.

The graphics produced by GTL are generated by template definitions that control
the graph format and appearance and specify the variable roles to represent in the
graph display. The graph can then be rendered by associating the templates with a
data source.
n The GTL templates are defined with PROC TEMPLATE. GTL includes
conditional statements that can be used to determine what graph features are
rendered. It also includes layout statements that specify the arrangement of
graph features, plot statements that request specific plot types (such as
histograms and scatter plots), and text and legend statements that specify
titles, footnotes, legends, and other text-based graph elements.
n The GTL templates are rendered using the SGRENDER procedure, which
specifies a data source that contains appropriate data values and the template
to use for rendering the graph.
n You can also modify predefined GTL templates that SAS delivers for use on the
SAS statistical procedures. For information about modifying existing templates,
refer to ODS Graphics Template Modification in SAS/STAT User’s Guide.

This manual provides a complete reference to the Graph Template Language. For
detailed usage information, consult the SAS Graph Template Language: User’s Guide.

Note: If you are also a SAS/GRAPH user, then you might want to consult the SAS
Graph Template Language: User’s Guide to learn about some of the distinctions
between ODS Graphics and SAS/GRAPH.

A Quick Example
The data set Sashelp.Class is delivered with SAS. It includes data columns named
Height and Weight, which store height and weight measures for a small sample of
subjects. The Graph Template Language can be used to generate a histogram that
shows the distribution of weight recorded in that data set:
 Graph Template Language (GTL) 5

The following SAS program produces the graph:

proc template;
define statgraph histogram;
begingraph;
layout overlay;
histogram weight;
endlayout;
endgraph;
end;
run;

ods graphics / width=450px;

proc sgrender data=sashelp.class
template=histogram;
run;
n The DEFINE STATGRAPH statement on PROC TEMPLATE opens a definition
block for defining a graphics template named HISTOGRAM. The HISTOGRAM
template is stored in the template folder (also called the “template store,” by
default located in Sasuser.Templat).
n The template definition for HISTOGRAM specifies two GTL statements within a
BEGINGRAPH/ENDGRAPH block: LAYOUT OVERLAY and HISTOGRAM.
n The LAYOUT OVERLAY statement is one of the most fundamental layout
statements. It can overlay the results of one or more plot statements, each of
which shares the same plot area, axes, and legends. The layout in this example
specifies only a single element: a HISTOGRAM with bars showing the
distribution of observations of the data column named Weight.
n The ENDLAYOUT statement ends the layout block, the ENDGRAPH statement
ends the graph definition, and the END statement ends the template definition.
 6 Chapter 1 / Overview

n The ODS GRAPHICS statement uses the WIDTH= option to set a width for the
output graph. Because the HEIGHT= option is not specified, GTL manages the
graph’s aspect ratio and sets an appropriate height.
n The DATA= option on PROC SGRENDER specifies Sashelp.Class as the data
source for the graph. TEMPLATE= specifies HISTOGRAM as the template
definition to use for rendering the graph.

Template Compilation
A GTL template describes the structure and appearance of a graph to be produced,
similar to how a TABLE template describes the organization and content of a table.

All templates are stored, compiled programs. The following source program
produces a simple GTL template named SCATTER:
proc template;
define statgraph scatter;
begingraph;
layout overlay;
scatterplot x=height y=weight;
endlayout;
endgraph;
end;
run;

When this code is submitted, the statement keywords and options are parsed, just
as with any other procedure. If no syntax error is detected, then an output template
named SCATTER is created and stored in the default template folder
Sasuser.Templat. No graph is produced. Note the following:
n Any required arguments in the template must be specified. In this example, X=
and Y= in the SCATTERPLOT statement must specify variables for the analysis,
but no checking for the existence of these variables is done at compile time.
(Unlike other SAS procedures, PROC TEMPLATE does not perform a compile
and then run sequence, which includes variable validation.)
n No reference to an input data set appears in the template.

Run-Time Actions
To produce a graph, a GTL template must be bound to a data source using the
SGRENDER procedure. The following example uses SGRENDER to bind the
SCATTER template to the SAS data set Sashelp.Class, which is delivered with SAS:
proc sgrender data=sashelp.class
template=scatter;
run;
 Components of a Graph 7

Generally, an ODS data object is constructed by comparing the template references

to column names with columns that exist in the current data set. In the current
example, Sashelp.Class contains columns named Height and Weight. Because these
column names match the columns that are named in template SCATTER, columns
Height and Weight are added to the data object. The other columns in
Sashelp.Class are ignored. (It is possible for a template to define new computed
columns based on existing columns.)

After all the observations have been read, the data object and template definition
are passed to a graph renderer, which produces the graph image. The image is then
automatically integrated into the ODS destination. The visual properties of the
graph are determined by the ODS style that is in effect.

Note: Template SCATTER is a restrictive definition: it can create a plot only with
columns named Height and Weight. A GTL template can be made more flexible by
introducing dynamic variables or macro variables that supply variables and other
information at run time. For more information, see “Flexible Templates” on page
14.

Components of a Graph
In general, a graph is made of up of the following parts:
n titles and footnotes

n one or more cells that contain a composite of one or more plots

n legends, which can reside inside or outside a cell

The following figure shows the different parts of a graph:

8 Chapter 1 / Overview

Figure 1.1 Components of a Graph

1 Graph
a visual representation of data. The graph can contain titles, footnotes, legends,
and one or more cells that have one or more plots.
2 Cell
a distinct rectangular subregion of a graph that can contain plots, text, and
legends.
3 Title
descriptive text that is displayed above any cell or plot areas in the graph.
4 Plot
a visual representation of data such as a scatter plot, a series line, a bar chart, or
a histogram. Multiple plots can be overlaid in a cell.
5 Legend
refers collectively to the legend border, one or more legend entries (where each
entry has a symbol and a corresponding label) and an optional legend title.
6 Axis
refers collectively to the axis line, the major and minor tick marks, the major tick
mark values, and the axis label. Each cell has a set of axes that are shared by all
the plots in the cell. In multi‐cell graphs, the columns and rows of cells can
share common axes if the cells have the same data type.
7 Footnote
 Graphical Layouts 9

descriptive text that is displayed below any cell or plot areas in the graph.

Graphical Layouts
One of most powerful features of GTL is the syntax built around hierarchical
statement blocks called “layouts.” The outermost layout block determines
n The overall organization of the graph—whether it uses a single-cell or a multi-
cell display.
n What statements are allowed in the block. Generally, layout blocks can contain
plots, lines of text, a legend, or even another layout.
n How the contained statements interact.

Table 1.1 Outermost Layouts in GTL

Layout Description

OVERLAY General purpose layout for displaying 2-D plots in a single-

cell.

OVERLAY3D Layout for displaying 3-D plots in a single-cell.

OVERLAYEQUATED Specialized OVERLAY with equated axes.

REGION General purpose layout for displaying single-cell graphs

that does not use axes.

GRIDDED Basic grid of plots. All cells are independent.

LATTICE Advanced multi-cell layout. Axes can be shared across

columns or rows and be external to grid. Many grid labeling
and alignment features.

DATALATTICE Generates a classification panel from the values of 1 or 2

classifiers.

DATAPANEL Generates a classification panel from the values of n

classifiers.

GLOBALLEGEND Specialized layout for creating a compound legend that

contains multiple discrete legends.

For example, the following graph is a two-cell graph produced using the LAYOUT
LATTICE statement as the outermost template in the layout.
10 Chapter 1 / Overview

The LAYOUT LATTICE statement is typically used to create a multi-cell layout of

plots that are aligned across columns and rows. In the following template, which
produced the graph, plot statements are specified within nested LAYOUT
OVERLAY statements. Thus, the LATTICE automatically aligns the plot areas and
tick display areas in the plots. The LATTICE layout is a good layout to choose when
you want to compare the results of related plots.
proc template;
define statgraph lattice;
begingraph;
entrytitle "Car Performance Profile";
layout lattice / border=true pad=10 opaque=true
rows=1 columns=2 columngutter=3;
layout overlay;
scatterplot x=horsepower y=mpg_city /
group=origin name="cars";
regressionPlot x=horsepower y=mpg_city / degree=2;
endlayout;

layout overlay;
scatterplot x=weight y=mpg_city / group=origin;
regressionPlot x=weight y=mpg_city / degree=2;
endlayout;

sidebar;
discretelegend "cars";
endsidebar;
endlayout;
endgraph;
end;
run;

For detailed information about each layout, see the chapter for that layout type.
 Axes 11

Plots
The plots in GTL are classified in different ways, depending on the context of the
discussion.

Within layout blocks, plots are often classified according to graphical dimension:
whether they are projected in two or three visual dimensions. Thus, plots in GTL are
often referred to as 2-D or 3-D plots, based on their graphical dimensions, not their
data dimensions.

Relative to their input data, plots are classified according to the statements that
calculate summary statistics from raw input data, and those that use calculated
statistics as input parameters on the plot statement. Thus, many GTL plot
statements have two versions: BARCHART and BARCHARTPARM, HISTOGRAM
and HISTOGRAMPARM, and so on. The main distinction between such plots is the
nature of the input data that they accept:
n The “non-parm” version (for example, BARCHART) computes its values from
raw, unsummarized data. For example, a BARCHART computes the summary
values it needs for the bars in the chart. Such plots are often referred to as
“computed plots.”
n The “parm” version (for example, BARCHARTPARM) does not summarize or
compute values from the input data but instead simply renders the input data it
is given. Thus, the input data must be pre-summarized, perhaps by a SAS
procedure. The “parm” version of plots, often referred to as “parameterized
plots,” produce the same result as the non-parm version. However, they do not
perform the calculations or data summarizations needed to achieve the result.

Chapter 5, “Key Concepts for Using Plots,” on page 201 discusses general concepts
that apply across plot types. For detailed information about a particular plot, see
the chapter for that plot.

Axes
GTL uses various criteria to determine the displayed axis features for a graph.
Generally, axis features are based on the following criteria:
n the layout type

n the order of plot statements in the layout and the options specified on those
statements
n the use of “primary” and “secondary” axes on the plots (when secondary axes
are supported)
12 Chapter 1 / Overview

n the plot type

n the column(s) of data that contribute to defining the axis range

n the data formats for the contributing data columns

Depending on the layout type, 2-D plots can have up to four independent axes that
can be displayed: X, Y, X2, and Y2. The X and Y axes are considered the primary
axes, and the X2 and Y2 axes are considered the secondary axes. By default, the X2
and Y2 axes are not displayed. When requested, the secondary axes can be
displayed as copies of the primary axes, or data can be mapped separately to them.
The following figure identifies the X, Y, X2, and Y2 axes.

All 3-D plots display the standard X, Y, and Z axes.

For more information about axis features in GTL, see Chapter 7, “Axis Features in
Layouts,” on page 1039.
 Legends 13

Legends
Many plot statements support a GROUP= option that partitions the data into
unique values, performs separate analysis, if necessary, and automatically assigns
distinct visual properties to each group value. The visual properties of group values
are defined by the style in effect.

Legends are not automatically displayed for plots with group values. Rather, an
appropriate legend statement must be added to the template to generate the
desired legend. In the following example, a legend is added to display markers and
line patterns that show the association between the group values from a scatter
plot and corresponding linear regression lines. The example shows the mechanism
that GTL uses to associate a legend with its corresponding plot(s): a name is
assigned to each plot that must be represented in the legend, and these names are
then used as arguments for the legend statement (in this case, MERGEDLEGEND).
proc template;
define statgraph scatterfit;
begingraph;
entrytitle "Linear Regression By Gender";
layout overlay;
scatterplot x=height y=weight / group=sex name="scat";
regressionplot x=height y=weight/ group=sex name="reg";
mergedlegend "scat" "reg" / border=true;
endlayout;
endgraph;
end;
run;

For more information about managing legends in GTL, see SAS Graph Template
Language: User’s Guide.
 14 Chapter 1 / Overview

Flexible Templates
Several features in GTL can make template definitions less restrictive on input data
and more general in nature. These features enable a single compiled template to
produce many output variations.

Expressions and Functions

In GTL, expressions can be used to compute constants and data columns. The
expressions must be enclosed in an EVAL construct. Within the expression, you can
use DATA step functions, arithmetic operators, and other special functions
supported by GTL.

Expressions are also useful in text statements like ENTRY and ENTRYTITLE. Both
of these statements support rich text and have special text commands such as
{SUP}, {SUB}, and {UNICODE}, which enable subscripting, superscripting, and
Unicode characters.

The following template shows how the ± symbol is included in the title line using
its hexadecimal Unicode value. Also, new data columns are computed for the upper
and lower error bars of the scatter plot, based on the input columns MeanWeight
and STDERR.
proc template;
define statgraph expression;
begingraph;
entrytitle "Errorbars show " {unicode "00B1"x} "2 SE";
layout overlay;
scatterplot x=age y=meanweight /
yerrorlower=eval(meanweight - 2*stderr)
yerrorupper=eval(meanweight + 2*stderr);
seriesplot x=age y=meanweight;
endlayout;
endgraph;
end;
run;

For more information about using expressions, see Chapter 21, “Expressions,” on
page 1539. For more information about using functions, see Chapter 22, “Functions,”
on page 1543.

Dynamic Variables and Macro Variables

An extremely useful technique for generalizing templates is to define dynamic
variables, macro variables, or both. The dynamic variables and macro variables
 Flexible Templates 15

resolve when the template is executed. The following PROC TEMPLATE

statements can be used in a DEFINE STATGRAPH block:

Template
Statement Purpose Value supplied by...

DYNAMIC defines one or more dynamic either of the following:

variables n DYNAMIC= suboption of ODS=
option of FILE PRINT
n DYNAMIC statement of PROC
SGRENDER

MVAR defines one or more macro %LET or CALL SYMPUT( )

variables

NMVAR defines one or more macro %LET or CALL SYMPUT( )

variables that resolve to a
number or numbers

NOTES provides information about the user-supplied text

graph definition

The following example defines a template named DYNAMICS that can create a
histogram and density plot for any variable. It defines both macro variables and
dynamic variables for run-time substitution. No data-dependent information is hard
coded in the template.

Note: You can initialize macro variables with %LET statements and dynamic
variables with SGRENDER’s DYNAMIC statement.

proc template;
define statgraph dynamics;
mvar SYSDATE9 SCALE;
nmvar BINS;
dynamic VAR VARLABEL;
begingraph;
entrytitle "Histogram of " VAR;
entrytitle "with Normal Distribution";
layout overlay / xaxisopts=(label=VARLABEL);
histogram VAR / scale=SCALE nbins=BINS;
densityplot VAR / normal();
endlayout;
entryfootnote halign=right "Created: " SYSDATE9 /
textattrs=GraphValueText;
endgraph;
end;
run;

%let bins=6;
%let scale=count;
proc sgrender data=sashelp.class
 16 Chapter 1 / Overview

template=dynamics;
dynamic var="Height" varlabel="Height in Inches";
run;

For more information about using dynamic variables and macro variables, see
Chapter 20, “Dynamic Variables and Macro Variables,” on page 1535.

Conditional Logic
Using conditional logic, you can create templates that have multiple visual results
or output representations, depending on existing conditions. The evaluation of a
logical expression must generate one or more complete statements (not portions of
statements). All conditional logic uses one of the following constructs:

IF Block IF/ELSE Block

if (condition) if (condition)
statement(s); statement(s);
endif; else
statement(s);
endif;

In the IF statement, condition must be enclosed in parentheses. The condition can

be any standard SAS expression involving arithmetic, logical operators, comparison
operators, Boolean operators, or concatenation operators. The expression can also
use SAS DATA step functions. The expression resolves to a single numeric value,
which is true or false.

In the following example, a histogram is conditionally overlaid with a normal

distribution curve, a Kernel Density Estimate distribution curve, both, or neither:
proc template;
define statgraph conditional;
dynamic VAR VARLABEL BINS CURVE;
begingraph;
entrytitle "Histogram of " VAR;
layout overlay / xaxisopts=(label=VARLABEL);
 Output 17

histogram VAR / nbins=BINS;

if (upcase(CURVE) in ("ALL" "KERNEL"))

densityplot VAR / kernel() name="k"
legendlabel="Kernel"
lineattrs=(pattern=dash);
endif;

if (upcase(CURVE) in ("ALL" "NORMAL"))

densityplot VAR / normal() name="n"
legendlabel="Normal";
endif;

discretelegend "n" "k";

endlayout;
endgraph;
end;
run;

Note that the legend syntax does not have to be made conditional. At run time,
each plot name in the legend is checked. If the plot does not exist, then its name is
removed from the legend name list. If no names appear in the DISCRETELEGEND
statement, then the legend “drops out” and the histogram size is adjusted to fill the
remaining space.

For more information about using conditional logic, see Chapter 23, “Conditional
Logic,” on page 1555.

Output
When using GTL, you focus primarily on defining template definitions that produce
specific graphs and generate a particular output layout. Ultimately, you must also
customize the graphical environment to get the exact output that you desire. The
ODS GRAPHICS statement is available for customizing the graphical environment,
and ODS styles enable you to manage the output appearance.

ODS GRAPHICS Statement

The ODS GRAPHICS statement is used to modify the environment in which
graphics templates are executed. The ODS GRAPHICS statement is used to control
n whether ODS Graphics is enabled

n the type and name of the image created

n the size of the image

n whether features such as scaling and anti-aliasing are used.

The following ODS GRAPHICS statement uses the HEIGHT= and WIDTH= options
to set an aspect ratio for the output image.
 18 Chapter 1 / Overview

ods graphics on / height=175px width=200px;

proc sgrender data=sashelp.class
template=scatter;
run;
ods graphics off;

For more information about using the ODS GRAPHICS statement in GTL, see SAS
Graph Template Language: User’s Guide. For a more complete discussion of the ODS
GRAPHICS statement, see “ODS GRAPHICS Statement” on page 1561.

ODS Styles
When any graphics template is executed, there is always an ODS style in effect
that governs the appearance of the output. The following ODS statement sends
graphics output to the RTF output destination using the LISTING style:
ods rtf style=listing;

ods graphics on / height=175px width=200px border=off;

proc sgrender data=sashelp.class
template=scatter;
run;
ods graphics off;

ods rtf close;

Support for ODS styles is highly integrated into GTL syntax. By default, the
graphical appearance features of most plot and text statements are mapped to
corresponding style elements and associated attributes. Because of this, your
output tables and graphs always have a reasonable overall appearance. Moreover,
output for a given ODS destination has a consistent look (for example, table colors
and graph colors do not clash).

The following figures show how a graph’s appearance can be changed by using
references to style elements to set the graph’s appearance options. This technique
permits changes in graph appearance by style modification instead of graphical
template modification. The graphs in the figures are generated with the following
GTL statement:
contourplotparm x=x y=y z=density /
contourtype=fill nhint=9
colormodel=ThreeColorRamp ;

The following style template shows the definition for the ThreeColorRamp style
element:
style ThreeColorRamp /
endcolor = GraphColors("gramp3cend")
neutralcolor = GraphColors("gramp3cneutral")
startcolor = GraphColors("gramp3cstart");
 About the Examples in This Documentation 19

For more information about the use of ODS styles in GTL, see SAS Graph Template
Language: User’s Guide. For a more complete discussion of ODS styles, see SAS
Output Delivery System: User’s Guide.

About the Examples in This

Documentation
The example programs that are shown in this document often provide all of the
code that you need to generate the graphs that are shown in the figures. We
encourage you to copy and paste the example code into your SAS session and
generate the graphs for yourself. The examples are written to be runnable in the
SAS windowing environment and in SAS Studio. Unless otherwise noted, the
examples use the default ODS destination. In the SAS windowing environment, the
default ODS destination is ODS HTML. For information about results in SAS Studio,
see SAS Studio: User’s Guide for your version of SAS Studio. You can find the
documentation for all versions of SAS Studio on the SAS Studio documentation
page on support.sas.com.
 20 Chapter 1 / Overview

If you generate the example graphs using an HTML destination, they are typically
rendered as 640 pixel by 480 pixel images using the HTMLBlue style. Because of
size limitations, the graphs in this document are not shown in their default size.
They are scaled down to meet the size requirements of our documentation
production system. When graphs that are produced with ODS graphics are reduced
in size, several automatic processes take place to optimize the appearance of the
output. Among the differences between default size graphs and smaller graphs are
that the smaller graphs have scaled down font sizes. Also, their numeric axes might
display a reduced number of ticks and tick values. Thus, the graphs that you
generate from the example programs will not always look identical to the graphs
that are shown in the figures. However, both graphs will accurately represent the
data.

When you produce your own graphical output, you can change the graph size and
attributes, if needed. The SAS Graph Template Language: User’s Guide explains how
to set fonts, DPI, anti-aliasing, and other features that contribute to producing
professional-looking graphics of any size in any output format.

Examples and Resources on the Web

The SAS website contains a large number of examples that can help you visualize
and code your graphs. The examples cover a range of SAS technologies including
the ODS Graphics procedures.
n Graphically Speaking is a blog focused on using ODS Graphics for data
visualization in SAS. The blog covers topics related to the ODS Graphics
procedures, the SAS Graph Template Language, and the SAS ODS Graphics
Designer.
https://blogs.sas.com/content/graphicallyspeaking/
n The SAS Training Post is a blog that provides tutorials, tips, and practical
information about SAS. Dr. Robert Allison contributed a great deal of
SAS/GRAPH and ODS Graphics data visualization examples to this blog during
his tenure at SAS. His blog is no longer active, but it provides an extensive set of
very useful examples.
https://blogs.sas.com/content/sastraining/author/robertallison/
n The SAS Knowledge Base contains an abundance of searchable samples and
SAS Notes. You can browse by topic, search for a particular note or a particular
technology such as the name of a procedure, and conduct other searches.
https://support.sas.com/en/knowledge-base.html

Note: The SAS Knowledge Base content is currently available only in English.

n The Graphics Samples Output Gallery in the SAS Knowledge Base is a collection
of graphs organized by SAS procedure. The graphs link to the source code. The
gallery is maintained by SAS Technical Support.
 Examples and Resources on the Web 21

https://support.sas.com/en/knowledge-base/graph-samples-gallery.html
n The Focus Area Graphics site provides a simple interface to business and
analytical graphs. The site is maintained by the SAS Data Visualization team.
https://support.sas.com/rnd/datavisualization/index.htm
n You can share your questions, suggestions, and experiences related to graphics
on the Graphics Programming community site.
https://communities.sas.com/t5/Graphics-Programming/bd-p/sas_graph.

In addition, SAS offers instructor-led training and self-paced e-learning courses to

help you get started with platform graphics software. The SAS Training site
provides information about the courses that are available.

sas.com/training.
22 Chapter 1 / Overview
 23

PART 2

Graph Block

Chapter 2
BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
24
 25

2
BEGINGRAPH Statement

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
BEGINGRAPH Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Dictionary

BEGINGRAPH Statement
Defines the outermost container for a graph template that is defined with GTL-statements.

Requirements: All STATGRAPH template definitions must start with a BEGINGRAPH statement
and end with an ENDGRAPH statement.
The BEGINGRAPH block must contain one and only one layout block.
The layout block and its nested layouts, if any, must contain at least one plot.

Syntax
BEGINGRAPH </options>;
<GTL-global-statements>
GTL-layout-block
<GTL-global-statements>
ENDGRAPH;
26 Chapter 2 / BEGINGRAPH Statement

Summary of Optional Arguments

Appearance options
ATTRPRIORITY=AUTO | COLOR | NONE
specifies a priority for cycling the group attributes.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the graph background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the graph.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the graph.
DATACOLORS=(color-list)
specifies the list of fill colors that will replace the graph data colors from
the GraphData1–GraphDataN style elements.
DATACONTRASTCOLORS=(color-list)
specifies the list of contrast colors that will replace the graph data
contrast colors from the GraphData1–GraphDataN style elements.
DATAFILLPATTERNS=(fill-pattern-list)
specifies the list of fill patterns that will replace the graph data fill
patterns from the GraphData1–GraphDataN style elements, if specified.
DATALINEPATTERNS=(line-pattern-list)
specifies the list of line patterns that will replace the graph data line
patterns from the GraphData1–GraphDataN style elements.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of all plots in the template that support
data skins.
DATASYMBOLS=(marker-symbol-list)
specifies the list of marker symbols that will replace the graph data
marker symbols from the marker symbols that are defined in the
GraphData1–GraphDataN style elements.
DESIGNHEIGHT=DEFAULTDESIGNHEIGHT | dimension
specifies the design height of the graph.
DESIGNWIDTH=DEFAULTDESIGNWIDTH | dimension
specifies the design width of the graph.
DRAWSPACE= GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a global drawing space and drawing units for all of the draw
statements within this BEGINGRAPH block.
OPAQUE=TRUE | FALSE
specifies whether the graph background is opaque or transparent.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the graph border.
SUBPIXEL=AUTO | OFF | ON
specifies whether subpixel rendering is used for drawing smooth curved
lines or for spacing bars more precisely.

Axis options
 BEGINGRAPH Statement 27

AXISBREAKSYMBOL=BRACKET | NOTCH | SLANTEDLEFT | SLANTEDRIGHT |

SQUIGGLE | SPARK | Z
specifies a symbol to use on the axis lines to indicate a break in the axis.
AXISBREAKTYPE=FULL | AXIS
specifies whether the axis break is indicated in the full display or only on
the axis line.
AXISLINEEXTENT=FULL | DATA | number
specifies the extent of the axis line for all axes.
DISCRETEAXISOFFSETPAD=TRUE | FALSE
specifies whether additional padding is added to the minimum and
maximum axis offsets for discrete axes.

Label options
LABELPLACEMENT=AUTO | GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in
the graphs.
SAPLACEMENTOPTS=(placement-options)
specifies the options for the label-placement algorithm when
LABELPLACEMENT=SA.

Midpoint options
INCLUDEMISSINGDISCRETE=TRUE | FALSE
specifies whether missing values are displayed on a discrete axis.

Optional Arguments
ATTRPRIORITY=AUTO | COLOR | NONE
specifies a priority for cycling the group attributes.

AUTO
honors the current state of the attribute priority rotation pattern as specified
in the active style or in the ODS GRAPHICS statement.

COLOR
changes the current setting of attribute priority rotation pattern to the color-
priority pattern by cycling through the list of colors while holding the marker
symbol, line pattern, or fill pattern constant. When all of the colors are
exhausted, the MarkerSymbol, LineStyle, or FillPattern attribute increments
to the next element, and then the colors in the list are repeated. This pattern
repeats as needed.

NONE
changes the current setting of attribute priority rotation pattern to the
default pattern, which cycles progressively through the attribute lists.

Tip Use NONE if you want groups to be distinguished by color, marker, and
line changes for all ODS styles that use color.

Default The attribute priority pattern that is specified in the active style or
in the ODS GRAPHICS statement.
28 Chapter 2 / BEGINGRAPH Statement

Interactions This option overrides the attribute priority rotation pattern that is
specified in the current style and the ATTRPRIORITY= option in the
ODS GRAPHICS statement.

The default lists of data colors, contrast colors, marker symbols,

and line patterns are set in the active style’s GraphData1–
GraphDataN elements.

The individual attributes in these lists can be overridden with the

BEGINGRAPH options DATACOLORS=,
DATACONTRASTCOLORS=, DATALINEPATTERNS=, and
DATASYMBOLS=.

The ATTRPRIORITY= option affects the cycling of the style

attributes for GROUP=, CYCLEATTRS=TRUE, and explicit style
references such as MARKERATTRS=GraphData2.

See “Attribute Rotation Patterns” in SAS Graph Template Language:

User’s Guide

AXISBREAKSYMBOL=BRACKET | NOTCH | SLANTEDLEFT | SLANTEDRIGHT |

SQUIGGLE | SPARK | Z
specifies a symbol to use on the axis lines to indicate a break in the axis.

Note: This feature applies to SAS 9.4M3 and to later releases.

The following figure shows an example of each symbol on a horizontal linear

axis for ranges 1–4 and 6–10.

Default SQUIGGLE

Restriction This option applies to linear and time axes only.

Requirements The AXISBREAKTYPE= option must be set to AXIS for this option
to have any effect.

You must use the INCLUDERANGES= option to specify ranges for

the axis for this option to have any effect.

The DISPLAY= option for the axis must include the axis line for
this option to have any effect.

AXISBREAKTYPE=FULL | AXIS
specifies whether the axis break is indicated in the full display or only on the
axis line.
 BEGINGRAPH Statement 29

Note: This feature applies to SAS 9.4M3 and to later releases.

The following figure shows an example of each type for ranges 50–52 and 56–73
on a linear horizontal axis.

Figure 2.1 Axis Break Types FULL and AXIS

Default FULL

Requirements You must use the INCLUDERANGES= axis option to specify

ranges for the axis for this option to have any effect.

The axis line or the plot wall outline must be displayed for AXIS
to have any effect. Otherwise, FULL is used.

Interaction When AXIS is specified, if the secondary axis line or the plot wall
outline is displayed, then the axis break symbol is displayed on
both the primary and the secondary axis. Otherwise, the break
symbol is displayed only on the primary axis, as shown in Figure
2.2 on page 29.

Notes The axis break indicators pass through inner margin areas.

No attempt is made to avoid collisions between the axis break

indicators and other graphical elements.

Tip When you use AXIS, use the AXISBREAKSYMBOL= option to

change the break symbol.

AXISLINEEXTENT=FULL | DATA | number

specifies the extent of the axis line for all axes.

Note: This feature applies to SAS 9.4M3 and to later releases.

FULL
specifies axis lines that extend along the entire length of the axis.

DATA
specifies axis lines that extend through the data range from the minimum
offset to the maximum offset.
30 Chapter 2 / BEGINGRAPH Statement

number
specifies, as a decimal proportion, how much the axis line extends from
DATA toward FULL. A value of 0 is equivalent to DATA, and a value of 1 is
equivalent to FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the
axis line at the midpoint positions of the minimum and maximum
bars. In that case, you can specify a numeric value to lengthen the
axis line so that it extends to the full width of both bars.

The following figure shows a simple example of each value for the X axis and Y-
axis lines. The light-blue dashed lines depict the minimum and maximum offsets
that are set on the axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Tips The graph wall outline might appear to be an axis line. In that case,
use the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in
the layout statement to suppress the wall outline.

Use the LINEEXTENT= axis option to control the axis line extent on
a per-axis basis.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the graph background.

style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style attribute named Color or ContrastColor is used.

Default The GraphBackground:Color style reference.

Interaction This option has no effect when OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the graph.

Default The ODS GRAPHICS statement BORDER= option setting, which is

TRUE by default.
 BEGINGRAPH Statement 31

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the graph.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element .

“Line Options” on page 1602 for available line-options.

DATACOLORS=(color-list)
specifies the list of fill colors that will replace the graph data colors from the
GraphData1–GraphDataN style elements.

(color-list)
a space-separated list of colors, enclosed in parentheses. You can use a style
attribute reference such as GraphData3:color, a color name, or an RGB,
CMYK, HLS, or HSV (HSB) color code to specify a color. The list can contain
a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

Example datacolors=(CXFF0000 green blue)

When this option is specified, the fill colors rotate through this color list rather
than through the colors that are defined in the GraphData1–GraphDataN style
elements. For information about the attribute rotation patterns, see “Attribute
Rotation Patterns” in SAS Graph Template Language: User’s Guide.

Default The colors that are defined in the GraphData1–GraphDataN style

elements.

Interaction Where applicable, the COLOR= suboption of the FILLATTRS= option

overrides the DATACOLORS= option.

Tip When specifying data attributes for multiple plots, the attributes
might not be assigned consistently in some cases. To ensure
consistent attribute assignments, specify option
CYCLEATTRS=FALSE in the layout statement.

DATACONTRASTCOLORS=(color-list)
specifies the list of contrast colors that will replace the graph data contrast
colors from the GraphData1–GraphDataN style elements.

(color-list)
a space-separated list of contrast colors, enclosed in parentheses. You can
use a style attribute reference such as GraphData3:color, a color name, or an
32 Chapter 2 / BEGINGRAPH Statement

RGB, CMYK, HLS, or HSV (HSB) color code to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

Example datacontrastcolors=(red graphdata3:contrastcolor CXFF0000)

When this option is specified, the contrast colors cycle through this color list
rather than through the contrast colors that are defined in the GraphData1–
GraphDataN style elements. For information about the attribute rotation
patterns, see “Attribute Rotation Patterns” in SAS Graph Template Language:
User’s Guide.

Default The contrast colors that are defined in the GraphData1–GraphDataN

style elements.

Interaction Where applicable, the COLOR= suboption of the MARKERATTRS=

option or the LINEATTRS= option overrides the
DATACONTRASTCOLORS= option.

Tip When specifying data attributes for multiple plots, the attributes
might not be assigned consistently in some cases. To ensure
consistent attribute assignments, specify option
CYCLEATTRS=FALSE in the layout statement.

DATAFILLPATTERNS=(fill-pattern-list)
specifies the list of fill patterns that will replace the graph data fill patterns
from the GraphData1–GraphDataN style elements, if specified.

Note: This feature applies to SAS 9.4M5 and to later releases.

(fill-pattern-list)
a space-separated list of fill patterns, enclosed in parentheses. You can use a
style attribute reference such as GraphData3:fillpattern or a two-character
fill-pattern name such as L1, X5, and so on, to specify a fill pattern. The list
can contain a mix of style attribute references and pattern names.

Requirement The list of fill patterns must be enclosed in parentheses.

Example datafillpatterns=(graphdata4:fillpattern L5 X2)

When this option is specified, the fill patterns rotate through this fill-pattern list
rather than through the fill patterns that are defined in the GraphData1–
GraphDataN style elements, if any. For information about the attribute rotation
patterns, see “Attribute Rotation Patterns” in SAS Graph Template Language:
User’s Guide.

Default The fill patterns that are defined in the GraphData1–GraphDataN

style elements, if any.

Interaction Where applicable, the PATTERN= suboption of the

FILLPATTERNATTRS= option overrides the DATAFILLPATTERNS=
option.
 BEGINGRAPH Statement 33

Tip When specifying data attributes for multiple plots, the attributes
might not be assigned consistently in some cases. To ensure
consistent attribute assignments, specify option
CYCLEATTRS=FALSE in the layout statement.

DATALINEPATTERNS=(line-pattern-list)
specifies the list of line patterns that will replace the graph data line patterns
from the GraphData1–GraphDataN style elements.

(line-pattern-list)
a space-separated list of line patterns, enclosed in parentheses. You can use
a style attribute reference such as GraphData3:LineStyle, a line pattern
number, or a line pattern name (where applicable) to specify a pattern. The
list can contain a mix of style attribute references, line pattern numbers, and
line pattern names.

Requirement The list of line patterns must be enclosed in parentheses.

When this option is specified, the line patterns cycle through this line-pattern
list rather than through the line patterns that are defined in the GraphData1–
GraphDataN style elements. When the patterns in line-pattern-list are
exhausted, the patterns repeat.

Default The line patterns that are defined in the GraphData1–GraphDataN

style elements.

Interaction Where applicable, the PATTERN= suboption of the LINEATTRS=

option overrides the DATALINEPATTERNS= option.

Tip When specifying data attributes for multiple plots, the attributes
might not be assigned consistently in some cases. To ensure
consistent attribute assignments, specify option
CYCLEATTRS=FALSE in the layout statement.

Example datalinepatterns=(solid dash 16 26)

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of all plots in the template that support data
skins. The following plot statements support data skins:

BARCHART HISTOGRAM SCATTERPLOT

BARCHARTPARM HISTOGRAMPARM SCATTERPLOTMATRIX
BOXPLOT LINECHART SERIESPLOT
BOXPLOTPARM NEEDLEPLOT STEPPLOT
BUBBLEPLOT PIECHART VECTORPLOT
DROPLINE POLYGONPLOT WATERFALLCHART
HIGHLOWPLOT REFERENCELINE

Default The GraphSkins:DataSkin style attribute, if it is specified in the

current style. If the current style does not specify the
GraphSkins:DataSkin style attribute, then the default is NONE.
34 Chapter 2 / BEGINGRAPH Statement

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data skin
is not applied to that plot. In that case, use the DATASKINMAX=
option in your ODS GRAPHICS statement to increase the maximum
limit.

Interaction This option is overridden by the DATASKIN= option in the individual

plot statements.

Note Applying data skins to a graph that has a very large number of
markers can negatively impact performance.

DATASYMBOLS=(marker-symbol-list)
specifies the list of marker symbols that will replace the graph data marker
symbols from the marker symbols that are defined in the GraphData1–
GraphDataN style elements.

(marker-symbol-list)
a space-separated list of marker symbols, enclosed in parentheses. You can
use a style attribute reference such as GraphData5:markerSymbol or a
marker symbol name to specify a marker. The list can contain a mix of style
attribute references and marker symbol names.

Requirement The list of marker symbols must be enclosed in parentheses.

When this option is specified, the marker symbols cycle through this marker
symbol list rather than through the line patterns that are defined in the
GraphData1–GraphDataN style elements. When the symbols in marker-symbol-
list are exhausted, the symbols repeat.

Default The marker symbols that are defined in the GraphData1–

GraphDataN style elements.

Interaction Where applicable, the SYMBOL= suboption of the MARKERATTRS=

option overrides the DATASYMBOLS= option.

Example datasymbols=(circle square triangle star)

DESIGNHEIGHT=DEFAULTDESIGNHEIGHT | dimension
specifies the design height of the graph.

Default DEFAULTDESIGNHEIGHT. This value is obtained from the SAS

Registry key ODS ð ODS Graphics ð Design Height when the
graph is rendered. The initial value of this registry key is 480px.

Restriction The minimum dimension value that you can set is 2 pixels. If a
smaller setting is specified, then the default design height is used.

Interactions The design height can be overridden at run time with a render
height that is specified with the HEIGHT= option in the ODS
GRAPHICS statement (external to the template). Also, the ODS
destination statement’s IMAGE_DPI= option can affect the height.
 BEGINGRAPH Statement 35

You can change the value of the Design Height registry key in the
SAS registry. However, doing so affects the design height of all
templates that do not include an explicit dimension for the design
height. You can also change the height setting in the graph style,
but doing so affects the height of all templates that use that style.

See “dimension” on page xiii

DESIGNWIDTH=DEFAULTDESIGNWIDTH | dimension
specifies the design width of the graph.

Default DEFAULTDESIGNWIDTH. This value is obtained from the SAS

Registry key ODS ð ODS Graphics ð Design Width when the graph
is rendered. The initial value of this registry key is 640px.

Restriction The minimum dimension value that you can set is 2 pixels. If a
smaller setting is specified, then the default design width is used.

Interactions The design width can be overridden at run time with a render width
that is specified with the WIDTH= option in the ODS GRAPHICS
statement (external to the template). Also, the ODS destination
statement’s IMAGE_DPI= option can affect the width.

You can change the value of the Design Width registry key in the
SAS registry. However, doing so affects the design width of all
templates that do not include an explicit dimension for the design
width. You can also change the width setting in the graph style, but
doing so affects the width of all templates that use that style.

See “dimension” on page xiii

DISCRETEAXISOFFSETPAD=TRUE | FALSE
specifies whether additional padding is added to the minimum and maximum
axis offsets for discrete axes. When set to TRUE, an additional 5 pixels of
padding are added to the minimum and maximum axis offsets.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default TRUE

Restriction This option applies only to discrete axes.

Tip This option is useful for heat maps when you want the heat map to
occupy the entire plot area. In that case, in addition to setting this
option to FALSE, set OFFSETMIN= and OFFSETMAX= to 0 for the
discrete axes.

DRAWSPACE= GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a global drawing space and drawing units for all of the draw
statements within this BEGINGRAPH block.
36 Chapter 2 / BEGINGRAPH Statement

Default LAYOUTPERCENT

Tip Individual draw statements within this BEGINGRAPH block can

override this global setting.

See “About the Drawing Space and Drawing Units” on page 1398

INCLUDEMISSINGDISCRETE=TRUE | FALSE
specifies whether missing values are displayed on a discrete axis.

Default FALSE

Interaction This option affects all charts and plots within the template.

See “boolean ” on page xii for other Boolean values that you can use.

LABELPLACEMENT=AUTO | GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in the
graphs. The following labels are affected:
n data labels for needle plots, scatter plots, series plots, step plots, and vector
plots
n vertex labels for line charts

n curve labels when the curve label is positioned at the start or end of the
curve

AUTO
always selects GREEDY.

GREEDY
specifies the Greedy method for managing label collision. The Greedy
method tries different placement combinations in order to find an optimal
approximation that avoids collisions. Label placement using this method is
often less optimal than label placement using the Simulated Annealing (SA)
method. However, depending on the number of data points and the potential
for label collisions, the Greedy process can be significantly faster.

SA
specifies the Simulated Annealing method for managing label collision. The
SA method attempts to determine the global minimization-of-cost function,
which is based on a simulated annealing algorithm. The resulting label
placement is usually better than placement using the Greedy method.
However, depending on the number of data points and the potential for label
collisions, the SA method can be significantly slower.

Restriction For BANDPLOT and LINECHART, the SA method has no effect

on the curve labels when the CURVELABELPOSITION= option
specifies START or END.

Default The value specified by the ODS GRAPHICS statement

LABELPLACEMENT= option, which is AUTO by default.

Restriction The data label placement algorithm is not aware of bar labels,
curve labels, box plot outlier labels, and marker characters.
 BEGINGRAPH Statement 37

Collisions between these elements and data labels might occur

regardless of the LABELPLACEMENT= setting.

Interactions This option overrides the ODS GRAPHICS statement

LABELPLACEMENT= option.

This option affects a plot’s labels only when

DATALABELPOSITION=AUTO is in effect for that plot.

The data label font size might be reduced in order to avoid

overlapping labels and markers. Starting with SAS 9.4M3, when a
broken axis is used, the data-label font size is not scaled during
label placement.

OPAQUE=TRUE | FALSE
specifies whether the graph background is opaque or transparent.

Note: This feature applies to SAS 9.4M3 and to later releases.

When this option is set to FALSE, the graph background is transparent.

Default TRUE

Restriction A transparent background is supported only by the PNG, EMF, PDF,

and SVG output formats. The PS output format supports a
transparent background when the graph is rendered as a PNG image.
It does not support a transparent background when the graph is
rendered as vector-graphics output.

Interaction When this option is set to FALSE, the BACKGROUNDCOLOR=

option has no effect.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the graph border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more of the following name-value-pair
options enclosed in parentheses:

LEFT=dimension specifies the amount of extra space added to the

left side.
RIGHT=dimension specifies the amount of extra space added to the
right side.
TOP=dimension specifies the amount of extra space added to the
top.
BOTTOM=dimension specifies the amount of extra space added to the
bottom.
38 Chapter 2 / BEGINGRAPH Statement

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 10 pixels.

Note The default units for dimension are pixels.

See “dimension” on page xiii

SAPLACEMENTOPTS=(placement-options)
specifies the options for the label-placement algorithm when
LABELPLACEMENT=SA. Placement options can be any of the following:

MAXITERATIONS=positive-integer
specifies the maximum number of iterations for the SA label-placement
algorithm.

Default 100

WEIGHTS=(keyword-number-list)
specifies the relative weight to give to a particular cost when determining
the best label position. The keyword number list is a space-separated list of
keyword = number pairs.

The following keywords can be used:

LABEL assigns a weight to the overlapping of labels

MARKER assigns a weight to the overlapping of markers and
labels
OUTOFBOUND assigns a weight to labels that are out-of-bounds or
clipped
PRIORITY assigns a weight to the priority of each potential label
position
OBSTACLE assigns a weight to the overlapping of labels with drop
lines, needles, reference lines, series lines, step lines,
and vector lines

The higher the number, the more weight is assigned to the specified cost. For
example, if MARKER is given more weight than OBSTACLE, avoiding marker
collisions is given a higher priority than avoiding line collisions.

Default A weight of 1.0 is assigned to each keyword

Restriction A dynamic variable reference or a run-time macro variable

reference cannot be used to specify a keyword value.

Example saplacementopts=(maxiterations=100
weights=(LABEL=2.0 OBSTACLE=10.0))

SEED=positive-integer
specifies a random number seed for the Simulated Annealing algorithm.
 BEGINGRAPH Statement 39

Default 1234567

Range 0–2147483647 (2 31–1), where 0 specifies the current Java time as

the seed value

Restriction This option applies only when LABELPLACEMENT=SA.

SUBPIXEL=AUTO | OFF | ON
specifies whether subpixel rendering is used for drawing smooth curved lines or
for spacing bars more precisely.

Note: Starting with SAS 9.4M3, this option controls subpixel rendering only for
image output. For vector-graphics output, subpixel rendering is always enabled.

AUTO
the system determines whether to use subpixel rendering. In SAS 9.4M2 and
in earlier releases, the system uses the default rendering for the rendering
technology. Starting with SAS 9.4M3, for image output, if the SUBPIXEL=
option is explicitly set in an ODS GRAPHICS statement, the system honors
its setting. Otherwise, the system determines whether to use subpixel
rendering based on the following criteria:
n If a SCATTERPLOT or SCATTERPLOTMATRIX statement is used,
subpixel rendering is OFF for the graph.
n If neither a SCATTERPLOT nor a SCATTERPLOTMATRIX statement is
used, subpixel rendering is turned ON for the graph if one or more of the
following statements is also used:

BANDPLOT DENSITYPLOT LOESSPLOT

BARCHART HEATMAP PBSPLINEPLOT
BARCHARTPARM HEATMAPPARM POLYGONPLOT
BOXPLOT HIGHLOWPLOT REGRESSIONPLOT
BOXPLOTPARM HISTOGRAM SERIESPLOT
BUBBLEPLOT HISTOGRAMPARM WATERFALLCHART
CONTOURPLOTPARM LINECHART
n For all other cases, subpixel rendering is turned OFF.

OFF
never uses subpixel rendering.

Note OFF is valid starting with SAS 9.4M3.

ON
always uses subpixel rendering, when applicable, for image output when
rendering graphs.

Default AUTO

Restrictions In SAS 9.4M2 and in earlier releases, subpixel rendering can be

used only for the following statements: BANDPLOT, BARCHART,
BARCHARTPARM, DENSITYPLOT, LINECHART, LOESSPLOT,
40 Chapter 2 / BEGINGRAPH Statement

PBSPLINEPLOT, REGRESSIONPLOT, and SERIESPLOT. Starting

with SAS 9.4M3, subpixel rendering can be used for all plots and
charts.

Starting with SAS 9.4M3, this option is ignored for vector-graphics

output.

Requirement Anti-aliasing must be enabled for this option to have any effect.

Interaction Starting with SAS 9.4M3, this option overrides the SUBPIXEL=
option in the ODS GRAPHICS statement.

Tips If anti-aliasing is disabled, use the ANTIALIAS=ON option in the

ODS GRAPHICS statement to enable it.

Anti-aliasing is disabled automatically when the resources

required for anti-aliasing exceed a preset threshold. When anti-
aliasing is disabled for all or part of a graph, subpixel rendering is
disabled for the entire graph. A note is written to the SAS log that
provides information about how to use the ANTIALIASMAX=
option in an ODS GRAPHICS statement to re-enable anti-aliasing.

See “Using Subpixel Rendering” in SAS Graph Template Language: User’s

Guide

“ODS GRAPHICS Statement” on page 1561 for information about

the ANTIALIAS= and ANTIALIASMAX= options.

Details

About the BEGINGRAPH Statement

All template definitions in the Graphics Template Language must start with a
BEGINGRAPH statement and end with an ENDGRAPH statement. Within a
BEGINGRAPH block, one and only one GTL layout block is required. It can be a
LATTICE, GRIDDED, OVERLAY, OVERLAYEQUATED, OVERLAY3D, REGION,
DATALATTICE, or DATAPANEL layout block. This layout block and its nested
layouts, if any, must contain at least one plot statement. It can contain other nested
layout blocks.

The GTL global statements apply to the entire template and can include
ENTRYTITLE and ENTRYFOOTNOTE statements, attribute maps, draw
statements, conditional statements, and so on. Any of these global statements can
precede or follow the GTL layout block.

Changing the Size of Your Graph

By default, graphs are rendered at 640px by 480px (4:3 aspect ratio). To change
the output size for a single graph, use the DESIGNWIDTH= and DESIGNHEIGHT=
options in the BEGINGRAPH statement for that graph. For example, the template in
the “Example Program” on page 42 uses DESIGNHEIGHT= to change the graph
 BEGINGRAPH Statement 41

height to 320px. To prevent the graph width from automatically scaling to preserve
the 4:3 aspect ratio, it uses DESIGNWIDTH= to maintain the 640px width. In this
instance, the setting renders each graph cell as a 320px by 320px square. (The cells
are square in this case, but the resulting cell size depends on the graph definition
and would not be the same for all graphs.)

Note: To change the graph sizes for all templates in the current SAS session, you
can use the WIDTH= and HEIGHT= options in the ODS GRAPHICS statement. Size
settings in the ODS GRAPHICS statement override size settings in the
BEGINGRAPH statement and remain in effect unless they are changed in another
ODS GRAPHICS statement. You can also use WIDTH= and HEIGHT= settings in the
graph style to modify the graphs sizes across template definitions. Be aware,
however, that if you explicitly manage the graph output size, then the graph
elements might be scaled so that the size specification is honored.

The following template defines a square graph (equal height and width, 1:1 aspect
ratio) by setting the design width equal to the internal default height (480px). The
setting is made with DESIGNWIDTH=DEFAULTDESIGNHEIGHT:

Note: A “square graph” means that the output graph’s width and height are equal.
That does not imply that the X axis and Y-axis lengths are equal if the graph
contains only one cell.

proc template;
define statgraph squareplot;
dynamic title xvar yvar;
begingraph / designwidth=defaultDesignHeight;
entrytitle title;
layout overlayequated / equatetype=square;
scatterplot x=xvar y=yvar;
regressionplot x=xvar y=yvar;
endlayout;
endgraph;
end;
run;

If this template is executed with the following SGRENDER procedure statement,

then a 480px by 480px graph is created:
proc sgrender data=mydata template="squareplot" ;
dynamic title="Square Plot" xvar="time1" yvar="time2";
run;

If the ODS GRAPHICS statement’s WIDTH= or HEIGHT= options change the render
width or render height, then the squareplot template’s 1:1 aspect ratio would still be
honored. Thus, both of the following GRENDER procedure statements would create
a 550px by 550px graph:
ods graphics / width=550px;
proc sgrender data=mydata template="squareplot" ;
dynamic title="Square Plot" xvar="time1" yvar="time2";
run;

ods graphics / height=550px;

 42 Chapter 2 / BEGINGRAPH Statement

proc sgrender data=mydata template="squareplot" ;

dynamic title="Square Plot" xvar="time1" yvar="time2";
run;

Example: BEGINGRAPH Statement

The following graph was generated by the “Example Program” on page 42:

Example Program
The BEGINGRAPH statement block is a required outermost container for any graph
template. One of its purposes is to support options that apply to the entire graph.
For example, the default graph size that a template produces is typically 640x480
pixels. If you need a different size, then you can declare the alternative size on this
statement. To do so, use the DESIGNWIDTH= option, or the DESIGNHEIGHT=
option, or both. This program shows one way to set the width and height of two
graph cells to be equal.
proc template;
define statgraph begingraph;
dynamic XVAR YVAR;
begingraph / designwidth=640px designheight=320px;
layout lattice / columns=2;
layout overlayequated / equatetype=square;
entry "Linear Regression Fit" /
valign=top textattrs=(weight=bold);
scatterplot x=XVAR y=YVAR / datatransparency=0.5;
regressionplot x=XVAR y=YVAR;
endlayout;
layout overlayequated / equatetype=square;
entry "Loess Fit" /
valign=top textattrs=(weight=bold);
scatterplot x=XVAR y=YVAR / datatransparency=0.5;
loessplot x=XVAR y=YVAR;
 BEGINGRAPH Statement 43

endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=begingraph;

dynamic title="Square Plot"
xvar="SepalLength" yvar="SepalWidth";
run;
44 Chapter 2 / BEGINGRAPH Statement
 45

PART 3

Layout Statements

Chapter 3
Summary of Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Chapter 4
Layout Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
46
 47

3
Summary of Layout Statements

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Single-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Multi-cell Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Data-driven Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Legend Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Overview
Layout blocks always begin with the LAYOUT keyword followed by a keyword
indicating the purpose of the layout. All layout blocks end with an ENDLAYOUT
statement.

The following sections summarize the available layouts. To learn more about a
layout, see the chapter devoted to that layout.

Single-cell Layouts
Graphics
Allowed and
Cells
Layout (Description) Produced Comments Example

OVERLAY (Single Cell) 2-D (1 cell) General purpose layout

for superimposing 2-D
plots
 48 Chapter 3 / Summary of Layout Statements

Graphics
Allowed and
Cells
Layout (Description) Produced Comments Example

OVERLAYEQUATED 2-D (1 cell) Specialized OVERLAY

(Single Cell) with equated axes

PROTOTYPE (Single 2-D (1 cell) Specialized LAYOUT

Cell) used only as child layout
of DATAPANEL or
DATALATTICE

REGION (Single Cell) 2-D (1 cell) General purpose layout

for displaying a single-
cell plot that does not
use axes

OVERLAY3D (Single 3-D (1 cell) General purpose 3-D

Cell) layout for superimposing
3-D plots.

Multi-cell Layouts
Graphics
Layout Allowed and
(Description) Cells Produced Comments Example

LATTICE 2-D (1 or more All cells must be predefined.

(Advanced Multi- cells) Axes can be shared across
cell) columns or rows and be
external to grid. Many grid
labeling and alignment
features.

GRIDDED 2-D (1 or more All cells must be predefined.

(Simple Multi- cells) Axes independent for each
cell) cell. Very simple multi-cell
container.
 Legend Layout 49

Data-driven Layouts
Graphics
Layout Allowed and
(Description) Cells Produced Comments Example

DATAPANEL 2-D (1 or more Displays a panel of similar

(Classification cells) graphs based on data subsets
Panel) by classification variable(s).
Number of cells is based on
crossings of n classification
variable(s).

DATALATTICE 2-D (1 or more Displays a panel of similar

(Classification cells) graphs based on data subsets
Panel) by classification variable(s).
Number of cells is based on
crossings of 1 or 2
classification variables.

Legend Layout
Layout Cells Produced Comments

GLOBALLEGEND 1 cell for a legend Specialized layout for creating a

compound legend that contains multiple
discrete legends.
50 Chapter 3 / Summary of Layout Statements
 51

4
Layout Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
LAYOUT DATALATTICE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
LAYOUT DATAPANEL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
LAYOUT GLOBALLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
LAYOUT GRIDDED Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
LAYOUT LATTICE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
LAYOUT OVERLAY Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
LAYOUT OVERLAYEQUATED Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
LAYOUT OVERLAY3D Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
LAYOUT PROTOTYPE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
LAYOUT REGION Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
INNERMARGIN Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Dictionary

LAYOUT DATALATTICE Statement

Creates a grid of graphs based on one or two classification variables and a graphical prototype. By
default, a separate instance of the prototype (a data cell) is created for each possible combination of
the classification variables.

Restriction: You can specify only one LAYOUT PROTOTYPE block in the LAYOUT
DATALATTICE block. If you specify more than one, then only the last prototype
block specified is honored. The remaining prototype blocks are ignored.
Requirement: You must specify at least one ROWVAR= option or one COLUMNVAR= option. You
can specify both.
52 Chapter 4 / Layout Statements

Syntax
LAYOUT DATALATTICE ROWVAR=class-variable
COLUMNVAR=class-variable < /options >;
LAYOUT PROTOTYPE </options>;
GTL-statements;
ENDLAYOUT;
<SIDEBAR </options>;
GTL-statements;
ENDSIDEBAR;>
ENDLAYOUT;

LAYOUT DATALATTICE COLUMNVAR=class-variable < /options >;

layout-prototype-block ;
<sidebar-block(s)> ;
ENDLAYOUT;

LAYOUT DATALATTICE ROWVAR=class-variable < /options >;

layout-prototype-block ;
<sidebar-block(s)> ;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid.
CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
COLUMNHEADERS=TOP | BOTTOM | BOTH
specifies where to position the outside column heading.
HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
HEADERLABELATTRS=style-element | style-element (text-options) | (text-
options)
 LAYOUT DATALATTICE Statement 53

specifies the color and font attributes of the data labels.

HEADERLABELLOCATION=OUTSIDE | INSIDE
indicates whether the cell header is placed within each cell (INSIDE) or
as row and column headers external to the lattice (OUTSIDE).
HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).
HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-
separated list in order to save space.
HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell
header when HEADERPACK=TRUE.
HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before
splitting the text to the next line.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.
SORTORDER=(COLUMNVAR=sort-order-keyword | ROWVAR=sort-order-
keyword)
specifies the order of the cells along the columns and rows.

Axis options
COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.
COLUMN2DATARANGE=AUTO | UNIONALL | UNION
specifies how the X2-axes of instances of the graph-prototype are scaled.
COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.
COLUMNDATARANGE=AUTO | UNIONALL | UNION
specifies how the X-axes of instances of the graph-prototype are scaled.
ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.
ROW2DATARANGE=AUTO | UNIONALL | UNION
specifies how the Y2-axes of instances of the graph-prototype are scaled.
ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.
ROWDATARANGE=AUTO | UNIONALL | UNION
specifies how the Y-axes of instances of the graph-prototype are scaled.
54 Chapter 4 / Layout Statements

Inset options
INSET=(variable-list)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Layout options
COLUMNS=integer
specifies the number of columns in the layout.
COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE
specifies the content of the cell headers.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the ROWVAR and
COLUMNVAR variables that contain a missing value.
PANELNUMBER=positive-integer
specifies the number of the panel to produce.
ROWGUTTER=dimension
specifies the amount of empty space between the rows.
ROWHEADERS=RIGHT | LEFT | BOTH
specifies where to position the outside row heading.
ROWS=integer
specifies the number of rows in the layout.
ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially
filled grid.
START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid from the top left or bottom
left corner.

Required Arguments
You must specify at least one of the following arguments. You can specify both.
ROWVAR=class-variable
specifies the classification variable for the rows. One row of cells is created for
each unique value of the row class variable.

See ROWS= option and “Managing Rows and Columns” on page 75

COLUMNVAR=class-variable
specifies the classification variable for the columns. One column is created of
each unique value of the column class variable.

See COLUMNS= option and “Managing Rows and Columns” on page 75

 LAYOUT DATALATTICE Statement 55

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid. Use this option in conjunction
with the CELLWIDTHMIN= option to set the minimum cell size.

The overall size of the panel is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the number of cells in the grid
increases, the size of each cell decreases. At some point the cell becomes so
small that a meaningful graph cannot be rendered. This option sets the
minimum height threshold for all cells. If the actual cell height becomes smaller,
then no panel is drawn.

Default 100px

See “dimension” on page xiii

56 Chapter 4 / Layout Statements

CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid. Use this option in conjunction
with the CELLHEIGHTMIN= option to set the minimum cell size.

The overall size of the panel is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the number of cells in the grid
increases, the size of each cell decreases. At some point the cell becomes so
small that a meaningful graph cannot be rendered. This option sets the
minimum width threshold for all cells. If the actual cell width becomes smaller,
then no panel is drawn.

Default 100px

See “dimension” on page xiii

COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.

Restriction Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options..

COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.

Restriction This option is needed only if you use a plot statement that
supports a secondary X2 axis. If you do not use that statement’s
XAXIS= option to map data to the X2 axis, then this option is
ignored. For more information about how data are mapped to the
axes, see “Plot Data Are Mapped to a Designated Axis” on page
1040

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options.

COLUMNDATARANGE=AUTO | UNIONALL | UNION

specifies how the X-axes of instances of the graph-prototype are scaled.

AUTO
selects the X-axis scale based on the COLUMNWEIGHT= option.

The X-axis scale is determined, as follows:

n When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.

n When COLUMNWEIGHT=PROPORTIONAL and the column axis is

discrete, UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the X-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
 LAYOUT DATALATTICE Statement 57

UNION
scales the X-axis data ranges separately for each column on a per-panel
basis.

The scaling does not span multiple panels.

Default AUTO

Tip Use the COLUMNAXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMN2DATARANGE=AUTO | UNIONALL | UNION

specifies how the X2-axes of instances of the graph-prototype are scaled.

AUTO
automatically selects the X2-axis scale based on the COLUMNWEIGHT=
option and the column axis type.

The X2-axis scale is selected as follows:

n When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.

n When COLUMNWEIGHT=PROPORTIONAL and the column axis is

discrete, UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the X2-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).

UNION
scales the X2-axis data ranges separately for each column on a per-panel
basis.

The scaling does not span multiple panels.

Default AUTO

Interaction This option is needed only if you use a plot statement that supports
a secondary X2 axis. If you do not use that statement’s XAXIS=
option to map data to the X2 axis, then this option is ignored. For
more information about how data are mapped to the axes, see “Plot
Data Are Mapped to a Designated Axis” on page 1040

Tip Use the COLUMN2AXISOPTS= option to control shared axis

features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
58 Chapter 4 / Layout Statements

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page xiii

COLUMNHEADERS=TOP | BOTTOM | BOTH

specifies where to position the outside column heading.

TOP
specifies that column heading text appears at the top of the layout.

BOTTOM
specifies that column heading text appears at the bottom of the layout.

BOTH
specifies that column heading text alternates between the top and bottom
of the layout column by column.

Default TOP

Interaction HEADERLABELLOCATION= OUTSIDE must be set for this

option to have any effect.

COLUMNS=integer
specifies the number of columns in the layout.

Defaults If this option is not specified, then the number of columns is

dynamically adjusted to equal the number of classifier values for
the COLUMVAR= variable.

If this option is specified, that many columns are created. If the

number of COLUMNVAR classifier values is greater than the
specified number of columns, then no graph is created for some
classifier values. If the number of classifier values is smaller than
the specified number of columns, then extra empty columns are
created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows,
the cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= option affects how the columns are populated.

The PANELNUMBER= option enables you to create multiple

smaller grids that completely partition the classifier values.

COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.

EQUAL
all columns have equal width.
 LAYOUT DATALATTICE Statement 59

PROPORTIONAL
each column has a width that is proportional to the number of discrete
midpoint values that it contains.

Restriction At least one column axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When COLUMNDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect,

COLUMNDATARANGE=AUTO is interpreted as UNION.

If all of the following conditions are true, then the discrete axis
is used to proportion the columns: PROPORTIONAL is in effect,
both the X and X2 axes are used, and only one of the two axes is
discrete. If both axes are discrete, then the X axis is used to
proportion the columns.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in COLUMNAXISOPTS=
and COLUMN2AXISOPTS=. The axis offsets are always set to
one-half of the midpoint spacing regardless of plot type.

Default EQUAL

HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphHeaderBackground:Color style reference.

Interaction HEADEROPAQUE= TRUE must be in effect in order for the color to

be seen.

HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default TRUE

Tip The border attributes are controlled by the GraphBorderLines style

element.

See “boolean ” for other Boolean values that you can use.

HEADERLABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the data labels.
60 Chapter 4 / Layout Statements

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE

specifies the content of the cell headers.

NAMEVALUE
displays the classification variable name and value as a name=value pair in
each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=NAMEVALUE produces cell headers such as the
following:

Country=CANADA
Product=TABLE

VALUE
displays the classification variable value only in each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=VALUE produces cell headers such as the
following:

CANADA
TABLE

NONE
suppresses the cell headers.

Default NAMEVALUE

HEADERLABELLOCATION=OUTSIDE | INSIDE
indicates whether the cell header is placed within each cell (INSIDE) or as row
and column headers external to the lattice (OUTSIDE).

Default OUTSIDE

HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).

Default TRUE

Interaction When this option is set to FALSE, the background color for cell
headers is not used.

See “boolean ” on page xii for other Boolean values that you can use.
 LAYOUT DATALATTICE Statement 61

HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated list
in order to save space.

Note: This option is valid starting in SAS 9.4M1.

The following figure shows the effect of HEADERPACK= on the cell headers in
one row of a data lattice. The data lattice contains classification variables
Country and Year.

The following figure shows the effect of HEADERPACK= on the cell headers in
one row of a data lattice. The data lattice contains classification variables
Country, Year, and Product.

Default FALSE

Note If the length of the cell header text exceeds the available width, then
the text is truncated.

Tip If truncation occurs, then use the HEADERSPLITCOUNT= option to

split the cell header text into multiple lines.

See “boolean ” on page xii for other Boolean values that you can use.

HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell header
when HEADERPACK=TRUE.

Note: This option is valid starting in SAS 9.4M1.

Default A comma followed by a space

Interaction This option is ignored when HEADERPACK=FALSE.

HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before splitting
the text to the next line.
62 Chapter 4 / Layout Statements

Note: This option is valid starting in SAS 9.4M1.

The following figure shows how HEADERSPLITCOUNT=2 splits the cell header
value in a data panel of classification variables Country, Year, and Product.

Default The cell header text is not split

Interaction This option is ignored when HEADERPACK=FALSE.

Note If the length of the cell header text exceeds the available width, then
the text is truncated.

Tip Use the HEADERSEPARATOR= option to specify a different

separator.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the ROWVAR and
COLUMNVAR variables that contain a missing value.

TRUE
any crossing of the class variables that includes a missing value produces a
row or column of cells in the grid.

FALSE
any crossing of the class variables that includes a missing value does not
produce a row or columns of cells in the grid.

By default, missing class values are included in the classification. When the data
contains missing classification values, cells are created for the missing classes.
The classification headers for the missing values are blank for missing string
values or a dot for missing numeric values. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing values. If you
want to keep the missing values, then you can create a format that specifies
more meaningful headings for the missing classes. For an example, see “Missing
Class Values” in SAS Graph Template Language: User’s Guide.

Note: ODS Graphics does not support Unicode values in user-defined formats
in SAS 9.4M2 and in earlier releases. Starting with SAS 9.4M3, ODS Graphics
supports Unicode values in user-defined formats only if they are preceded by
the (*ESC*) escape sequence. Example: "(*ESC*){unicode beta}". ODS
Graphics does not support an escape character that is defined in an ODS
ESCAPECHAR statement in user-defined formats.

Default TRUE

See “boolean ” on page xii for other Boolean values that you can use.

INSET=(variable-list)
specifies what information is displayed in an inset. The variable-list defines one
or more variables whose names and values appear as a small table in the data
cells. The variables can be either numeric or character. Variable names are
separated by spaces.
 LAYOUT DATALATTICE Statement 63

Requirement No predefined information is available for the inset. You must

create the desired inset information as part of your input data. See
“Creating Your Inset Data” on page 76.

Note The variable values are associated with the data cells by data
order. That is, the first observation from all the variables in
variable-list are used in the first data cell, the second observation
from all variables in variable-list are used in the second data cell,
and so on. If a value is missing for an observation, then the
corresponding name-value pair is skipped in the affected data cell.

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See “Adding Insets to Your Graph” in SAS Graph Template Language:

User’s Guide

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The
appearance options can be any one or more of the following values:

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the inset is automatically aligned within the layout.

NONE
does not automatically align the inset.

This inset’s position is set by the HALIGN= and VALIGN=appearance-

options.

AUTO
attempts to center the inset in the area that is farthest from any
surrounding markers.

Data cells might have different inset placements.

(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with
the data cell’s other graphics features.

The location-list is a space-separated list that can contain any of these

locations: TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT,
BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Example AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the HALIGN= and

VALIGN= options are ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background.

style-reference
specifies a style reference in the form style-element : style-attribute.
64 Chapter 4 / Layout Statements

Only the COLOR and CONTRASTCOLOR style attributes are valid.

Default The background is transparent. No color is assigned.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

CONTENTDISPLAY=LABELVALUE | VALUE
specifies whether the variable information that is displayed in the inset
includes the column label and value, or only the column value.

Note: This feature applies to SAS 9.4M2 and to later releases.

If a label is not assigned to a column, then the column name is used as the
label for that column. Consider the following inset data:
F
Obs _TYPE_ Value Pr > F
1 SS1 94.359 <.0001

The following figure shows the effect that the CONTENTDISPLAY= option
has on the content of an inset that displays this data.

Default LABELVALUE

Tip Use the SEPARATOR= option to specify a separator other than the
default blank space.

DATASCHEME=LIST | MATCHED
specifies the scheme that was used to merge the inset information into the
analysis data.

LIST
one-to-one merging (no BY statement) was used to merge the inset and
analysis data.

The variable values are associated with the cells of the data lattice by
using data order. That is, the inset variable values in the first observation
are used in the inset for the first cell, the inset variable values in the
second observation are used in the inset for the second cell, and so on.

MATCHED
match-merging (using a BY statement) was used to merge the inset and
analysis data.
 LAYOUT DATALATTICE Statement 65

Defaults LIST if the data is a SAS data set

MATCHED if the data is a CAS in-memory table

Tip MATCHED is the preferred data scheme for merging the inset and
analysis data.

See “Adding Insets to Classification Panels” in SAS Graph Template

Language: User’s Guide

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.

Default LEFT

Interaction This option has an effect only when this layout is nested within a
REGION layout or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When OPAQUE=FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

SEPARATOR="string"
specifies a new separator for the column label and value.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default A blank space

Interaction This option is ignored when CONTENTDISPLAY=VALUE.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset, excluding the title.

Default The GraphDataText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TITLE="string"
specifies a title for the inset.

The title is added at the top of the inset and spans the full inset width.

Note Space is not reserved for the title when this value is not specified.
66 Chapter 4 / Layout Statements

Tip Text properties for the title string can be specified with
TITLEATTRS=.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the inset.

Default TOP

Interaction This option has effect only when this layout is nested within a
REGION layout or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

Requirements The options must be enclosed in parentheses.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.
 LAYOUT DATALATTICE Statement 67

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

68 Chapter 4 / Layout Statements

Note The default units for dimension are pixels.

See “dimension” on page xiii

PANELNUMBER=positive-integer
specifies the number of the panel to produce. This option enables you to
partition a large grid into a number of smaller grids under these conditions:
n You set a grid size explicitly (ROWS= and COLUMNS= when ROWVAR and
COLUMNVAR variables are specified; ROWS= when only a ROWVAR
variable is specified; COLUMNS= when only a COLUMNVAR variable is
specified)
n The grid size is smaller in one or both of the dimensions of the default
dynamically generated grid.
n You execute the template N times and increment the panel number each
time. N is determined by CEIL(all rows * all columns / grid rows * grid
columns).

Default 1

Example Suppose ROWVAR=R (R has 10 unique values) and COLUMNVAR=C

(C has 11 unique values). The dynamic grid has 10 rows and 11 columns
and you would have to make the HEIGHT=and WIDTH= quite large to
enable 110 plots to be displayed. By setting some smaller grid size, say
ROWS=3 and COLUMNS=4, and by making the value of
PANELNUMBER= a dynamic or macro variable, you can create 10
panels (9 with 12 data cells and 1 with 2 data cells) that collectively
display all 110 possible crossings. You simply invoke PROC SGRENDER
or a DATA step 10 times, incrementing the dynamic value for
PANELNUMBER each time.

ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1218 for a list of options..

ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

Interaction This option is needed only if you use a plot statement that
supports a secondary Y2 axis. If you do not use that statement’s
YAXIS= option to map data to the Y2 axis, then this option is
ignored. For more information about how data are mapped to the
axes, see “Plot Data Are Mapped to a Designated Axis” on page
1040
 LAYOUT DATALATTICE Statement 69

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1218 for a list of options..

ROWDATARANGE=AUTO | UNIONALL | UNION

specifies how the Y-axes of instances of the graph-prototype are scaled.

AUTO
automatically selects the Y-axis scale based on the ROWWEIGHT= option
and the column axis type.

The Y-axis scale is selected as follows:

n When ROWWEIGHT=EQUAL (default), UNIONALL is selected.

n When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,

UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the Y-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).

UNION
scales the Y-axis data ranges separately for each row in the layout on a per-
panel basis.

The scaling does not span multiple panels.

Default AUTO

Tip Use the ROWAXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROW2DATARANGE=AUTO | UNIONALL | UNION

specifies how the Y2-axes of instances of the graph-prototype are scaled.

AUTO
automatically selects the Y2-axis scale based on the ROWWEIGHT= option
and the column axis type.

The Y2-axis scale is selected as follows:

n When ROWWEIGHT=EQUAL (default), UNIONALL is selected.

n When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,

UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the Y2-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).

UNION
scales the Y2-axis data ranges separately for each row in the layout on a
per-panel basis.

The scaling does not span multiple panels.

70 Chapter 4 / Layout Statements

Default AUTO

Interaction This option is needed only if you use a plot statement that supports
a secondary Y2 axis. If you do not use that statement’s YAXIS=
option to map data to the Y2 axis, then this option is ignored. For
more information about how data are mapped to the axes, see “Plot
Data Are Mapped to a Designated Axis” on page 1040

Tip Use the ROW2AXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page xiii

ROWHEADERS=RIGHT | LEFT | BOTH

specifies where to position the outside row heading.

RIGHT
specifies that row heading appears at the right of the layout.

LEFT
specifies that row heading appears at the left of the layout.

BOTH
specifies that row heading alternates between the right and left of the
layout row by row.

Default RIGHT

Requirement HEADERLABELLOCATION= OUTSIDE must be set for this option

to have any effect.

ROWS=integer
specifies the number of rows in the layout.

Defaults If this option is not specified, then the number of rows is

dynamically adjusted to equal the number of classifier values for
the ROWVAR= variable.

If this option is specified, then the specified number of rows is

created. If the number of ROWVAR classifier values is greater than
the specified number of rows, then no graph is created for some
classifier values. If the number of classifier values is smaller than
the specified number of rows, then extra empty rows are created.
 LAYOUT DATALATTICE Statement 71

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows,
the cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= option affects how the rows are populated.

Tip The PANELNUMBER= option enables you to create multiple

smaller grids that completely partition the classifier values.

ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.

EQUAL
all rows have equal height.

PROPORTIONAL
each row has a height that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one row axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When ROWDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect, ROWDATARANGE=AUTO

is interpreted as UNION.

If all of the following conditions are true, then the discrete axis
is used to proportion the rows: PROPORTIONAL is in effect,
both the Y and Y2 axes are used, and only one of the two axes is
discrete. When both axes are discrete, the Y axis is used to
proportion the rows.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in ROWAXISOPTS= and
ROW2AXISOPTS=. The axis offsets are always set to one-half
of the midpoint spacing regardless of plot type.

Default EQUAL

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Default FALSE

Note If SHRINKFONTS=TRUE, the font scaling applies to all child layouts

within the parent layout. If a child layout also specifies
SHRINKFONTS=TRUE, the fonts for the children of that child layout
are further scaled down.

See “boolean ” on page xii for other Boolean values that you can use.
72 Chapter 4 / Layout Statements

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled grid.

TRUE
skips empty cells and "snaps" the external axes to the nearest data cell, both
vertically and horizontally.

Though the empty cells are not displayed, the data cells in the grid are not
enlarged to fill the area.

FALSE
displays external axes at their normal locations, even if there are empty cells
at one or more of the locations.

Whenever the number of unique COLUMNVAR= classifier values (data cells) is

not evenly divisible by the COLUMNS= value, or the number of unique
ROWVAR= classifier values (data cells) is not evenly divisible by the ROWS=
value, then one or more panels is partially filled with data cells and padded with
empty cells to complete the grid.

Here is an example of a data lattice that consists of 4 column-data cells and 3

row-data cells arranged in a 4-column, 2-row grid. The following figure shows
the default appearance of the last panel:

When SKIPEMPTYCELLS=TRUE, the empty padding cells of all panels are

removed and external axis ticks and tick values snap to the data cells:

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.
 LAYOUT DATALATTICE Statement 73

SORTORDER=(COLUMNVAR=sort-order-keyword | ROWVAR=sort-order-
keyword)
specifies the order of the cells along the columns and rows.

COLUMNVAR
the column role.

ROWVAR
the row variable role.

sort-order-keyword
a sort-order keyword.

The keyword can be one of the following:

AUTO
sorts using DATA for character data and ascending unformatted for
numeric data.

DATA
retains the data order.

Note In SAS 9.4M6 and SAS 9.4M5, when data stored in a CAS in-
memory table is used and DATA is specified, unpredictable results
might occur. To generate consistent graphs in that case, specify
ASCENDINGFORMATTED or DESCENDINGFORMATTED. Starting
with SAS Viya 3.5 and SAS Studio 5.2, when data stored in a CAS
in-memory table is used and DATA is specified, DATA is ignored,
and AUTO is used instead.

ASCENDINGFORMATTED
sorts in ascending order, using the formatted values.

DESCENDINGFORMATTED
sorts in descending order, using the formatted values.

Defaults AUTO for all roles if the data is a SAS data set

ASCENDINGFORMATTED for all roles if the data is a CAS in-memory

table

Tip The placement of the cells within the layout also depends on the
starting location, which is controlled by the START= option.

START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid from the top left or bottom left
corner. If ROWVAR=R has values in the sort order 1 and 2 and COLUMNVAR=C
has values in the sort order a and b, then START=BOTTOMLEFT is populated as
follows:
74 Chapter 4 / Layout Statements

START=TOPLEFT is populated as follows:

Default TOPLEFT

SIDEBAR Statement Options

ALIGN=BOTTOM | TOP | LEFT | RIGHT
specifies the sidebar’s location within the layout. You can specify up to four
SIDEBAR blocks in a LAYOUT DATALATTICE, one for each of the bottom, top,
left, and right sidebar positions.
n The LAYOUT DATALATTICE automatically aligns a sidebar with the layout
columns or rows.
n Only one statement (such as ENTRY or DISCRETELEGEND) or one layout
block (such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create
multi-line text in a sidebar, nest ENTRY statements within a LAYOUT
GRIDDED block.

Default BOTTOM

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the
sidebar boundaries, set this option to FALSE.

See “boolean ” on page xii for other Boolean values that you can use.

Details

Statement Description
The LAYOUT DATALATTICE statement makes it easy to create a grid of graphs,
based on the values of one or two classifications variables. To create a grid that is
based on more than two classification variables, or to have more control over the
grid layout, use LAYOUT DATAPANEL instead.

By default, the number of cells in the layout is determined by the number of value
pairings that are possible for the classification values plus any empty cells needed
to complete the last row or column of the grid. The contents of each data cell are
based on a graph prototype that you specify in the graph-prototype-block. You can
enhance the display using one or more sidebar-statement-blocks. For classification
 LAYOUT DATALATTICE Statement 75

variables that have many values, you can use the COLUMNS= and or ROWS=
options and the PANELNUMBER= option to generate multiple panel displays.

Classification variables for the layout are specified on the ROWVAR= argument (to
specify a row variable), or the COLUMNVAR= argument (to specify a column), or
both arguments to specify both a column and a row variable. The graph prototype
for each data cell’s contents is specified within a “LAYOUT PROTOTYPE Block” on
page 77 block, and sidebars are specified within SIDEBAR blocks. The LAYOUT
PROTOTYPE and SIDEBAR blocks are nested within the LAYOUT DATALATTICE
block.

By default, the first data cell to be filled is in the layout’s top left corner. Use the
START= option to change the starting data cell to the bottom left corner.

Rather than display the header labels outside the grid, you can set
HEADERLABELLOCATION= INSIDE to display them inside the grid, as shown in the
following figure:

Note: The DATALATTICE layout is designed to be the outermost layout in the

template.

Managing Rows and Columns

If you do not explicitly manage columns and rows using the COLUMNS= and
ROWS= options, then the default layout behavior is as follows:
n If both ROWVAR= and COLUMNVAR= are specified, then a data cell is created
for each of the value pairings that are possible for the classification values of
the specified variables. If the ROWVAR variable has R distinct values and the
COLUMNVAR variable has C distinct values, then the dimension of grid
produced is R x C.
76 Chapter 4 / Layout Statements

n If only the ROWVAR variable is used, then an R x 1 grid is produced.

n If only the COLUMNVAR variable is used, then a 1 x C grid is produced.

If the class variable is of type character, then its values are returned in data order.
To control the ordering of the values, you can sort the input data by the
classification variables. If the class variable is of type numeric, then the values are
displayed in ordinal order.

Formats can be assigned to class variables to create classification levels (for

example, an AGEGROUPFMT. format for numeric AGE). In this case, the
classification is performed after the format is applied. For numeric data, the order is
ordinal, based on the first value in each class.

Use the INCLUDEMISSINGCLASS option to control whether cells are displayed

when any value crossing contains a missing value.

The output size does not grow automatically as the number of cells increases. To
set a panel size for the current template, use the DESIGNHEIGHT= and
DESIGNWIDTH= options in the BEGINGRAPH statement. To set a panel size for all
templates in the current SAS session, use the HEIGHT= and WIDTH= options in the
ODS GRAPHICS statement. Size settings in the ODS GRAPHICS statement
override size settings in the BEGINGRAPH statement. The default output width is
640px, and the default output height is 480px.

As the number of cells in the grid increases, the size of each cell decreases. At some
point the cells might become so small that a meaningful graph cannot be rendered.
The CELLHEIGHTMIN= and CELLWIDTHMIN= options set a threshold for the
smallest cell. If the actual cell height or width becomes smaller, then no panel is
drawn. The default minimum cell size is CELLHEIGHTMIN=100px and
CELLWIDTHMIN=100px.

Using the default panel size and cell size, the DATALATTICE layout accommodates
a grid of about 24 cells (6 columns by 4 rows). If you know that the number of cells
is larger, then you should increase the overall panel size, or decrease the minimum
cell size, or both. You can also use ROWS= , COLUMNS= , and PANELNUMBER=
options to partition your data so that a number of smaller grids are produced that
cumulatively show all of the value crossings.

Creating Your Inset Data

When you use the INSET= option to insert an inset, no predefined information is
available for the inset. You must create the desired inset information as part of your
input data. This is most typically done as follows:
n Create a separate data set for the inset columns making sure that the column
names are different from the other columns used in graph. The number
observations of inset data should match the number of cells in the classification
panel. The ordering of the inset observations should be the same as the
population order of the classification panel’s cells, taking into account the
ROWVAR= and COLUMNVAR= arguments and the START= option. Typically,
the number of observations for the inset data is smaller than the other input
data for the graph.
n Merge the inset data set with the data set for the graph using a DATA or PROC
SQL step. Do not match-merge the observations of the two data sets (no BY
 LAYOUT DATALATTICE Statement 77

processing). The resulting data set typically has the inset columns padded with
missing values.
n Use the merged data set to produce the graph, specifying the inset column
names in this option’s variable-list.

LAYOUT PROTOTYPE Block

You must specify a single LAYOUT PROTOTYPE block within the LAYOUT
DATALATTICE block, using the following syntax:

LAYOUT PROTOTYPE </options>;

GTL-statement(s);
ENDLAYOUT;

The LAYOUT PROTOTYPE block determines the graphical content of each data cell
and is repeated within each data cell, based on the subsets of the classification
variables.

For more information about the LAYOUT PROTOTYPE block and the list of
available options, see “LAYOUT PROTOTYPE Statement” on page 186.

SIDEBAR Blocks
A LAYOUT DATALATTICE enables you to display one or more sidebars outside of
the axes. A sidebar spans across columns or rows and is useful for displaying
information that applies to all of the columns or all of the rows. For example,
sidebars are useful for displaying a legend.

A SIDEBAR statement has the following syntax:

SIDEBAR </ options>;

GTL-statement(s);
ENDSIDEBAR;

The following example shows a SIDEBAR block that displays a legend at the top of
the layout grid.

sidebar / align=top;
discretelegend 'p' 'a' / across=2;
endsidebar;

Example: LAYOUT DATALATTICE

This example shows the result of using row and column classification variables. In
this case, a four-column, three-row data lattice is created:
n The classification values are placed as row or column labels by default.

n The ROWDATARANGE=UNION option assures that an axis range is computed

separately for each row using the data ranges of all Y= columns in that row. This
facilitates the visual comparison of the data cells.
 78 Chapter 4 / Layout Statements

n A SIDEBAR block is used to place the legend at the bottom of the lattice.

Example Graph
The following graph was generated by the “Example Program” on page 78:

Example Program
proc template;
define statgraph layoutdatalattice;
begingraph;
entrytitle "Annual Furniture Sales Comparisons";
layout datalattice rowvar=country columnvar=year /
rowdatarange=union
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(display=(tickvalues) griddisplay=on
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
 LAYOUT DATAPANEL Statement 79

sidebar / align=bottom;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

class country year month;
var actual predict;
output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatalattice;

run;

LAYOUT DATAPANEL Statement

Creates a grid of graphs based on one or more classification variables and a graphical prototype. By
default, a separate instance of the prototype (a data cell) is created for each actual combination of
the classification variables.

Restriction: You can specify only one LAYOUT PROTOTYPE block in the LAYOUT DATAPANEL
block. If you specify more than one, then only the last prototype block specified is
honored. The remaining prototype blocks are ignored.
Tip: The DATAPANEL layout should be the outermost layout in the template.

Syntax
LAYOUT DATAPANEL CLASSVARS=(class-var1…class-varN) < /options > ;
LAYOUT PROTOTYPE < /options >;
GTL-statements;
ENDLAYOUT;
<SIDEBAR < / options >;
GTL-statement(s);
ENDSIDEBAR;
<… more-sidebar-statement-blocks …> >
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
80 Chapter 4 / Layout Statements

specifies the color of the layout background.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid.
CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.
HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.
HEADERLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the data labels.
HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).
HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-
separated list in order to save space.
HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell
header when HEADERPACK=TRUE.
HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before
splitting the text to the next line.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.
SORTORDER=(COLUMNVAR=sort-order-keyword | ROWVAR=sort-order-
keyword)
specifies the order of the cells along the columns and rows.

Axis options
COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.
COLUMN2DATARANGE=AUTO | UNIONALL | UNION
 LAYOUT DATAPANEL Statement 81

specifies how the X2-axes of instances of the graph-prototype are scaled.

COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.
COLUMNDATARANGE=AUTO | UNIONALL | UNION
specifies how the X-axes of instances of the graph-prototype are scaled.
ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.
ROW2DATARANGE=AUTO | UNIONALL | UNION
specifies how the Y2-axes of instances of the graph-prototype are scaled.
ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.
ROWDATARANGE=AUTO | UNIONALL | UNION
specifies how the Y-axes of instances of the graph-prototype are scaled.

Inset options
INSET=(variable-list)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Layout options
COLUMNS=integer
specifies the number of columns in the layout.
COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.
HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE
specifies the content of the cell headers.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the CLASSVARS
variables that contain a missing value.
ROWGUTTER=dimension
specifies the amount of empty space between the rows.
ROWS=integer
specifies the number of rows in the layout.
ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row
priority.
PANELNUMBER=positive-integer
specifies the number of the panel to produce.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially
filled grid.
SPARSE=TRUE | FALSE
82 Chapter 4 / Layout Statements

specifies whether crossings of the class variables include only the

crossings in the data or all possible crossings.
START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid cells from the top left or
bottom left corner.

Role options
ROLENAME=(role-name-list)
specifies user-defined roles for information contained in data columns.

Required Argument
CLASSVARS=(column-list)
specifies a list of classification variables. By default, a data cell is created for
each crossing of these variables in the input data. The total number of grid cells
created is the result of a crosstabulation table of all the classification variables
plus any empty cells needed to complete the last row or column of the grid. You
can request that data cells be generated for all possible crossings, even when
the class variables have no values at those crossings. For more information, see
the SPARSE= option.

If the class variable is of type character, then its values are returned in data
order. To control the ordering of the values, you can sort the input data by the
classification variables. If the class variable is of type numeric, then the values
are displayed in ordinal order.

Formats can be assigned to class variables to create classification levels (for

example, an AGEGROUPFMT. format for numeric AGE). In this case, the
classification is performed after the format is applied. For numeric data, the
order is ordinal, based on the first value in each class.

Use the INCLUDEMISSINGCLASS option to control whether cells are displayed

when any value crossing contains a missing value.

The output size does not grow automatically as the number of cells increases.
To set a panel size for the current template, use the DESIGNHEIGHT= and
DESIGNWIDTH= options in the BEGINGRAPH statement. To set a panel size for
all templates in the current SAS session, use the HEIGHT= and WIDTH= options
in the ODS GRAPHICS statement. Size settings in the ODS GRAPHICS
statement override size settings in the BEGINGRAPH statement. The default
output width is 640px, and the default output height is 480px.

As the number of cells in the grid increases, the size of each cell decreases. At
some point the cells might become so small that a meaningful graph cannot be
rendered. The CELLHEIGHTMIN= and CELLWIDTHMIN= options set a threshold
for the smallest cell. If the actual cell height or width becomes smaller, then no
panel is drawn. The default minimum cell size is CELLHEIGHTMIN=100px and
CELLWIDTHMIN=100px.

Using the default panel size and cell size, the DATAPANEL layout
accommodates a grid of about 24 cells (6 columns by 4 rows). If you know that
the number of cells is larger, then you should increase the overall panel size, or
decrease the minimum cell size, or both. You can also use ROWS= ,
COLUMNS= , and PANELNUMBER= options to partition your data so that a
 LAYOUT DATAPANEL Statement 83

number of smaller grids are produced that cumulatively show all of the value
crossings.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

HEADERBORDER=TRUE | FALSE
specifies whether a border is drawn around the header cells.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default TRUE

Tips The border attributes are controlled by the GraphBorderLines style

element.
84 Chapter 4 / Layout Statements

HEADERBORDER=FALSE does not remove the border around empty

cells. If your panel has empty cells and you want to remove the cell
borders, you can specify SKIPEMPTYCELLS=TRUE with
HEADERBORDER=FALSE to remove the empty cells and the cell
borders.

See “boolean ” for other Boolean values that you can use.

CELLHEIGHTMIN=dimension
specifies the minimum height of a cell in the grid. Use this option in conjunction
with the CELLWIDTHMIN= option to set the minimum cell size.

The overall size of the panel is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the number of cells in the grid
increases, the size of each cell decreases. At some point the cell becomes so
small that a meaningful graph cannot be rendered. This option sets the
minimum height threshold for all cells. If the actual cell height becomes smaller,
then no panel is drawn.

Default 100px

See “dimension” on page xiii

CELLWIDTHMIN=dimension
specifies the minimum width of a cell in the grid. Use this option in conjunction
with the CELLHEIGHTMIN= option to set the minimum cell size.

The overall size of the panel is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the number of cells in the grid
increases, the size of each cell decreases. At some point the cell becomes so
small that a meaningful graph cannot be rendered. This option sets the
minimum width threshold for all cells. If the actual cell width becomes smaller,
then no panel is drawn.

Default 100px

See “dimension” on page xiii

COLUMNAXISOPTS=(axis-options)
specifies X-axis options for all columns.

Restriction Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options..

COLUMN2AXISOPTS=(axis-options)
specifies X2-axis options for all columns.

Restriction This option is needed only if you use a plot statement that
supports a secondary X2 axis. If you do not use that statement’s
XAXIS= option to map data to the X2 axis, then this option is
ignored. For more information about how data are mapped to the
 LAYOUT DATAPANEL Statement 85

axes, see “Plot Data Are Mapped to a Designated Axis” on page

1040

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” for a list of options.

COLUMNDATARANGE=AUTO | UNIONALL | UNION

specifies how the X-axes of instances of the graph-prototype are scaled.

AUTO
selects the X-axis scale based on the COLUMNWEIGHT= option.

The X-axis scale is determined, as follows:

n When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.

n When COLUMNWEIGHT=PROPORTIONAL and the column axis is

discrete, UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the X-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).

UNION
scales the X-axis data ranges separately for each column on a per-panel
basis.

The scaling does not span multiple panels.

Default AUTO

Tip Use the COLUMNAXISOPTS= option to control shared axis features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMN2DATARANGE=AUTO | UNIONALL | UNION

specifies how the X2-axes of instances of the graph-prototype are scaled.

AUTO
automatically selects the X2-axis scale based on the COLUMNWEIGHT=
option and the column axis type.

The X2-axis scale is selected as follows:

n When COLUMNWEIGHT=EQUAL (default), UNIONALL is selected.

n When COLUMNWEIGHT=PROPORTIONAL and the column axis is

discrete, UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the X2-axis data ranges across all layout columns and panels (when
PANELNUMBER= is in effect).
86 Chapter 4 / Layout Statements

UNION
scales the X2-axis data ranges separately for each column on a per-panel
basis.

The scaling does not span multiple panels.

Default AUTO

Interaction This option is needed only if you use a plot statement that supports
a secondary X2 axis. If you do not use that statement’s XAXIS=
option to map data to the X2 axis, then this option is ignored. For
more information about how data are mapped to the axes, see “Plot
Data Are Mapped to a Designated Axis” on page 1040

Tip Use the COLUMN2AXISOPTS= option to control shared axis

features.

See The COLUMNWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page xiii

COLUMNS=integer
specifies the number of columns in the layout.

Defaults If this option is not specified and ROWS= is specified, then the
number of data cells (and columns) increases dynamically to allow
all classifier values to be presented.

If both this option and ROWS= are specified, then a grid of that size
is created, regardless of the number of classifier values. If the
number of classifier values is greater than the grid size, then no
graphs are created for some classifier values. If the number of
classifier values is small and the grid size large, then there might be
empty cells created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows,
the cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= and ORDER= options affect how the rows are
populated.

The PANELNUMBER= option enables you to create multiple

smaller grids that completely partition the classifier values.
 LAYOUT DATAPANEL Statement 87

COLUMNWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the columns widths.

EQUAL
all columns have equal width.

PROPORTIONAL
each column has a width that is proportional to the number of discrete
midpoint values that it contains.

Restriction At least one column axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When COLUMNDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect,

COLUMNDATARANGE=AUTO is interpreted as UNION.

If all of the following conditions are true, then the discrete axis
is used to proportion the columns: PROPORTIONAL is in effect,
both the X and X2 axes are used, and only one of the two axes is
discrete. If both axes are discrete, then the X axis is used to
proportion the columns.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in COLUMNAXISOPTS=
and COLUMN2AXISOPTS=. The axis offsets are always set to
one-half of the midpoint spacing regardless of plot type.

Default EQUAL

HEADERBACKGROUNDCOLOR=style-reference | color
specifies the background color of the cell headers.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphHeaderBackground:Color style reference.

Interaction HEADEROPAQUE= TRUE must be in effect in order for the color to

be seen.

HEADERLABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the data labels.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

88 Chapter 4 / Layout Statements

HEADERLABELDISPLAY=NAMEVALUE | VALUE | NONE

specifies the content of the cell headers.

NAMEVALUE
displays the classification variable name and value as a name=value pair in
each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=NAMEVALUE produces cell headers such as the
following:

Country=CANADA
Product=TABLE

VALUE
displays the classification variable value only in each cell header.

Example If the classification variables are Country and Product, then

HEADERLABEL=VALUE produces cell headers such as the
following:

CANADA
TABLE

NONE
suppresses the cell headers.

Default NAMEVALUE

HEADEROPAQUE=TRUE | FALSE
specifies whether the background for cell headers is opaque (TRUE) or
transparent (FALSE).

Default TRUE

Interaction When this option is set to FALSE, the background color for cell
headers is not used.

See “boolean ” on page xii for other Boolean values that you can use.

HEADERPACK=TRUE | FALSE
specifies whether the header cells are consolidated into a comma-separated list
in order to save space.

Note: This option is valid starting in SAS 9.4M1.

The following figure shows the effect of HEADERPACK= on the cell headers in
one row of a data lattice. The data lattice contains classification variables
Country and Year.

The following figure shows the effect of HEADERPACK= on the cell headers in
one row of a data lattice. The data lattice contains classification variables
Country, Year, and Product.
 LAYOUT DATAPANEL Statement 89

Default FALSE

Note If the length of the cell header text exceeds the available width, then
the text is truncated.

Tip If truncation occurs, then use the HEADERSPLITCOUNT= option to

split the cell header text into multiple lines.

See “boolean ” on page xii for other Boolean values that you can use.

HEADERSEPARATOR="string"
specifies one or more characters to place between each value in the cell header
when HEADERPACK=TRUE.

Note: This option is valid starting in SAS 9.4M1.

Default A comma followed by a space

Interaction This option is ignored when HEADERPACK=FALSE.

HEADERSPLITCOUNT=positive-integer
specifies the number of headers to consolidate on a header line before splitting
the text to the next line.

Note: This option is valid starting in SAS 9.4M1.

The following figure shows how HEADERSPLITCOUNT=2 splits the cell header
value in a data panel of classification variables Country, Year, and Product.
90 Chapter 4 / Layout Statements

Default The cell header text is not split

Interaction This option is ignored when HEADERPACK=FALSE.

Note If the length of the cell header text exceeds the available width, then
the text is truncated.

Tip Use the HEADERSEPARATOR= option to specify a different

separator.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether to include grid cells for crossings of the CLASSVARS
variables that contain a missing value.

TRUE
any crossing of the class variables that includes a missing value produces a
row or column of cells in the grid.

FALSE
any crossing of the class variables that includes a missing value does not
produce a row or columns of cells in the grid.

By default, missing class values are included in the classification. When the data
contains missing classification values, cells are created for the missing classes.
The classification headers for the missing values are blank for missing string
values or a dot for missing numeric values. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing values. If you
want to keep the missing values, then you can create a format that specifies
more meaningful headings for the missing classes. For an example, see “Missing
Class Values” in SAS Graph Template Language: User’s Guide.

Note: ODS Graphics does not support Unicode values in user-defined formats
in SAS 9.4M2 and in earlier releases. Starting with SAS 9.4M3, ODS Graphics
supports Unicode values in user-defined formats only if they are preceded by
the (*ESC*) escape sequence. Example: "(*ESC*){unicode beta}". ODS
Graphics does not support an escape character that is defined in an ODS
ESCAPECHAR statement in user-defined formats.

Default TRUE

See “boolean ” on page xii for other Boolean values that you can use.

INSET=(variable-list)
specifies what information is displayed in an inset. The variable-list defines one
or more variables whose names and values appear as a small table in the data
cells. The variables can be either numeric or character. Variable names are
separated by spaces.

Requirement No predefined information is available for the inset. You must

create the desired inset information as part of your input data. See
“Creating Your Inset Data” on page 76.

Note The variable values are associated with the data cells by data
order. That is, the first observation from all the variables in
 LAYOUT DATAPANEL Statement 91

variable-list are used in the first data cell, the second observation
from all variables in variable-list are used in the second data cell,
and so on. If a value is missing for an observation, then the
corresponding name-value pair is skipped in the affected data cell.

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See “Adding Insets to Your Graph” in SAS Graph Template Language:

User’s Guide

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The
appearance options can be any one or more of the following values:

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the inset is automatically aligned within the layout.

NONE
does not automatically align the inset.

This inset’s position is set by the HALIGN= and VALIGN=appearance-

options.

AUTO
attempts to center the inset in the area that is farthest from any
surrounding markers.

Data cells might have different inset placements.

(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with
the data cell’s other graphics features.

The location-list is a space-separated list that can contain any of these

locations: TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT,
BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Example AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the HALIGN= and

VALIGN= options are ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background.

style-reference
specifies a style reference in the form style-element : style-attribute.

Only the COLOR and CONTRASTCOLOR style attributes are valid.

Default The background is transparent. No color is assigned.

92 Chapter 4 / Layout Statements

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

CONTENTDISPLAY=LABELVALUE | VALUE
specifies whether the variable information that is displayed in the inset
includes the column label and value, or only the column value.

Note: This feature applies to SAS 9.4M2 and to later releases.

If a label is not assigned to a column, then the column name is used as the
label for that column. Consider the following inset data:
F
Obs _TYPE_ Value Pr > F
1 SS1 94.359 <.0001

The following figure shows the effect that the CONTENTDISPLAY= option
has on the content of an inset that displays this data.

Default LABELVALUE

Tip Use the SEPARATOR= option to specify a separator other than the
default blank space.

DATASCHEME=LIST | MATCHED
specifies the scheme that was used to merge the inset information into the
analysis data.

LIST
one-to-one merging (no BY statement) was used to merge the inset and
analysis data.
The variable values are associated with the cells of the data lattice by
using data order. That is, the inset variable values in the first observation
are used in the inset for the first cell, the inset variable values in the
second observation are used in the inset for the second cell, and so on.

MATCHED
match-merging (using a BY statement) was used to merge the inset and
analysis data.

Defaults LIST if the data is a SAS data set

MATCHED if the data is a CAS in-memory table

 LAYOUT DATAPANEL Statement 93

Tip MATCHED is the preferred data scheme for merging the inset and
analysis data.

See “Adding Insets to Classification Panels” in SAS Graph Template

Language: User’s Guide

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.

Default LEFT

Interaction This option has an effect only when this layout is nested within a
REGION layout or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When OPAQUE=FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

SEPARATOR="string"
specifies a new separator for the column label and value.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default A blank space

Interaction This option is ignored when CONTENTDISPLAY=VALUE.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset, excluding the title.

Default The GraphDataText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TITLE="string"
specifies a title for the inset.

The title is added at the top of the inset and spans the full inset width.

Note Space is not reserved for the title when this value is not specified.

Tip Text properties for the title string can be specified with
TITLEATTRS=.
94 Chapter 4 / Layout Statements

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the inset.

Default TOP

Interaction This option has effect only when this layout is nested within a
REGION layout or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

Requirements The options must be enclosed in parentheses.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.

ROWMAJOR
fills the data cells by rows, from the starting position.

COLUMNMAJOR
fills the data cells by columns, from the starting position.

Default ROWMAJOR

Interaction The starting point for rendering data cells is controlled by the
START= option. See the START= option for examples.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.
 LAYOUT DATAPANEL Statement 95

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0
96 Chapter 4 / Layout Statements

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

PANELNUMBER=positive-integer
specifies the number of the panel to produce. This option enables you to
partition a large grid into a number of smaller sized grids under these conditions:
n You set a grid size explicitly ( ROWS= and COLUMNS= options).

n The grid size (gridrows x gridcolumns) is smaller than the total number of
classifier levels.
n You execute the template N times and increment the panel number each
time. N is determined by CEIL(total-classification-levels / gridrows
x gridcolumns).

Default 1

Example Suppose there are two classifiers (CLASS1 has 10 unique values and
CLASS2 has 11 unique values). By setting some smaller grid size, say
ROWS=3 and COLUMNS=4, and making the value of
PANELNUMBER= a dynamic or macro variable, you can create 10
panels (9 panels with 12 data cells and 1 panel with 2 data cells) that
collectively display all 110 possible crossings. You simply invoke PROC
SGRENDER or a DATA step 10 times, incrementing the dynamic value
for PANELNUMBER each time.

ROLENAME=(role-name-list)
specifies user-defined roles for information contained in data columns. The role
name list is a space-separated list role-name=column pairs.

Default no user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles.

Example The following example assigns the column Obs to the user-defined
role TIP1.
ROLENAME=(TIP1=OBS)

ROWAXISOPTS=(axis-options)
specifies Y-axis options for all rows.
 LAYOUT DATAPANEL Statement 97

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1218 for a list of options..

ROW2AXISOPTS=(axis-options)
specifies Y2-axis options for all rows.

Requirement Axis options must be enclosed in parentheses and separated by

spaces.

Interaction This option is needed only if you use a plot statement that
supports a secondary Y2 axis. If you do not use that statement’s
YAXIS= option to map data to the Y2 axis, then this option is
ignored. For more information about how data are mapped to the
axes, see “Plot Data Are Mapped to a Designated Axis” on page
1040

See “Axis Options for LAYOUT DATALATTICE and LAYOUT

DATAPANEL” on page 1218 for a list of options..

ROWDATARANGE=AUTO | UNIONALL | UNION

specifies how the Y-axes of instances of the graph-prototype are scaled.

AUTO
automatically selects the Y-axis scale based on the ROWWEIGHT= option
and the column axis type.

The Y-axis scale is selected as follows:

n When ROWWEIGHT=EQUAL (default), UNIONALL is selected.

n When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,

UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the Y-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).

UNION
scales the Y-axis data ranges separately for each row in the layout on a per-
panel basis.

The scaling does not span multiple panels.

Default AUTO

Tip Use the ROWAXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROW2DATARANGE=AUTO | UNIONALL | UNION

specifies how the Y2-axes of instances of the graph-prototype are scaled.
98 Chapter 4 / Layout Statements

AUTO
automatically selects the Y2-axis scale based on the ROWWEIGHT= option
and the column axis type.

The Y2-axis scale is selected as follows:

n When ROWWEIGHT=EQUAL (default), UNIONALL is selected.

n When ROWWEIGHT=PROPORTIONAL and the row axis is discrete,

UNION is selected. Otherwise, UNIONALL is selected.

UNIONALL
scales the Y2-axis data ranges across all layout rows and panels (when
PANELNUMBER= is in effect).

UNION
scales the Y2-axis data ranges separately for each row in the layout on a
per-panel basis.

The scaling does not span multiple panels.

Default AUTO

Interaction This option is needed only if you use a plot statement that supports
a secondary Y2 axis. If you do not use that statement’s YAXIS=
option to map data to the Y2 axis, then this option is ignored. For
more information about how data are mapped to the axes, see “Plot
Data Are Mapped to a Designated Axis” on page 1040

Tip Use the ROW2AXISOPTS= option to control shared axis features.

See The ROWWEIGHT= option.

The PANELNUMBER= option for information about how to create

multiple panels.

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page xiii

ROWS=integer
specifies the number of rows in the layout.

Defaults If this option is not specified and COLUMNS= is specified, then the
number of data cells (and rows) increases dynamically to allow all
classifier values to be presented.

If both this option and COLUMNS= are specified, then a grid of that
size is created, regardless of the number of classifier values. If the
number of classifier values is greater than the grid size, then no
graphs are created for some classifier values. If the number of
 LAYOUT DATAPANEL Statement 99

classifier values is small and the grid size large, then there might be
empty cells created.

Interactions The overall grid size is constrained by the HEIGHT= and WIDTH=
options in the ODS GRAPHICS statement. As the grid size grows,
the cell size shrinks. To control the minimum size of a cell use the
CELLHEIGHTMIN= and CELLWIDTHMIN= options.

The START= and ORDER= options affect how the rows are
populated.

Tip The PANELNUMBER= option enables you to create multiple

smaller grids that completely partition the classifier values.

ROWWEIGHT=EQUAL | PROPORTIONAL
specifies how weights are assigned to the row heights.

EQUAL
all rows have equal height.

PROPORTIONAL
each row has a height that is proportional to the number of discrete midpoint
values that it contains.

Restriction At least one row axis must be discrete in order for

PROPORTIONAL to have any effect. Otherwise, EQUAL is used.

Interactions When ROWDATARANGE=UNIONALL, PROPORTIONAL is

ignored and EQUAL is used.

When PROPORTIONAL is in effect, ROWDATARANGE=AUTO

is interpreted as UNION.

If all of the following conditions are true, then the discrete axis
is used to proportion the rows: PROPORTIONAL is in effect,
both the Y and Y2 axes are used, and only one of the two axes is
discrete. When both axes are discrete, the Y axis is used to
proportion the rows.

When PROPORTIONAL is in effect, the OFFSETMIN= and

OFFSETMAX= axis options are ignored in ROWAXISOPTS= and
ROW2AXISOPTS=. The axis offsets are always set to one-half
of the midpoint spacing regardless of plot type.

Default EQUAL

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Default FALSE

Note If SHRINKFONTS=TRUE, the font scaling applies to all child layouts

within the parent layout. If a child layout also specifies
SHRINKFONTS=TRUE, the fonts for the children of that child layout
are further scaled down.
100 Chapter 4 / Layout Statements

See “boolean ” on page xii for other Boolean values that you can use.

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether the external axes skip the empty cells in a partially filled grid.

TRUE
skips empty cells and "snaps" the external axes to the nearest data cell, both
vertically and horizontally.

Though the empty cells are not displayed, the data cells in the grid are not
enlarged to fill the area.

FALSE
displays external axes at their normal locations, even if there are empty cells
at one or more of the locations.

Whenever the total number of classifier crossings (data cells) is not evenly
divisible by the panel size (columns * rows), the last panel is partially filled with
data cells and padded with empty cells to complete the grid.

Here is an example of a data panel that consists of 16 data cells arranged in a 4-

column, 3-row grid. The following figure shows the default appearance of the
last panel:

When SKIPEMPTYCELLS=TRUE, the empty padding cells of the last panel are
removed and external axis ticks and tick values snap to the data cells:
 LAYOUT DATAPANEL Statement 101

Note that SKIPEMPTYCELLS=TRUE removes only the empty padding cells on

the last panel. It does not remove any data cells that have no crossing values
and therefore no graph (these data cells are displayed when SPARSE=TRUE).

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

SORTORDER=(COLUMNVAR=sort-order-keyword | ROWVAR=sort-order-
keyword)
specifies the order of the cells along the columns and rows.

COLUMNVAR
the column role.

ROWVAR
the row variable role.

sort-order-keyword
a sort-order keyword.

The keyword can be one of the following:

AUTO
sorts using DATA for character data and ascending unformatted for
numeric data.

DATA
retains the data order.

Note In SAS 9.4M6 and SAS 9.4M5, when data stored in a CAS in-
memory table is used and DATA is specified, unpredictable results
might occur. To generate consistent graphs in that case, specify
ASCENDINGFORMATTED or DESCENDINGFORMATTED. Starting
with SAS Viya 3.5 and SAS Studio 5.2, when data stored in a CAS
in-memory table is used and DATA is specified, DATA is ignored,
and AUTO is used instead.
102 Chapter 4 / Layout Statements

ASCENDINGFORMATTED
sorts in ascending order, using the formatted values.

DESCENDINGFORMATTED
sorts in descending order, using the formatted values.

Defaults AUTO for all roles if the data is a SAS data set

ASCENDINGFORMATTED for all roles if the data is a CAS in-memory

table

Tip The placement of the cells within the layout also depends on the
starting location, which is controlled by the START= option.

SPARSE=TRUE | FALSE
specifies whether crossings of the class variables include only the crossings in
the data or all possible crossings.

FALSE
specifies that data cells are created only for crossings of the class variables
that are in the data.

TRUE
specifies that the number of data cells is the product of the unique values
for each classification variable.

By default, if a crossing of the class variables has a missing value as part of the
data, then a data cell is created for it.

Here is an example of a classification panel where the classification variables

are COUNTRY and STATE. There are 3 distinct values of COUNTRY (Canada,
Mexico, and U.S.A.) Within Canada and Mexico there are 4 states, and within
U.S.A. there are 8 states. All state names are unique to each country. Therefore,
there are 16 unique STATE values and 48 unique crossings of COUNTRY and
STATE, but there are data for only 16 of the crossings.

Assume that a data panel layout is created with COLUMNS=6 and

SPARSE=TRUE, meaning to display all possible crossings. This is what the first
row would look like. Blank data cells are added whenever there are no data
values for a crossing:

When SPARSE=FALSE the crossings of the classifiers with no data are

automatically removed. This compacts the display:
 LAYOUT DATAPANEL Statement 103

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

START=TOPLEFT | BOTTOMLEFT
indicates whether to start populating the grid cells from the top left or bottom
left corner. If ORDER=ROWMAJOR (the default) and START=TOPLEFT (the
default), then a 2 row 2 column grid is populated as shown in the following
figure.

If ORDER=ROWMAJOR (the default) and START=BOTTOMLEFT, then a 2 row 2

column grid is populated as shown in the following figure.

If ORDER=COLUMNMAJOR and START=BOTTOMLEFT, then a 2 row 2 column

grid is populated as shown in the following figure.
104 Chapter 4 / Layout Statements

If ORDER=COLUMNMAJOR and START=TOPLEFT, then a 2 row 2 column grid

is populated as shown in the following figure.

Default TOPLEFT

SIDEBAR Optional Argument

ALIGN=BOTTOM | TOP | LEFT | RIGHT
specifies the sidebar’s location within the layout. You can specify up to four
SIDEBAR blocks in a LAYOUT DATAPANEL, one for each of the bottom, top,
left, and right sidebar positions.
n The LAYOUT DATAPANEL automatically aligns a sidebar with the layout
columns or rows.
n Only one statement (such as ENTRY or DISCRETELEGEND) or one layout
block (such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create
multi-line text in a sidebar, nest ENTRY statements within a LAYOUT
GRIDDED block.

Default BOTTOM

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the
sidebar boundaries, set this option to FALSE.

See “boolean ” on page xii for other Boolean values that you can use.

Details

Statement Description
The LAYOUT DATAPANEL statement creates a grid of graphs, based on the values
of one or more classifications variables. The main differences between this layout
and the DATALATTICE layout are that this layout supports more than two
classification variables, and it provides more control over the grid layout.
 LAYOUT DATAPANEL Statement 105

By default, the number of cells in the layout is determined by a crosstabulation

table of all the classification variables plus any empty cells needed to complete the
last row or column of the grid. The contents of each data cell are based on a graph
prototype that you specify in the graph-prototype-block. You can enhance the
display using one or more sidebar-statement-blocks. For classification variables
that have many values, you can use the COLUMNS= option or the ROWS= option,
or both with the PANELNUMBER= option to generate multiple panel displays.

The order of the value pairings for the classification variables is determined by the
order that the variables are specified on the CLASSVARS= argument. The last
named variable’s values vary most rapidly (like nested DO loops). Variable values
are always returned in data order.

By default, the first data cell to be filled is in the layout’s top left corner, and data
cells are filled from left-to-right, top-to-bottom. Use the START= option to change
the starting data cell to the bottom left corner, and use the ORDER= option to
determine whether data cells fill by column or by row. See the START= option for
illustrations on how START= and ORDER= interact to manage the fill sequence for
data cells.

Note: The DATAPANEL layout is designed to be the outermost layout in the

template.

Prototype Block
You must specify a single graph-prototype-block within the LAYOUT DATAPANEL
block, using the following syntax:

LAYOUT PROTOTYPE </options>;

GTL-statements;
ENDLAYOUT;

The graph-prototype-block determines the graphical content of each data cell and
is repeated within each data cell, based on the subsets of the classification
variables.

For more information about the LAYOUT PROTOTYPE block and the list of
available options, see “LAYOUT PROTOTYPE Statement” on page 186.

Sidebar Blocks
A LAYOUT DATAPANEL enables you to display sidebars outside of the axis areas. A
sidebar spans across columns or rows and is useful for displaying information that
applies to all of the columns or all of the rows. For example, sidebars are useful for
displaying a legend.

A SIDEBAR statement has the following syntax:

SIDEBAR </ options >;

GTL-statement(s);
ENDSIDEBAR;
 106 Chapter 4 / Layout Statements

The following example shows a SIDEBAR block that displays a legend at the top of
the layout grid.

sidebar / align=top;
discretelegend 'p' 'a' / across=2;
endsidebar;

Creating Your Inset Data

When you use the INSET= option to insert an inset, no predefined information is
available for the inset. You must create the desired inset information as part of your
input data. This is most typically done as follows (see the chapter on classification
panels and the chapter on insets in SAS Graph Template Language: User’s Guide for
complete examples):
n Create a separate data set for the inset columns making sure that the column
names are different from the other columns used in graph. The number
observations of inset data should match the number of cells in the classification
panel. The ordering of the inset observations should be the same as population
order of the cells of the classification panel, taking into account the
CLASSVARS= argument and the ORDER= and START= options. Typically, the
number of observations for the inset data is smaller than the other input data
for the graph.
n Merge the inset data set with the data set for the graph using a DATA or PROC
SQL step. Do not match-merge the observations of the two data sets (no BY
processing). The resulting data set typically has the inset columns padded with
missing values.
n Use the merged data set to produce the graph, specifying the inset column
names in this option’s variable-list.

Example: LAYOUTDATAPANEL Statement

This example shows a four-column, three-row data panel using two classification
variables. With this layout, each data cell is subsetted and labeled with the values
of the classification variables.
n The ROWDATARANGE=UNION option assures that an axis range is computed
separately for each row using the data ranges of all Y= columns in that row. This
facilitates the visual comparison of the data cells.
n A SIDEBAR block is used to place the legend at the bottom of the lattice.

Example Graph
The following graph was generated by the “Example Program” on page 107:
 LAYOUT DATAPANEL Statement 107

Example Program
proc template;
define statgraph layoutdatapanel;
begingraph;
entrytitle "Annual Furniture Sales Comparisons";
layout datapanel classvars=(country year) /
columns=4 rows=3 rowdatarange=union
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(display=(tickvalues) griddisplay=on
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
sidebar / align=top;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

class country year month;
 108 Chapter 4 / Layout Statements

var actual predict;

output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatapanel;

run;

LAYOUT GLOBALLEGEND Statement

Creates a compound legend containing multiple discrete legends positioned at the bottom of a
graph.

Restrictions: Only one global legend is allowed in a graph.

Continuous legends are not supported inside the global legend.
Requirements: You must place the GLOBALLEGEND block directly inside the BEGINGRAPH block.
Do not embed it in any other GTL statement block.
When the LAYOUT GLOBALLEGEND block is used, all of the template's legend
statements must be specified within the LAYOUT GLOBALLEGEND block. Any
legend statements that are specified outside of the LAYOUT GLOBALLEGEND
block are ignored.
See: “DISCRETELEGEND, MERGEDLEGEND Statements” on page 1303

Syntax
LAYOUT GLOBALLEGEND </options>;
<discreteLegend-statement(s) | mergedLegend-statement(s)>;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
DISPLAYCLIPPED=TRUE | FALSE
specifies whether the global legend is displayed when any portion of its
nested legends cannot be fully rendered because of space constraints.
GUTTER=dimension
specifies the gap between nested layouts.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
 LAYOUT GLOBALLEGEND Statement 109

specifies the amount of extra space that is added inside the layout
border.
TYPE=ROW | COLUMN
specifies whether nested legends are arranged into a single row or
column.
WEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the preferred space allocation for the nested legends.

Legend title options

LEGENDTITLEPOSITION=LEFT | TOP
specifies the position of each nested legend’s title.
TITLE="string"
specifies a title for the global legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the global legend title.

Location options
HALIGN=CENTER | LEFT | RIGHT
specifies the layout’s horizontal alignment within the graph area that is
defined by the BEGINGRAPH block.

Optional Arguments
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the global legend is displayed when any portion of its nested
legends cannot be fully rendered because of space constraints. When the graph
size is reduced, parts of a nested legend (title, legend symbol, or legend value)
might be clipped (truncated). When clipping occurs and this option is FALSE,
the entire global legend is removed from the graph and the space for it is
reclaimed by the remainder of the graph. When this option is TRUE, the global
110 Chapter 4 / Layout Statements

legend always appears, even if some parts of the nested legends have been
clipped.

Default FALSE

Interaction This option overrides any DISPLAYCLIPPED option that is set on its
nested legend statements.

See “boolean ” on page xii for other Boolean values that you can use.

GUTTER=dimension
specifies the gap between nested layouts.

Default 0

Note The default units for dimension are pixels.

See “dimension” on page xiii

HALIGN=CENTER | LEFT | RIGHT

specifies the layout’s horizontal alignment within the graph area that is defined
by the BEGINGRAPH block.

Default CENTER

Note When CENTER is in effect and the outermost layout is an overlay-type

layout, the global legend is centered below the wall area if it can fit
within the wall width.

LEGENDTITLEPOSITION=LEFT | TOP
specifies the position of each nested legend’s title. Specifying LEFT places each
title to the left of the legend items for that legend. Specifying TOP places each
title above the legend items for that legend.

Default LEFT

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.
 LAYOUT GLOBALLEGEND Statement 111

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Notes The default units for dimension are pixels.

Starting with SAS 9.4M1, the default padding between the global
legend and the plot area (including the axes) is increased to 10 pixels. If
the new default padding is not desirable, then use this option to adjust
it.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0
112 Chapter 4 / Layout Statements

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

TITLE="string"
specifies a title for the global legend.

Default No title is displayed for the global legend.

Restriction The string must be enclosed in quotation marks.

Tip The title for the global legend is independent of the titles for its
nested legends.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the global legend title.

Default The GraphLabelText style element.

Interaction For this option to have any effect, the TITLE= option must also be
specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

TYPE=ROW | COLUMN
specifies whether nested legends are arranged into a single row or column.

Default ROW

Interaction When this option is set to ROW, the relative width of each legend is
determined by the setting for the WEIGHTS= option.

WEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the preferred space allocation for the nested legends.

UNIFORM
allocates an equal amount of space for all nested legends.

PREFERRED
allocates the preferred amount of space for each nested legend.

(weight-list)
a space-separated list of preferred space allocations, enclosed in
parentheses. The list can combine numbers with the keyword PREFERRED.
Each number is a proportional weight for the corresponding nested legend
(the weights do not have to sum to 1.0). Keyword PREFERRED specifies that
the preferred space should be allocated to the corresponding nested legend.
The order of the weights that are specified in the list should correspond to
 LAYOUT GLOBALLEGEND Statement 113

the order of the legend statements that are nested in the GLOBALLEGEND
layout.

Default UNIFORM

Restriction The option is supported only for TYPE=ROW.

Tip When a weight-list is specified, all the legends using PREFERRED

get their preferred space. Any remaining space is divided among the
legends, in proportion to the numeric values specified in the weight-
list.

Details
A global legend layout can contain multiple discrete or merged legends. Continuous
legends are not supported inside the global legend block.

A global legend is placed at the bottom of the graph, just above the footnote(s). All
of the discrete or merged legend statements that are nested within the global
legend block are arranged into a single row or column, depending on the setting for
the TYPE= option.

Depending on the outermost layout type and the legend content, the legend is
centered on the graph wall area or on the graph output area. For example, if the
outermost layout is an OVERLAY layout, then when positioning the legend, the
GLOBALLEGEND statement first attempts to center the legend on the graph wall
area. If that position causes the legend to be clipped, then it attempts to center the
legend on the entire output area instead. In that case, the legend might appear to be
slightly off-center with respect to the graph.

Only one global legend block is permitted in a graph. The block must be located
within the BEGINGRAPH block, but outside of the outermost layout block.

When a global legend block is used, only the legend statements that are specified
within the global legend block are displayed in the graph. Any legend statements
that are specified outside of the global legend block are ignored.

Example: LAYOUT GLOBALLEGEND Statement

The following graph was generated by the “Example Program” on page 114:
 114 Chapter 4 / Layout Statements

Example Program
proc template;
define statgraph globallegend;
begingraph;
entrytitle "Prediction Ellipses";
layout overlay;
scatterplot x=petallength y=petalwidth / group=species
name="sp";
ellipse x=petallength y=petalwidth / type=predicted alpha=0.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth / type=predicted alpha=0.05
name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
endlayout;
layout globalLegend / type=column title="Sample Global Legend";
discretelegend "sp" / title="Species:";
discretelegend "p80" "p95" / title="Predictions:";
endLayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=globallegend;

run;
 LAYOUT GRIDDED Statement 115

LAYOUT GRIDDED Statement

Assembles the results of nested GTL-statements into a grid.

Note: Empty cells might be omitted from the grid. See “Details” on page 122.

Syntax
LAYOUT GRIDDED </options>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
COLUMNGUTTER=dimension
specifies the amount of empty space between the columns.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Grid options
COLUMNS=integer
specifies the number of columns in the layout.
ROWS=integer
specifies the number of rows in the layout.

Layout options
ROWGUTTER=dimension
116 Chapter 4 / Layout Statements

specifies the amount of empty space between the rows.

Legend options
LOCATION=INSIDE | OUTSIDE
specifies whether the legend appears inside or outside the plot area when
nested within an overlay-type layout.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent
when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies this layout’s horizontal alignment within its parent when nested
within an overlay-type or REGION layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies this layout’s vertical alignment within its parent when nested
within an overlay-type or REGION layout.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row
priority.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent when
nested within an overlay-type layout.

NONE
does not automatically align the layout within its overlay-type parent layout.

This layout’s position within its parent layout is therefore set by the
HALIGN= and VALIGN= options.

AUTO
within the overlay-type parent layout, attempts to center the layout in the
area that is farthest from any surrounding data point markers.

This option is available only if the parent layout contains a scatter plot.
Otherwise, it is ignored.

(location-list)
within the parent layout, restricts the layout’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the parent layout’s other graphics features.

The location-list is a space-separated list that can contain any of these

locations: TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT,
BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Default NONE
 LAYOUT GRIDDED Statement 117

Restriction This option is ignored if this layout statement is the outermost

layout or if the parent layout is not an overlay-type layout.

Interactions When this option is not NONE and the parent layout is an overlay-
type layout, the HALIGN= and VALIGN= options are ignored.

This option is ignored if LOCATION= OUTSIDE.

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type layout.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

COLUMNGUTTER=dimension
specifies the amount of empty space between the columns.

Default 0
118 Chapter 4 / Layout Statements

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page xiii

COLUMNS=integer
specifies the number of columns in the layout. This option is used to create a
grid with a fixed number of columns, without concern for how many rows. For
example, the following settings ensure that columns 1 and 2 in the first row are
filled with content, as shown in the figure:
layout gridded / columns=2 order=rowmajor border=true ;
entry '1' /border=true;
entry '2' /border=true;
entry '3' /border=true;
endlayout;

Defaults If ORDER= ROWMAJOR, then the default is 1.

If ORDER=COLUMNMAJOR, then as many columns are created as

needed to satisfy the ROWS= request.

Restriction Assuming ORDER=ROWMAJOR, if COLUMNS=n and there are m

cells defined, and n > m, then only m columns are created (there are
n - m cells with zero size).

HALIGN=CENTER | LEFT | RIGHT | number

specifies this layout’s horizontal alignment within its parent when nested within
an overlay-type or REGION layout.

number
specifies the horizontal alignment as a fraction of the parent container’s
width.

Range 0 to 1, where 0 is all the way to the left and 1 is all the way to the
right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a
REGION layout, or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type or
REGION layout.
 LAYOUT GRIDDED Statement 119

LOCATION=INSIDE | OUTSIDE
specifies whether the legend appears inside or outside the plot area when
nested within an overlay-type layout.

Default INSIDE

Restriction This option has effect only when the GRIDDED layout block
appears within an overlay-type layout. For more information about
how child positions are determined in an overlay-type layout, see
“LAYOUT OVERLAY Statement” on page 159.

Interactions If this option is set to OUTSIDE, then the HALIGN= and VALIGN=
options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the layout is
positioned outside of the plot area.

The actual position is determined by this option’s setting plus the

settings for the AUTOALIGN= or HALIGN= and VALIGN= options.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.

ROWMAJOR
fills all the columns in a row, from left to right, before going to the next row.

COLUMNMAJOR
fills all the rows in a column, from top to bottom, before going to the next
column.

Default ROWMAJOR

Requirements If this option is set to COLUMNMAJOR, then the ROWS= option

must be specified to indicate how many rows to fill before
wrapping to the next column. The default number of rows is 1.

If this option is set to ROWMAJOR, then the COLUMNS= option

must be specified to indicate how many columns to fill before
wrapping to the next column. The default number of columns is 1.

Interactions The ROWS= option is ignored when ORDER=ROWMAJOR.

The COLUMNS= option is ignored when

ORDER=COLUMNMAJOR.

See ROWS= on page 121

120 Chapter 4 / Layout Statements

COLUMNS= on page 118

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0
 LAYOUT GRIDDED Statement 121

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page xiii

ROWS=integer
specifies the number of rows in the layout. This option is used to create a grid
with a fixed number of rows, without concern for how many columns. For
example, the following settings ensure that rows 1 and 2 in the first column are
filled with content, as shown in the figure:
layout gridded / rows=2 order=columnmajor border=true ;
entry '1' /border=true;
entry '2' /border=true;
entry '3' /border=true;
endlayout;

Defaults If ORDER= COLUMNMAJOR, then the default is 1.

If ORDER=ROWMAJOR, then this option is ignored and as many

rows are created as needed to satisfy the COLUMNS= request.
122 Chapter 4 / Layout Statements

Restriction Assuming ORDER=COLUMNMAJOR, if ROWS=n and there are m

cells defined, and n > m, then only m rows are created (there are n -
m cells with zero size).

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Default FALSE

Note If SHRINKFONTS=TRUE, the font scaling applies to all child layouts

within the parent layout. If a child layout also specifies
SHRINKFONTS=TRUE, the fonts for the children of that child layout
are further scaled down.

See “boolean ” on page xii for other Boolean values that you can use.

VALIGN=CENTER | TOP | BOTTOM | number

specifies this layout’s vertical alignment within its parent when nested within an
overlay-type or REGION layout.

number
specifies the vertical alignment as a fraction of the parent container’s height.

Range 0 to 1, where 0 is on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a
REGION layout, or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type or
REGION layout.

Details
A GRIDDED layout is commonly used to create small tables of text that are nested
within other layouts. The layout might also be used to span and center a single
entry (a legend, for example) across a set of grids. Or it might be used to display a
grid of graphs when there is no need to scale axis data ranges or align graphs across
grid cells.

The GRIDDED layout automatically decides how much area to allocate to cell
contents:
n Text items have a fixed size based on the amount of text and the font
properties.
 LAYOUT GRIDDED Statement 123

n Graphs take up the remaining space.

The layout’s grid size is determined by the COLUMNS= and ROWS= options. The
resulting columns and rows can be separated by areas called “gutters,” which are
controlled by the COLUMNGUTTER= and ROWGUTTER= options.

By default, the results of the GTL statements are placed into the grid sequentially
from left to right, wrapping to a new row each time the current row is filled. You can
use the ORDER= option to fill cells from top to bottom down a column. In that case,
the layout cells wrap to a new column each time the current column is filled. The
GTL statements can include text statements, plot statements, and layout blocks.
Each statement or layout block provides content for one cell in the grid.

If a statement or layout block for a grid cell does not produce any output, then the
space for that cell might not be retained as an empty cell in the grid. In that case,
the empty cell is removed, and the remaining cells (if any) fill the gap in the grid. A
statement produces no output when required data for that statement does not
resolve. A layout block produces no output when it contains no statements or when
none of the statements contained in the block produce any output.

Examples

Example 1: Creating a 2-By-2 Grid of Plots

This example shows you how to create a simple 2-by-2 grid of plots. Each plot in
the grid must be the same size and have independent axes. The GRIDDED layout is
the ideal choice for these requirements. The following graph was generated by the
“Example Program” on page 124:
 124 Chapter 4 / Layout Statements

Example Program
/* Define the graph template */
proc template;
define statgraph graphgridded;
begingraph / subpixel=on;
entrytitle "Automotive Data";
layout gridded / columns=2 rows=2 rowgutter=5 columngutter=5;
/* Define graph for row 1 cell 1 */
layout overlay / border=on pad=5
xaxisopts=(label="Mean Highway MPG"
griddisplay=on gridattrs=(color=lightgray pattern=dot))
yaxisopts=(label="Type" labelposition=top);
barchart x=type y=Mpg_Highway / stat=mean orient=horizontal
barlabel=true barlabelformat=5.0;
endlayout;

/* Define graph for row 1 cell 2 */

layout overlay / border=on pad=5
yaxisopts=(label="Origin" labelposition=top)
xaxisopts=(label="Models Produced"
griddisplay=on gridattrs=(color=lightgray pattern=dot));
barchart x=Origin y=Models / name="cars" stat=freq
barlabel=true orient=horizontal;
endlayout;

/* Define graph for row 2 cell 1*/

layout overlay / border=on pad=5
xaxisopts=(label="Curb Weight (lbs)"
griddisplay=on gridattrs=(color=lightgray pattern=dot)
linearopts=(tickvaluesequence=(start=2000 end=8000
increment=1000)))
yaxisopts=(label="Horsepower" labelposition=top
griddisplay=on gridattrs=(color=lightgray pattern=dot));
pbsplineplot x=weight y=horsepower;
scatterplot x=weight y=horsepower / datatransparency=0.8;
endlayout;

/* Define graph for row 2 cell 1*/

layout overlay / border=on pad=5
yaxisopts=(label="Drivetrain" labelposition=top
linearopts=(tickvaluesequence=(start=2000 end=8000
increment=1000)))
xaxisopts=(label="Models Produced"
griddisplay=on gridattrs=(color=lightgray pattern=dot));
barchart category=drivetrain / stat=freq
barlabel=true orient=horizontal;
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=graphgridded;

run;
 LAYOUT GRIDDED Statement 125

Example 2: Creating an Inset

The following graph was generated by the “Example Program” on page 125:

Example Program
The GRIDDED layout offers the best way to nest a table of information inside
another layout. In the GRIDDED layout, you can control the content, text
justification, and fonts of columns. Because this example nests the GRIDDED
layout within an OVERLAY layout, you can control where it appears within the plot
area. The AUTOALIGN= option enables you to specify a prioritized list of possible
positions where the layout should be drawn. The position actually used is the first
one that avoids collision with the histogram. Also, the GRIDDED layout is set to be
opaque so that the grid lines do not show through. This example also illustrates a
reusable template in the sense that it works for any numeric column specified by
the dynamic variable VAR. Also, SGE functions for computing the N, MEAN,
STDDEV of the column are used in the table to compute the statistics as the
template is executed.
proc template;
define statgraph inset;
dynamic VAR;
begingraph;
entrytitle "Distribution of " VAR;
layout overlay / yaxisopts=(griddisplay=on);
histogram VAR / scale=percent;
layout gridded / columns=2
autoalign=(topleft topright) border=true
opaque=true backgroundcolor=GraphWalls:color;
 126 Chapter 4 / Layout Statements

entry halign=left "N";

entry halign=left eval(strip(put(n(VAR),12.0)));
entry halign=left "Mean";
entry halign=left eval(strip(put(mean(VAR),12.2)));
entry halign=left "Std Dev";
entry halign=left eval(strip(put(stddev(VAR),12.2)));
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=inset;

dynamic VAR="Weight";
run;

LAYOUT LATTICE Statement

Creates a grid of graphs that automatically aligns plot areas and tick display areas across grid cells
to facilitate data comparisons among graphs.

Note: Empty cells might be omitted from the lattice. See “Cell Contents” on page 144.

Syntax
LAYOUT LATTICE </options>;
GTL-statement(s) | cell-statement-block(s);
<COLUMNAXES;
COLUMNAXIS / axis-options;
<… more-COLUMNAXIS-statements …>
ENDCOLUMNAXES;>
<COLUMN2AXES;
COLUMNAXIS / axis-options;
<… more-COLUMNAXIS-statements …>
ENDCOLUMN2AXES;>
<ROWAXES;
ROWAXIS / axis-options;
<… more-ROWAXIS-statements …>
ENDROWAXES;>
<ROW2AXES;
ROWAXIS / axis-options;
<… more-ROWAXIS-statements …>
ENDROW2AXES;>
 LAYOUT LATTICE Statement 127

<COLUMNHEADERS;
GTL-statement(s);
ENDCOLUMNHEADERS;
<… more-header-statement-block(s) …> >
<SIDEBAR </options>;
GTL-statement(s);
ENDSIDEBAR;>
<… more-sidebar-statement-blocks …> >
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Cell options
COLUMNWEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the proportional width of each column relative to the overall
grid width, not including the headings and sidebars.
ROWWEIGHTS=UNIFORM | PREFERRED | (weight-list)
specifies the proportional height of each row relative to the overall grid
height, not including the headings and sidebars.

Column options
COLUMN2DATARANGE=DATA | UNION | UNIONALL
specifies how the X2-axis data ranges of graphs within the lattice
columns are scaled.
COLUMNDATARANGE=DATA | UNION | UNIONALL
128 Chapter 4 / Layout Statements

specifies how the X-axis data ranges of graphs within the lattice columns
are scaled.

Lattice options
COLUMNS=integer
specifies the number of columns in the lattice.
ROWS=integer
specifies the number of rows in the lattice.
SKIPEMPTYCELLS=TRUE | FALSE
specifies whether row and column axes skip the unused cells in a
partially filled lattice.

Layout options
ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent
when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies this layout’s horizontal alignment within its parent when nested
within an overlay-type or REGION layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies this layout’s vertical alignment within its parent when nested
within an overlay-type or REGION layout.

Panel options
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row
priority.

Row options
ROW2DATARANGE=DATA | UNION | UNIONALL
specifies how the Y2-axis data ranges of graphs within the lattice rows
are scaled.
ROWDATARANGE=DATA | UNION | UNIONALL
specifies how the Y-axis data ranges of graphs within the lattice rows are
scaled.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether this layout is automatically aligned within its parent when
nested within an overlay-type layout.

NONE
does not automatically align the layout within its overlay-type parent layout.

This layout’s position within its parent layout is therefore set by the
HALIGN= and VALIGN= options.
 LAYOUT LATTICE Statement 129

AUTO
within the overlay-type parent layout, attempts to center the layout in the
area that is farthest from any surrounding data point markers.

This option is available only if the parent layout contains a scatter plot.
Otherwise, it is ignored.

(location-list)
within the parent layout, restricts the layout’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the parent layout’s other graphics features.

The location-list is a space-separated list that can contain any of these

locations: TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT,
BOTTOMLEFT, BOTTOM, and BOTTOMRIGHT.

Default NONE

Restriction This option is ignored if this layout statement is the outermost

layout or if the parent layout is not an overlay-type layout.

Interaction When this option is not NONE and the parent layout is an overlay-
type layout, the HALIGN= and VALIGN= options are ignored.

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout. The child layout appears in either the top right or
top left position, based on which position has more “unoccupied”
area.

dynamic VAR STAT1 STAT2 STAT3;

layout overlay;
histogram VAR;
layout lattice / AUTOALIGN=(TOPRIGHT TOPLEFT)
columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

130 Chapter 4 / Layout Statements

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

COLUMNDATARANGE=DATA | UNION | UNIONALL

specifies how the X-axis data ranges of graphs within the lattice columns are
scaled.

DATA
scales the X-axis data ranges separately for each cell in the lattice.

UNION
scales the X-axis data ranges separately for each column in the lattice.

This setting is supported only if all plots across the column can share the
same data range and axis type. For more information, see “Plot Axis Types
Must Agree on Common Axes” on page 1049.

UNIONALL
scales the X-axis data ranges across all columns in the lattice.

This setting is supported only if all plots across the column can share the
same data range and axis type. For more information, see “Plot Axis Types
Must Agree on Common Axes” on page 1049.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous
and the graphs have the same orientation. If graphs cannot use the
same X-axis range or if COLUMNDATARANGE=DATA, then the X-
axis range of each graph is determined from the data.

If any plot statement in any cell contains a XAXIS=X2 option, then

that plot's X values are ignored whenever COLUMNDATARANGE=
is set to UNION or UNIONALL.
 LAYOUT LATTICE Statement 131

Tips If column axes are used and if a lattice cell contains a LAYOUT
OVERLAY statement with the XAXISOPTS= option specified, then
the XAXISOPTS= option is ignored. In that case, use the
COLUMNAXIS statement to specify the desired column X-axis
features. For more information, see “Row and Column Axis
Statements” on page 146 .

By default, each cell has its own axes, which are managed using axis
options in the layout statement for that cell. To consolidate the cell
X axes into column X axes, set this option to UNION or UNIONALL.
Then in the LAYOUT LATTICE statement block, specify a
COLUMNAXES block that includes a COLUMNAXIS statement for
each column X axis that needs to be managed.

COLUMN2DATARANGE=DATA | UNION | UNIONALL

specifies how the X2-axis data ranges of graphs within the lattice columns are
scaled.

DATA
scales the X2-axis data ranges separately for each cell in the lattice.

UNION
scales the X2-axis data ranges separately for each column in the lattice.

This setting is supported only if all plots across the column can share the
same data range and axis type. For more information, see “Plot Axis Types
Must Agree on Common Axes” on page 1049.

UNIONALL
scales the X2-axis data ranges across all columns in the lattice.

This setting is supported only if all plots across the column can share the
same data range and axis type. For more information, see “Plot Axis Types
Must Agree on Common Axes” on page 1049.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous
and the graphs have the same orientation. If graphs cannot use the
same X2-axis range or if COLUMN2DATARANGE=DATA, then the
X2-axis range of each graph is determined from the data.

If any plot statement in any cell contains a XAXIS=X option, then

that plot's X values are ignored whenever COLUMN2DATARANGE=
is set to UNION or UNIONALL.

Tips By default, each cell has its own axes, which are managed using axis
options in the layout statement for that cell. To consolidate the cell
X2 axes into column X2 axes, set this option to UNION or
UNIONALL. Then in the LAYOUT LATTICE statement block, specify
a COLUMN2AXES block that includes a COLUMNAXIS statement
for each column X2 axes that needs to be managed.

If column X2 axes are used and if a lattice cell contains a LAYOUT

OVERLAY statement with the X2AXISOPTS= option specified, then
the X2AXISOPTS= option is ignored. In that case, use one or more
132 Chapter 4 / Layout Statements

COLUMNAXIS statements to specify the desired column X2-axis

features. For more information, see “Row and Column Axis
Statements” on page 146 .

COLUMNGUTTER=dimension
specifies the amount of empty space that is between the columns.

Default 0

Note If there are n columns, then there are n-1 gutters.

See “dimension” on page xiii

COLUMNS=integer
specifies the number of columns in the lattice.

Defaults If ORDER= ROWMAJOR, then the default is 1.

If ORDER=COLUMNMAJOR, then as many columns are created as

are needed to satisfy the ROWS= request.

Interactions If both ROWS=n and COLUMNS=m are specified, then an n by m

grid of cells is created. If the number of statements that define cell
contents is greater than n x m, then the grid size does not expand
and some statements are not displayed. If the number of
statements that define cell contents is less than n x m, then the grid
will contain empty cells.

If this option is not defined and ORDER=COLUMNMAJOR, then the

number of columns is dynamically determined by the number of
defined cells.

COLUMNWEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the proportional width of each column relative to the overall grid
width, not including the headings and sidebars.

UNIFORM
equally divides the total available width among all of the columns.

PREFERRED
specifies that each column gets its preferred width.

Preferred widths are determined as follows:

n Columns that contain one or more vertically one-dimensional plots get
the maximum preferred width from the vertically one-dimensional plots.
n The remaining columns that do not contain vertically one-dimensional
plots get an equal amount of width from the remaining space.

Interaction If a one-dimensional box plot is specified in the preferred

column, then the box plot's BOXWIDTH= option is ignored.

Notes The PREFERRED option is used for LATTICE layouts that mix
one-dimensional and two-dimensional plots in the grid. It
 LAYOUT LATTICE Statement 133

enables the layout to compute the weights automatically for

columns that contain one-dimensional plots.

Examples of one-dimensional plots include axis tables, block

plots, fringe plots, and box plots with Y= values only.

(weight-list)
a space-separated list of column weights.

The list should contain a weight for each column, which can be expressed as
one of the following:

PREFERRED
specifies that the corresponding column gets its preferred width as
described previously.

Note The PREFERRED option should be used on columns that contain

only one-dimensional plots. Using the PREFERRED keyword on
columns that contain a mix of one-dimensional and two-
dimensional plots might cause unexpected results. In that case,
use a numeric weight instead.

Tip The PREFERRED option is particularly useful for axis tables.

number
specifies that the corresponding column gets a width that is based on the
proportion of the specified number to the total of the numbers in the
weight list.

For example, the following weight specifications are equivalent:

columnweights=(0.2 0.3 0.5)

columnweights=(2 3 5)

In these examples, the first column gets 20% of the available width, the
second column gets 30%, and the third column gets 50%.

If the list contains the PREFERRED keyword, then the number specifies
the proportion of the width that remains after the preferred width or
widths are calculated and subtracted from the available width.

Requirements The values in the weight list must be enclosed in parentheses.

If there are n columns in the grid, then the weight list should
contain n weights, one for each column.

All of the numbers in the list must be positive.

At least one number in the list must be nonzero.

Note The weights for all of the columns are specified as a

proportion and, as such, are not required to total 1.0.

Example Here is an example that specifies the preferred width for the
first column, 20% of the remaining width to the second
column, and 80% of the remaining width to the third column.
134 Chapter 4 / Layout Statements

columnweights=(preferred 0.2 0.8)

HALIGN=CENTER | LEFT | RIGHT | number

specifies this layout’s horizontal alignment within its parent when nested within
an overlay-type or REGION layout.

number
specifies the horizontal alignment as a fraction of the parent container’s
width.

Range 0 to 1, where 0 is all the way to the left and 1 is all the way to the
right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a
REGION layout, or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type or
REGION layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout and is positioned in the OVERLAY’s top right
corner.
dynamic VAR STAT1 STAT2 STAT3;
layout overlay;
histogram VAR;
layout lattice / VALIGN=TOP HALIGN=RIGHT
columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether data cells are populated by column priority or by row priority.

ROWMAJOR
fills all the columns in a row, from left to right, before going to the next row.
 LAYOUT LATTICE Statement 135

COLUMNMAJOR
fills all the rows in a column, from top to bottom, before going to the next
column.

Default ROWMAJOR

Requirements If this option is set to COLUMNMAJOR, then the ROWS= option

must be specified to indicate how many rows to fill before
wrapping to the next column. The default number of rows is 1.

If this option is set to ROWMAJOR, then the COLUMNS= option

must be specified to indicate how many columns to fill before
wrapping to the next column. The default number of columns is 1.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.
136 Chapter 4 / Layout Statements

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

ROWDATARANGE=DATA | UNION | UNIONALL

specifies how the Y-axis data ranges of graphs within the lattice rows are
scaled.

DATA
scales the Y-axis data ranges separately for each cell in the lattice.

UNION
scales the Y-axis data ranges separately for each row in the lattice.

This setting is supported only if all plots down the row can share the same
data range and axis type. For more information, see “Plot Axis Types Must
Agree on Common Axes” on page 1049.

UNIONALL
scales the Y-axis data ranges across all rows in the lattice.
 LAYOUT LATTICE Statement 137

This setting is supported only if all plots down the row can share the same
data range and axis type. For more information, see “Plot Axis Types Must
Agree on Common Axes” on page 1049.

Default DATA

Interactions The data ranges are scaled only if the data values are continuous
and the graphs have the same orientation. If graphs cannot use the
same vertical range or if ROWDATARANGE=DATA, then the
vertical range of each graph is determined from the data.

If any plot statement in any cell contains a YAXIS=Y2 option, then

that plot's Y values are ignored whenever ROWDATARANGE= is set
to UNION or UNIONALL.

Tips By default, each cell has its own axes, which are managed using axis
options in the layout statement for that cell. To consolidate the cell
Y axes into a row Y axis, set this option to UNION or UNIONALL.
Then in the LAYOUT LATTICE statement block, specify a
ROWAXES block that contains a ROWAXIS statement for each row
Y axis that needs to be managed.

If row axes are used and if a lattice cell contains a LAYOUT

OVERLAY with the YAXISOPTS= option specified, then the
YAXISOPTS option is ignored. In that case, use one or more
ROWAXIS statements to specify desired row Y-axis features. For
more information, see “Row and Column Axis Statements” on page
146 .

ROW2DATARANGE=DATA | UNION | UNIONALL

specifies how the Y2-axis data ranges of graphs within the lattice rows are
scaled.

DATA
scales the Y2-axis data ranges separately for each cell in the lattice.

UNION
scales the Y2-axis data ranges separately for each row in the lattice.

This setting is supported only if all plots down the row can share the same
data range and axis type. For more information, see “Plot Axis Types Must
Agree on Common Axes” on page 1049.

UNIONALL
scales the Y2-axis data ranges across all rows in the lattice.

This setting is supported only if all plots down the row can share the same
data range and axis type. For more information, see “Plot Axis Types Must
Agree on Common Axes” on page 1049.

Default DATA

Interactions The Y2-axis data ranges are scaled only if the data values are
continuous and the graphs have the same orientation. If graphs
cannot use the same vertical range or if ROW2DATARANGE=DATA,
then the vertical range of each graph is determined from the data.
138 Chapter 4 / Layout Statements

If any plot statement in any cell contains a YAXIS=Y2 option, then

that plot's Y values are ignored whenever ROW2DATARANGE= is
set to UNION or UNIONALL.

Tips By default, each cell has its own axes, which are managed using axis
options in the layout statement for that cell. To consolidate the cell
Y axes into row X2 axes, set this option to UNION or UNIONALL,
and then in the LAYOUT LATTICE statement block, specify a
ROW2AXES block that contains a ROWAXIS statement for each
row Y2 axis that needs to be managed.

If row Y2 axes are used and if a lattice cell contains a LAYOUT

OVERLAY with the Y2AXISOPTS= option specified, then the
Y2AXISOPTS option is ignored. In that case, use one or more
ROWAXIS statements to specify the desired row Y2-axis features.
For more information, see “Row and Column Axis Statements” on
page 146 .

ROWGUTTER=dimension
specifies the amount of empty space between the rows.

Default 0

Note If there are n rows, then there are n-1 gutters.

See “dimension” on page xiii

ROWS=integer
specifies the number of rows in the lattice.

Defaults If ORDER=COLUMNMAJOR, then the default is 1.

If ORDER=ROWMAJOR, then as many ROWS are created as

needed to satisfy the COLUMNS= request.

Interactions If both ROWS=n and COLUMNS=m are specified, then an n by m

grid of cells is created. If the number of statements that define cell
contents is greater than n x m, then the grid size does not expand
and some statements are not displayed. If the number of
statements that define cell contents is less than n x m, then the grid
will contain empty cells.

If ORDER=ROWMAJOR and ROWS= is not defined, then the

number of rows is dynamically determined by the number of
defined cells.

ROWWEIGHTS=UNIFORM | PREFERRED | (weight-list)

specifies the proportional height of each row relative to the overall grid height,
not including the headings and sidebars.

UNIFORM
equally divides the total available height among all of the rows.

PREFERRED
specifies that each row gets its preferred height.
 LAYOUT LATTICE Statement 139

Preferred height is determined as follows:

n Rows that contain one or more horizontally one-dimensional plots get
the maximum preferred height from the horizontally one-dimensional
plots.
n The remaining rows that do not contain horizontally one-dimensional
plots get an equal amount of height from the remaining space.

Notes The PREFERRED option is used for LATTICE layouts that mix one-
dimensional and two-dimensional plots in the grid. It enables the
layout to compute the weights automatically for rows that contain
one-dimensional plots.

Examples of one-dimensional plots include axis tables, block plots,

fringe plots, and box plots with Y= values only.

(weight-list)
a space-separated list of row weights.

The list should contain a weight for each row, which can be expressed as one
of the following:

PREFERRED
specifies that the corresponding row gets its preferred height as
described previously.

Note The PREFERRED option should be used on rows that contain only
one-dimensional plots. Using the PREFERRED keyword on rows
that contain a mix of one-dimensional and two-dimensional plots
might cause unexpected results. In that case, use a numeric weight
instead.

number
specifies that the corresponding row gets a height that is based on the
proportion of the specified number to the total of the numbers in the
weight list.

For example, the following weight specifications are equivalent:

rowweights=(0.2 0.3 0.5)

rowweights=(2 3 5)

In these examples, the first row gets 20% of the available height, the
second row gets 30%, and the third row gets 50%.

If the list contains the PREFERRED keyword, then the number specifies
the proportion of the height that remains after the preferred height or
heights are calculated and subtracted from the available height.

Requirements The values in the weight list must be enclosed in parentheses.

If there are n rows in the grid, then the weight list should
contain n weights, one for each row.

All of the numbers in the list must be positive.

140 Chapter 4 / Layout Statements

At least one number in the list must be nonzero.

Note The weights for all of the rows are specified as a proportion
and, as such, are not required to total 1.0.

Example Here is an example that specifies 25% of the available height

for the first row, 25% for the second row, and 50% for the third
row.
rowweights=(1 1 2)

SHRINKFONTS=TRUE | FALSE
specifies whether fonts in the layout are scaled down.

Default FALSE

Note If SHRINKFONTS=TRUE, the font scaling applies to all child layouts

within the parent layout. If a child layout also specifies
SHRINKFONTS=TRUE, the fonts for the children of that child layout
are further scaled down.

See “boolean ” on page xii for other Boolean values that you can use.

SKIPEMPTYCELLS=TRUE | FALSE
specifies whether row and column axes skip the unused cells in a partially filled
lattice.

FALSE
displays row and column axes at their normal locations.

TRUE
skips empty cells and “snaps” the row and column axes to the nearest
populated cell, both vertically and horizontally.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

VALIGN=CENTER | TOP | BOTTOM | number

specifies this layout’s vertical alignment within its parent when nested within an
overlay-type or REGION layout.

number
specifies the vertical alignment as a fraction of the parent container’s height.

Range 0 to 1, where 0 is on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option has effect only when this layout is nested within a
REGION layout, or when this layout is nested in an overlay-type
layout and AUTOALIGN=NONE.
 LAYOUT LATTICE Statement 141

See “LAYOUT OVERLAY Statement” on page 159 for more information

about how child positions are determined in an overlay-type or
REGION layout.

Example In the following example, the LATTICE layout is the child of the
OVERLAY layout. The child layout appears in either the top right or
top left position, based on which position has more "unoccupied"
area.
dynamic VAR STAT1 STAT2 STAT3;
layout overlay / height=500px width=600px;
histogram VAR;
layout lattice / VALIGN=TOP HALIGN=RIGHT
height=80px width=70px columns=1;
entry STAT1;
entry STAT2;
entry STAT3;
endlayout;
endlayout;

SIDEBAR Optional Arguments

ALIGN=TOP | BOTTOM | LEFT | RIGHT
specifies the alignment of the sidebar.

Default BOTTOM

Tip You can use ENTRY statements to place rotated text in the right or left
side bar.

SPACEFILL=TRUE | FALSE
specifies whether to fill all the area of the sidebar with its contents.

Default TRUE

Tip To prevent a layout block within the sidebar from expanding to the
sidebar boundaries, set this option to FALSE.

See “boolean ” on page xii for other Boolean values that you can use.

Details

Statement Description
The LAYOUT LATTICE statement creates a grid of graphs that are aligned across
rows and columns. For plot statements that are specified in the layout block or
nested in a LAYOUT OVERLAY statement, the LATTICE layout automatically aligns
the plot areas and tick display areas in the plots.

Note: To achieve the alignment, the LATTICE layout automatically aligns plot
areas and tick display areas across rows and columns. Also, it overrides axis offset
settings in the OVERLAY layouts that you specify in those rows and columns. (For
142 Chapter 4 / Layout Statements

details about offsets and the tick display area, see “Adjusting Axis Offsets” on page
1053.) If you do not want this alignment, then you might use LAYOUT GRIDDED
instead. For example, if you have a heterogeneous panel of graphs, such as a mix of
scatter plots, box plots, bar charts, or other types of graphs, then you might
consider using LAYOUT GRIDDED rather than LAYOUT LATTICE.

The layout can unify the scale of the data ranges that are displayed in the plots,
based on the values set for the COLUMNDATARANGE= and ROWDATARANGE=
options. If one or more plots within the template use the XAXIS= option to produce
independent X2 (top) axes, then the X2 data scales can be unified, based on the
values set for the COLUMN2DATARANGE= option. If one or more plots within the
template use the YAXIS= option to produce independent Y2 (right) axes, then the
Y2 data scales can be unified, based on the values set for the ROW2DATARANGE=
options. The data ranges can be scaled separately for each column, for each row, or
for both. Or they can be scaled across all columns, all rows, or all of both.

When the data-range scales are unified, you can simplify the lattice by displaying
only the row and column axes that apply to all of the graphs across the
corresponding columns or rows. See “Row and Column Axis Statements” on page
146 for more details.

The following figure shows the parts of the LATTICE layout with the default axis
display (internal axes are displayed).
 LAYOUT LATTICE Statement 143

This next figure shows the parts of the LATTICE layout when the graph display is
simplified so that only row and column axes are displayed.

Note: The figure shows secondary X (top) and secondary Y (right) axes. The layout
also enables you to generate independent X2 (top) and independent Y2 (right)
axes. For details, see “Row and Column Axis Statements” on page 146

The rows and columns can be separated by areas called “gutters,” which are
controlled by the COLUMNGUTTER= and ROWGUTTER= options. In addition, the
COLUMNWEIGHTS= and ROWWEIGHTS= options can be used to allocate a
proportion of available space to each row and column.

The LATTICE layout automatically decides how much area to allocate to cell
contents:
n Text items have a fixed size based on the amount of text and the font
properties.
n Graphs take up the remaining space.

The layout’s grid size is determined by the COLUMNS= and ROWS= options.

By default, the results of the GTL-statements are placed into the grid sequentially
from left to right, wrapping to a new row each time the current row is filled. You can
144 Chapter 4 / Layout Statements

use the ORDER= option to fill cells from top to bottom down a column. In that case,
the lattice cells wrap to a new column each time the current column is filled.

Cell Contents
The following general syntax is used to define the contents of each cell in a
LAYOUT LATTICE:

GTL-statement(s) | cell-statement-block(s)

A cell-statement-block, when used, has the following syntax:

CELL;
<CELLHEADER; GTL-statement(s); ENDCELLHEADER;>
GTL-statement(s);
ENDCELL;

The following guidelines apply to defining cell content:

n The contents of each cell is generated by GTL statements, which can be
specified independently or enclosed in a CELL block.
n Independent GTL statements include text statements, plot statements, or
layout blocks. Each independently specified GTL statement or layout block
provides content for one cell.
n A CELL block can include text statements, plot statements, or layout blocks.
Each CELL block provides content for one cell.
n Within a CELL block, you can use a CELLHEADER block to generate one or
more header lines within the cell. Specify each header line on a separate GTL
statement within the CELLHEADER block. The header block is typically used to
specify one or more text statements, but other statements are allowed within
the block. For example, you can specify a LAYOUT GRIDDED statement to
produce a grid of text for the header
n You can use only one CELLHEADER block per CELL block. If you specify more
than one, then only the last one is used.
n If you do not specify a CELLHEADER block in a CELL block, then the enclosed
GTL statements produce the same results that they would produce if they were
specified independently.
n If more than one plot statement is needed to generate contents for a cell, you
should place the plot statements in a layout block such as LAYOUT OVERLAY.
Otherwise, unexpected results might occur. This applies to independent GTL
statements and to GTL statements in a CELL block. See Figure 4.3 on page 145.

If a CELL block, or an independent statement or layout block for a lattice cell does
not produce any output, then the space for that cell might not be retained as an
empty cell in the lattice. In that case, the empty cell is removed, and the remaining
cells (if any) fill the gap in the lattice. A statement produces no output when
required data for that statement does not resolve. A layout block produces no
output when it contains no statements or when none of the statements contained
in the block produce any output.
 LAYOUT LATTICE Statement 145

The example code shows a LAYOUT LATTICE block that uses one GTL statement
and one CELL block to generate the two-column lattice shown in the following
figure:

Figure 4.1 Lattice Layout with Independent Plot Statements and a CELL Block

proc template;
define statgraph cellcontents;
begingraph;
layout lattice /
columngutter=5 columns=2;

/* independent plot statement - defines first cell */

scatterplot x=age y=height;

/* cell block - defines second cell */

cell;
cellheader;
entry "Cell Header" / border=true;
endcellheader;
/* two plot statements are needed - enclose
the statements in a LAYOUT OVERLAY block */
layout overlay;
scatterplot x=weight y=height;
referenceline y=53 / lineattrs=(pattern=shortdash)
curvelabellocation=inside curvelabel="Reference Line";
endlayout;
endcell;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=cellcontents;

run;
146 Chapter 4 / Layout Statements

Row and Column Axis Statements

Overview
By default, each cell in the lattice has its own internal axes, which are managed
using axis options in the layout statement for that cell. Here is an example of a
lattice with internal axes.

Notice that the axes ranges for each plot in the lattice differ. If the data ranges for
each plot are comparable, you can unity the axes to a common scale for all of the
plots in the lattice. In that case, you can also simplify and clarify your lattice by
creating row and column axes that are shared by the cells in the lattice. Here is an
example of a lattice with row and column axes.

Row and column axes are defined by row and column axis statement blocks in the
LAYOUT LATTICE statement block. Row and column axis statement blocks are
 LAYOUT LATTICE Statement 147

useful only if the data ranges across the affected columns or rows are comparable
and can be unified to a common scale. For example, row and column axes are not
supported if an affected lattice cell contains a LAYOUT OVERLAYEQUATED
statement. If the axis ranges are not unified for the affected columns or rows, then
the row and column axis statement blocks in the LAYOUT LATTICE statement
block are ignored.

Use the options shown in the following table to unify the axis data ranges in the
lattice.

Axis Option

X COLUMNDATARANGE=

X2 COLUMN2DATARANGE=

Y ROWDATARANGE=

Y2 ROW2DATARANGE=

Specifying Row and Column Axis Features

For columns, the features for the column X axes (bottom) are specified within a
COLUMNAXES block, which contains one COLUMNAXIS statement for each
column X axis that needs to be manage. The COLUMNAXIS statement provides a
DISPLAYSECONDARY= option, which enables you to display a column secondary X
(top) axis that mirrors the column primary X axis but can have different display
features. In that case, the axis features that you specify in the COLUMNAXIS
statement apply to both the column primary and column secondary X axes.

If one or more plots within the LAYOUT LATTICE statement block use the
XAXIS=X2 option to produce independent X2 (top) axes, then axis features for the
column X2 axes are specified within a COLUMN2AXES block, nesting one
COLUMNAXIS statement for each column that contains an X2 axis that needs to be
managed. Within the COLUMN2AXES block, the COLUMNAXIS statement’s
DISPLAYSECONDARY= option enables you to display a column secondary X2
(bottom) axis that mirrors the column primary X2 axis but can have different
display features. Here again, the axis features that you specify in the COLUMNAXIS
statement apply to both the column primary and the column secondary X2 axes.

Note: If you specify independent X and X2 scales at the same time, then the
DISPLAYSECONDARY= option is ignored in the COLUMNAXIS statement. This is
true whether the COLUMNAXIS statement is specified in a COLUMNAXES or
COLUMN2AXES block.

For both the COLUMNAXES and COLUMN2AXES blocks, if a lattice cell within the
column contains a LAYOUT OVERLAY statement with the XAXISOPTS= or
X2AXISOPTS= option specified, these LAYOUT OVERLAY statement options are
ignored. In that case, specify the desired axis features in one or more
COLUMNAXIS statements in the COLUMNAXES block.
148 Chapter 4 / Layout Statements

For rows, axis features for the row Y axes (left) are specified within a ROWAXES
block, nesting one ROWAXIS statement for each row that contains a Y axis that
needs to be managed. The ROWAXIS statement provides a DISPLAYSECONDARY=
option, which enables you to display a row secondary Y (right) axis that mirrors the
row primary Y axis but can have different display features. In that case, the axis
features that you specify in the ROWAXIS statement apply to both the row primary
and row secondary Y axes.

If one or more plots within the LAYOUT LATTICE statement block use the YAXIS=
option to produce independent Y2 (right) axes, then axis features for the row Y2
axes (right) are specified within a ROW2AXES block, nesting one ROWAXIS
statement for each row that contains a Y2 axis that needs to be managed. Within
the ROW2AXES block, the ROWAXIS statement’s DISPLAYSECONDARY= option
enables you to display a row secondary Y2 (left) axis that mirrors the row primary
Y2 axis but can have different display features. Here again, the axis features that
you specify in the ROWAXIS statement apply to both the row primary and row
secondary Y2 axes.

Note: If you specify independent Y and Y2 scales at the same time, then the
DISPLAYSECONDARY= option is ignored in the ROWAXIS statement. This is true
whether the ROWAXIS statement is specified in a ROWAXES or ROW2AXES block.

For both the ROWAXES and ROW2AXES blocks, if a lattice cell within the row
contains a LAYOUT OVERLAY statement with the YAXISOPTS= or Y2AXISOPTS=
option specified, these LAYOUT OVERLAY statement options are ignored. In that
case, specify the desired axis features in one or more ROWAXIS statements in the
ROWAXES block.

Syntax and Restrictions for the Axis Statements

The following table shows the general syntax for row and column axis-statement
blocks.

Row or Column X and Y Axes Row or Column X2 and Y2 Axes

COLUMNAXES; COLUMN2AXES;
COLUMNAXIS / axis-options; COLUMNAXIS / axis-options;
<…COLUMNAXIS-n;> <…COLUMNAXIS-n;>
ENDCOLUMNAXES; ENDCOLUMN2AXES;

ROWAXES; ROW2AXES;
ROWAXIS / axis-options; ROWAXIS / axis-options;
<…ROWAXIS-n;> <…ROWAXIS-n;>
ENDROWAXES; ENDROW2AXES;

In the LAYOUT LATTICE block, the following restrictions apply:

 LAYOUT LATTICE Statement 149

n If the LAYOUT LATTICE statement sets the row or column axis data range to
DATA, then the corresponding axes block is ignored. The data range must be set
to UNION or UNIONALL to create row and column axes.
n Only one COLUMNAXES block can be used to manage column X axes, and only
one COLUMN2AXES block can be used to manage column X2 axes. If more than
one of either block is specified, then only the last one of that block type is used.
n Within a COLUMNAXES or COLUMN2AXES block, one COLUMNAXIS
statement should be specified for each column that contains axes that you need
to manage. Both axes blocks can contain a COLUMNAXIS statement for the
same column. For example, to manage the axes in the first column of the lattice,
the COLUMNAXES block can contain a COLUMNAXIS statement that manages
the column’s X axes. The COLUMN2AXES block can contain a COLUMNAXIS
statement that manages the column’s X2 axes.
n Only one ROWAXES block can be used to manage row Y axes, and only one
ROWAXES block can be used to manage row Y2 axes. If more than one of either
block is specified, then only the last one of that block type is used.
n Within a ROWAXES or ROW2AXES block, one ROWAXIS statement should be
specified for each row that contains axes that need to be managed. Both axes
blocks can contain a ROWAXIS statement for the same row. For example, to
manage the axes in the first row of the lattice, the ROWAXES block can contain
a ROWAXIS statement that manages the row’s Y axes. The ROW2AXES block
can contain a ROWAXIS statement that manages the column’s Y2 axes.
n If the number COLUMNLAXIS or ROWAXIS statements is greater than the
number needed, then the extra statements are ignored. If the number of
statements is fewer than the number needed, then the additional COLUMNAXIS
or ROWAXIS statements are automatically generated with DISPLAY=NONE
option in effect for each.

For the list of axis-options, see “Row and Column Axis Options for LAYOUT
LATTICE” on page 1140 .

The following example shows a LAYOUT LATTICE block that uses a ROWAXES
block to set the row Y axis and to display grid lines for the row display.
begingraph;
layout lattice /
rowdatarange=union
columns=2;

/* axis definitions */
rowaxes;
rowaxis /griddisplay=on;
endrowaxes;

/* cell contents */
scatterplot x=x y=t;
scatterplot x=x y=y;

endlayout;
endgraph;

Here, the LAYOUT LATTICE statement specifies the ROWDATARANGE option to

unify the data ranges across rows in the lattice. Because LAYOUT LATTICE
150 Chapter 4 / Layout Statements

specifies COLUMNS=2 and there are two plot statements in the template, the
resulting graph has two columns and only one row. Thus, only one ROWAXIS
statement is needed in the ROWAXES block to specify axis attributes for that row
of graphs. A ROW2AXES block is not needed because neither SCATTERPLOT
statement in the template maps data to the Y2 axis.

For more information and examples that demonstrate how data are mapped to the
axes, see “Plot Data Are Mapped to a Designated Axis” on page 1040 .

Header Statements
Header statements are used to display one or more headers for the rows and
columns in a LATTICE layout. Each statement is specified as a block in the form
statement - ENDstatement. The header block is typically used to specify one or
more text statements, but other statements are allowed within the block. For
example, you could specify a LAYOUT GRIDDED statement to produce a grid of
text for the header.

The general syntax for a COLUMNHEADERS statement is

COLUMNHEADERS;
GTL-statement(s);
ENDCOLUMNHEADERS;

The following header statements are available:

Header Statement Description

COLUMNHEADERS specifies a header for the primary (bottom) column-

header position.

COLUMN2HEADERS specifies a header for the secondary (top) column-

header position.

ROWHEADERS specifies a header for the primary (left) row-header

position. ENTRY statements can be used to specify
rotated text.

ROW2HEADERS specifies a header for the secondary (right) row-

header position. ENTRY statements can be used to
specify rotated text.

n The LAYOUT LATTICE aligns headers with the columns, or the rows, or both.

n Each of the header blocks COLUMNHEADERS, COLUMN2HEADERS,

ROWHEADERS, and ROW2HEADERS can be used once in a LAYOUT LATTICE
block. If more than one block is specified for one of the statements, then only
the last specified block for that statement is used.

The following example shows a LAYOUT LATTICE block that uses a

COLUMNHEADERS block to display column headers above the left and right
columns in the lattice.
 LAYOUT LATTICE Statement 151

begingraph;
layout lattice / columns=2;

/* Lattice header definitions */

columnheaders;
entry "Left Column";
entry "Right Column";
endcolumnheaders;

/* cell contents */
scatter x=x y=t;
scatter x=x y=y;

endlayout;
endgraph;

Sidebar Statements
A LAYOUT LATTICE supports the display of a sidebar between a row or column
header and a row or column axis. (See the figures in “Statement Description” on
page 141 .) A sidebar spans across columns or rows and is useful for displaying
information that applies to all of the columns or all of the rows. For example,
sidebars are useful for displaying legends.

A SIDEBAR statement has the following syntax:

SIDEBAR </options>;
GTL-statement(s);
ENDSIDEBAR;

n You can specify up to four SIDEBAR blocks in a LAYOUT LATTICE, one for each
of the top, bottom, left, and right sidebar positions.
n If two or more SIDEBAR blocks have the same alignment, then the sidebar
information forms two or more columns (ALIGN=LEFT or ALIGN=RIGHT) within
the sidebar area. Or it forms two or more rows (ALIGN=TOP or
ALIGN=BOTTOM) within the sidebar area.
n Only one statement (such as ENTRY or DISCRETELEGEND) or one layout block
(such as LAYOUT GRIDDED) is allowed in a SIDEBAR block. To create multi-line
text in a sidebar, nest ENTRY statements within a LAYOUT GRIDDED block.
n The LAYOUT LATTICE automatically aligns a sidebar with the lattice columns
or rows.

The following example shows a LAYOUT LATTICE block that uses a SIDEBAR
block to display a top sidebar in the lattice.
begingraph;
layout lattice / columns=2;

sidebar / align=top;
layout gridded / border=true ;
entry "Top Sidebar" ;
entry "(spans both columns)";
endlayout;
endsidebar;
 152 Chapter 4 / Layout Statements

scatterplot x=x y=t;

scatterolot x=x y=y;

endlayout;
begingraph;

Examples

Example 1: Lattice with Internal Axes

This example shows you how to create a four-cell lattice of graphs (two rows, two
columns), where each graph has its own, independently managed X and Y axes. The
following features are used in this example:
n the following LAYOUT LATTICE statement options:

o The COLUMNS= option specifies the number of columns in the lattice.

o The COLUMNGUTTER= and ROWGUTTER= options specify the spacing
between the lattice cells.
n A COLUMN2HEADER block in the LAYOUT LATTICE statement block displays
headers at the top of each column. ENTRY statements provide the text for each
header.
n Four LAYOUT OVERLAY statement blocks in the LAYOUT LATTICE statement
block provide the content for each cell.
n Axis options in each LAYOUT OVERLAY statement manage the axes for each
plot.

Example Graph
The following graph was generated by the “Example Program” on page 153:
 LAYOUT LATTICE Statement 153

Example Program
/* Create the stock data for Microsoft for the years 2001 and 2002 */
data stockyear1 stockyear2;
set sashelp.stocks(where=(stock eq "Microsoft" and year(date) in
(2001 2002)));
volume = volume/1000000;
if year(date) = 2001 then
output work.stockyear1;
else if year(date)=2002 then
output work.stockyear2;
run;

data msstock;
merge work.stockyear1(rename=(date=year1date close=year1close
volume=year1vol))
work.stockyear2(rename=(date=year2date close=year2close
volume=year2vol));
format year1date year2date monname3. year1close year2close dollar6.;
run;

/* Create the graph template */

proc template;
define statgraph graphlattice;
begingraph;
entrytitle "Microsoft Stock Performance in 2001 and 2002";
layout lattice / columns=2 columngutter=5 rowgutter=5;
/* Column headers */
column2headers;
entry textattrs=GraphData1(weight=bold size=9pt) "Year 2001";
entry textattrs=GraphData2(weight=bold size=9pt) "Year 2002";
endcolumn2headers;
154 Chapter 4 / Layout Statements

/* Row 1 content (cells 1-2) */

layout overlay /
xaxisopts=(display=(ticks tickvalues)
timeopts=(viewmax='01DEC2001'd tickvalueformat=monname3.))
yaxisopts=(label="Closing Price");
seriesplot x=year1date y=year1close / display=all
smoothconnect=true
lineattrs=GraphData1 markerattrs=GraphData1;
endlayout;
layout overlay /
xaxisopts=(display=(ticks tickvalues)
timeopts=(viewmax='01DEC2002'd tickvalueformat=monname3.))
yaxisopts=(label="Closing Price");
seriesplot x=year2date y=year2close / display=all
smoothconnect=true
lineattrs=GraphData2 markerattrs=GraphData2;
endlayout;

/* Row 2 content (cells 3-4) */

layout overlay /
xaxisopts=(display=(ticks tickvalues)
timeopts=(viewmax='01DEC2001'd tickvalueformat=monname3.))
yaxisopts=(label="Volume (Millions)");
needleplot x=year1date y=year1vol / displaybaseline=off
lineattrs=GraphData1;
endlayout;
layout overlay /
xaxisopts=(display=(ticks tickvalues)
timeopts=(viewmax='01DEC2002'd tickvalueformat=monname3.))
yaxisopts=(label="Volume (Millions)");
needleplot x=year2date y=year2vol / displaybaseline=off
lineattrs=GraphData2;
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=msstock template=graphlattice;

run;

Example 2: Lattice with Row and Column Axes

In the lattice shown in “Example 1: Lattice with Internal Axes” on page 152, because
the axes in each cell are independently managed, the axis ranges can vary across
the plots. However, the axis ranges in each plot in that lattice are comparable. In
that case, you can use options COLUMNDATARANGE= and ROWDATARANGE= in
the LAYOUT LATTICE statement to create uniform axes for the plots in individual
columns or rows, or for all of the plots in all of the columns or rows. The axes are
scaled to accommodate the data ranges in all of the included plots. In addition, you
can use COLUMNAXES and ROWAXES statement blocks in the LAYOUT LATTICE
statement block to consolidate the individual internal axes into common row and
column axes. This example shows you how to modify the lattice in “Example 1:
 LAYOUT LATTICE Statement 155

Lattice with Internal Axes” on page 152 to use common row and column axes for the
plots in the lattice. The following features are used in this example:
n the following LAYOUT LATTICE statement options:

o The COLUMNS= option specifies the number of columns in the lattice.

o The COLUMNGUTTER= and ROWGUTTER= options specify the spacing
between the lattice cells.
o The COLUMNDATARANGE=UNION and ROWDATARANGE=UNION options
specify uniform axes for each row and column.
n A COLUMN2HEADER block in the LAYOUT LATTICE statement block displays
headers at the top of each column, and ENTRY statements provide the header
text.
n A COLUMNAXES statement block in the LAYOUT LATTICE statement block
specifies COLUMNAXIS statements, which manage each column axis. Likewise,
a ROWAXES statement block provides ROWAXIS statements, which manage
each row axis.
n Four LAYOUT OVERLAY statement blocks in the LAYOUT LATTICE statement
block provide the content for each cell. No axis options can be used in the
LAYOUT OVERLAY statements in this case because axes blocks are used to
manage common row and column axes.

Example Graph
The following graph was generated by the “Example Program” on page 156:
 156 Chapter 4 / Layout Statements

Example Program
/* Create the stock data for Microsoft for the years 2001 and 2002 */
data stockyear1 stockyear2;
set sashelp.stocks(where=(stock eq "Microsoft" and year(date) in
(2001 2002)));
volume = volume/1000000;
if year(date) = 2001 then
output work.stockyear1;
else if year(date)=2002 then
output work.stockyear2;
run;

data msstock;
merge work.stockyear1(rename=(date=year1date close=year1close
volume=year1vol))
work.stockyear2(rename=(date=year2date close=year2close
volume=year2vol));
format year1date year2date monname3. year1close year2close dollar6.;
run;

/* Create the graph template */

proc template;
define statgraph graphlattice;
begingraph;
entrytitle "Microsoft Stock Performance in 2001 and 2002";
layout lattice / columns=2 columndatarange=union columngutter=5
rowdatarange=union rowgutter=5;
/* Column headers */
column2headers;
entry textattrs=GraphData1(weight=bold size=9pt) "Year 2001";
entry textattrs=GraphData2(weight=bold size=9pt) "Year 2002";
endcolumn2headers;

/* Column and row axes */

columnaxes;
columnaxis / display=(ticks tickvalues)
timeopts=(viewmax='01DEC2001'd tickvalueformat=monname3.);
columnaxis / display=(ticks tickvalues)
timeopts=(viewmax='01DEC2002'd tickvalueformat=monname3.);
endcolumnaxes;
rowaxes;
rowaxis / label="Closing Price";
rowaxis / label="Volume (Millions)";
endrowaxes;

/* Row 1 content */
layout overlay;
seriesplot x=year1date y=year1close / display=all
smoothconnect=true
lineattrs=GraphData1 markerattrs=GraphData1;
endlayout;
layout overlay;
seriesplot x=year2date y=year2close / display=all
smoothconnect=true
 LAYOUT LATTICE Statement 157

lineattrs=GraphData2 markerattrs=GraphData2;
endlayout;

/* Row 2 content */
layout overlay;
needleplot x=year1date y=year1vol / displaybaseline=off
lineattrs=GraphData1;
endlayout;
layout overlay;
needleplot x=year2date y=year2vol / displaybaseline=off
lineattrs=GraphData2;
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=msstock template=graphlattice;

run;

Example 3: Lattice with Sidebar

This example shows a two-cell lattice (two columns, one row) with a legend placed
in a sidebar at the top of the graph. A ROWAXES block is used in the example to
create a row Y axis. The following features are used in this example:
n The ROWDATARANGE=UNION option in the LAYOUT LATTICE statement
ensures that the data ranges of all Y= columns in the row cells are merged into a
common axis range. This facilitates the visual comparison of the cells.
n A SIDEBAR block in the LAYOUT LATTICE statement block creates a sidebar at
the top of the lattice. A DISCRETELEGEND statement populates the sidebar
with a graph legend.
n The DISPLAYSECONDARY= option in the ROWAXIS statement displays a
secondary Y axis on the right. The secondary Y axis is a duplicate of the primary
Y axis, which makes it easier to read Y-axis values when viewing the bar chart in
the right cell.

Example Graph
The following graph was generated by the “Example Program” on page 158:
 158 Chapter 4 / Layout Statements

Example Program
proc template;
define statgraph layoutlattice;
begingraph;
entrytitle "Vehicle Gas Mileage";
entryfootnote "Averages of 428 models from 38 manufactures";
layout lattice / columns=2 rowdatarange=union;
layout overlay / cycleattrs=true;
barchart x=origin y=mpg_highway / stat=mean barwidth=.8
name="H" ;
barchart x=origin y=mpg_city / stat=mean barwidth=.5
name="C" ;
endlayout;
layout overlay / cycleattrs=true;
barchart x=type y=mpg_highway / stat=mean barwidth=.8;
barchart x=type y=mpg_city / stat=mean barwidth=.5;
endlayout;
sidebar / align=top;
discretelegend "H" "C" / border=false;
endsidebar;
rowaxes;
rowaxis / display=(tickvalues) displaysecondary=(tickvalues)
griddisplay=on;
endrowaxes;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars template=layoutlattice;
 LAYOUT OVERLAY Statement 159

run;

LAYOUT OVERLAY Statement

Builds a composite from one or more GTL-statements. The composite could be an entire graph. Or, if
this layout is nested in a GRIDDED or LATTICE layout, then the composite typically provides
contents for one cell in the parent layout.

Restrictions: You can add one or more 2-D plots to the graph area that the LAYOUT OVERLAY
statement creates, but all of the graphs will share the same set of axes.
3-D plots are not allowed.
Interaction: When nested within another layout type, the OVERLAY layout defines the graph
display for one cell of the parent layout. A separate OVERLAY layout is specified
for each cell.
Tips: If you want equal sized units on the X and Y axes, use the LAYOUT
OVERLAYEQUATED statement.
By default, the first plot specified in the LAYOUT OVERLAY block defines the axis
characteristics. If you want another plot to define the axis characteristics, specify
PRIMARY=TRUE in the statement for that plot.
Overlaid plots are stacked in the order in which they are specified. The last plot
specified is placed on top. When plots are stacked, a plots’s graphical data might
obscure the graphical data of the plots beneath it. In that case, you can change the
plot order or use transparency to make all of the graphical data visible.
The LAYOUT OVERLAY block is primarily used to specify 2-D plot statements, but
you can also specify other statements such as DISCRETELEGEND,
CONTINUOUSLEGEND, and ENTRY.

Syntax
LAYOUT OVERLAY </options>;
GTL-statements;
<INNERMARGIN </options(s)>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;>
<… more-innermargin-blocks …> >
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the plot’s wall area.
160 Chapter 4 / Layout Statements

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills
in nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.

Axes options
X2AXISOPTS=(axis-options)
specifies one or more X2 axis options.
XAXISOPTS=(axis-options)
specifies one or more X-axis options.
Y2AXISOPTS=(axis-options)
specifies one or more Y2 axis options.
YAXISOPTS=(axis-options)
specifies one or more Y-axis options.

Optional Arguments
ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the plot’s wall area. The ratio is expressed as a
positive decimal fraction representing wall-height divided by wall-width. For
example, 0.75 is a 3/4 aspect ratio and 1.0 is a square aspect ratio.

Default AUTO. The wall area is sized to the maximum area that can fill the
available space inside the OVERLAY layout.

Interaction When the LAYOUT OVERLAY statement is nested in a LAYOUT

LATTICE block, the ASPECTRATIO= option is ignored unless
ROWDATARANGE=DATA, COLUMNDATARANGE=DATA,
ROW2DATARANGE=DATA, and COLUMN2DATARANGE=DATA are
in effect in the LAYOUT LATTICE statement.

See “LAYOUT LATTICE Statement” on page 126

 LAYOUT OVERLAY Statement 161

BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.

FALSE
does not cycle the default visual attributes of multiple plots.

For example, if you overlay three series plots, then each series line has the
same default visual properties.

TRUE
attempts to use the GraphData1–GraphDataN style elements to assign
different visual properties to applicable plots (scatter plots and series plots
and others).

Some plots in the layout do not participate in the cycling (for example,
reference lines and drop lines).
162 Chapter 4 / Layout Statements

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 210

“boolean ” on page xii for other Boolean values that you can use.

Example In the following example, the first three series plots are assigned line
properties that are based on the GraphData1, GraphData2, and
GraphData3 style elements. The fourth series plot does not
participate in the attribute cycling because its LINEATTRS= option
assigns a line style.
layout overlay / cycleattrs=true;
seriesplot x=date y=var1;
seriesplot x=date y=var2;
seriesplot x=date y=var3;
seriesplot x=date y=var4 / lineattrs=GraphReference;
endlayout;

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.
 LAYOUT OVERLAY Statement 163

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
164 Chapter 4 / Layout Statements

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style- attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.

STANDARD
displays a filled wall.

The setting of the FrameBorder attribute of the GraphWalls style element

determines whether the wall outline is displayed.

ALL
displays a filled, outlined wall.

NONE
displays no wall and no wall outline.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall fill and outline for occupied cells
are set by the GraphWalls style element. For empty cells, the outline
color is set by the ContrastColor attribute of the GraphBorderLines
style element, the line style is solid, and the line thickness is 1 pixel.

XAXISOPTS=(axis-options)
specifies one or more X-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.
 LAYOUT OVERLAY Statement 165

See “Axis Options for LAYOUT OVERLAY” on page 1055 for a list of
axis options.

X2AXISOPTS=(axis-options)
specifies one or more X2 axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 1055 for a list of
axis options.

YAXISOPTS=(axis-options)
specifies one or more Y-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 1055 for a list of
options.

Y2AXISOPTS=(axis-options)
specifies one or more Y2 axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY” on page 1055 for a list of
options.

Details
The LAYOUT OVERLAY statement builds a composite using one or more GTL-
statements. You can specify one or more two-dimensional plots within the layout,
provided all plots can share the same type of axes. You can also specify one or
more insets, such as nested layout statements (for example, LAYOUT GRIDDED),
ENTRY statements, and legend statements (for example, CONTINUOUSLEGEND
or DISCRETELEGEND).

The following general logic applies to rendering the composite:

166 Chapter 4 / Layout Statements

Note: The details for positioning insets also apply to insets that are specified
within a LAYOUT REGION block.

n All plot statements are rendered first. Plot statement results are always
rendered in the plot area. The plots are stacked on top of one another in the
order in which they are specified, with the last one on top. It is possible for one
plot’s graphical data to obscure graphical data beneath it. You can control this
by selectively ordering the plot statements, or by using transparency on the
individual plots, or by doing both.
n The insets are rendered next, in the order in which they are specified. As with
the plot statements, it is possible for the insets to obscure the results of other
statements in the layout.
n To control the horizontal and vertical positioning of some insets, you can use
the inset statement’s AUTOALIGN= option, or its HALIGN= and VALIGN=
options. Each nested inset determines its own relative position in the parent
OVERLAY. This positioning achieves the best results for text-based insets
whose size can be easily fit within an open area of the graph wall. A large text-
based inset might not fit well, and an inset that contains a plot might be
dropped from the display without warning when the template is executed.
n Some insets, like legends, can be positioned inside or outside of the plot area
using the inset statement’s LOCATION= option. The inset’s AUTOALIGN= or
HALIGN= and VALIGN= settings are then relative to that location.

Generally, the first specified plot determines the layout’s default axis
characteristics. To enable another plot to define the axis characteristics, set
PRIMARY=TRUE for that plot. For more information about the default axis
characteristics, see “When Plots Share Data and a Common Axis” on page 1046. See
also the LAYOUT OVERLAYEQUATED and the LAYOUT OVERLAY3D statements.

An OVERLAY layout can also contain an inner margin, which is a nested region at
the top or bottom of the OVERLAY container. One or more inner margin plots can
be specified, and each is specified within an INNERMARGIN block. Within the
INNERMARGIN block, only one-dimensional plot statements such as BLOCKPLOT
and AXISTABLE can be specified. See “INNERMARGIN Statement” on page 193.

Example: Simple Overlay

This example shows how to create a simple overlay of two series plots using the
OVERLAY layout. The following figure shows the output.
 LAYOUT OVERLAY Statement 167

Here is the SAS code.

data workers;
format Date monyy5.;
input Date monyy5. Electric Masonry;
datalines;
JAN80 320.3 293.8
FEB80 315.7 285.8
MAR80 312.6 292
APR80 306.5 299.3
MAY80 308.6 301.7
JUN80 316.3 307.9
JUL80 319.5 310.7
AUG80 326.4 314.9
SEP80 330.8 312.7
OCT80 329.3 318.5
NOV80 330.6 307.7
DEC80 327.2 296.2
JAN81 316.2 259.2
FEB81 310.1 258.8
MAR81 308.5 271.5
APR81 311.1 281
MAY81 313.6 283.7
JUN81 318.3 289.3
JUL81 321.3 291.1
AUG81 327.4 295.9
SEP81 326.7 292.7
OCT81 326.4 282.6
NOV81 322.5 275.5
DEC81 318.6 260.2
JAN82 301.9 214.3
FEB82 296.1 224.8
MAR82 298.3 228.7
APR82 297.7 244.7
 168 Chapter 4 / Layout Statements

MAY82 303.5 260.4

JUN82 305 262.2
JUL82 307.6 270.4
;
proc template;
define statgraph layoutoverlay;
begingraph;
entrytitle "Trends in Employment Levels";
layout overlay / cycleattrs=true
xaxisopts=(display=(ticks tickvalues))
yaxisopts=(label="Number of Workers (thousands)");
seriesplot x=date y=electric /
curvelabel="Electrical"
curvelabellocation=outside;
seriesplot x=date y=masonry / curvelabel="Masonry"
curvelabellocation=outside;
endlayout;
endgraph;
end;
run;

proc sgrender data=workers template=layoutoverlay;

run;

LAYOUT OVERLAYEQUATED Statement

Builds a composite from one or more GTL-statements. The composite could be an entire graph. Or, if
this layout is nested in another layout, such as a GRIDDED layout, then the composite typically
provides contents for one cell in the parent layout. In an OVERLAYEQUATED layout, the display unit
of the X axis always equals the display unit of the Y axis.

Restrictions: All overlaid plots share common X and Y axes.

3-D plots are not allowed.
The axis type for both X and Y axes is always linear and is derived directly from the
data. Plot types that have discrete or binned axes cannot be used with this layout.
This includes BOXPLOT, BOXPLOTPARM, BARCHARTPARM, HISTOGRAM and
HISTOGRAMPARM. It also includes HEATMAP and HEATMAPPARM when at least
one of the plot variables is of type character, or TRUE is in effect for at least one of
the options DISCRETEX=, DISCRETEY=, XBINAXIS=, and YBINAXIS=.
You can add one or more of the following X-Y plots to the graph area that the
LAYOUT OVERLAYEQUATED statement creates: BANDPLOT,
CONTOURPLOTPARM, ELLIPSE, ELLIPSEPARM, LOESSPLOT, NEEDLEPLOT,
PBSPLINEPLOT, REGRESSIONPLOT, SCATTERPLOT, SERIESPLOT, STEPPLOT, or
VECTORPLOT. You can also add HEATMAP and HEATMAPPARM plots when both
plot variables are numeric, and FALSE is in effect for options DISCRETEX=,
DISCRETEY=, XBINAXIS=, and YBINAXIS. As long as one of these plots is present,
you can also add FRINGEPLOT, LINEPARM, MODELBAND, REFERENCELINE,
DROPLINE, DISCRETELEGEND, CONTINUOUSLEGEND, and text-based
statements such as ENTRY.
 LAYOUT OVERLAYEQUATED Statement 169

This layout has only two independent axes from a data standpoint, X and Y. If any
contained plot uses an X=X2 or Y=Y2 option, then the option is ignored and the
data is mapped to the X or Y axis. However, the X2 and Y2 axes can be displayed
using the DISPLAYSECONDARY= suboption of the XAXISOPTS= and YAXISOPTS=
options.
Interaction: When nested within another layout type, the OVERLAYEQUATED layout defines
the graph display for one cell of the parent layout. A separate OVERLAYEQUATED
layout is specified for each cell.
Tips: By default, the first plot specified in the LAYOUT OVERLAYEQUATED block
defines the axis characteristics. If you want another plot to define the axis
characteristics, specify PRIMARY=TRUE in the statement for that plot.
Overlaid plots are stacked in the order in which they are specified. The last plot
specified is placed on top. When plots are stacked, a plots’s graphical data might
obscure the graphical data of the plots beneath it. In that case, you can change the
plot order or use transparency to make all of the graphical data visible.

Syntax
LAYOUT OVERLAYEQUATED </options>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills
in nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
170 Chapter 4 / Layout Statements

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.

Axes options
COMMONAXISOPTS=(common-axis-options)
specifies one or more options to apply to both the X and Y equated axes.
EQUATETYPE=FIT | SQUARE | SQUAREDATA | EQUATE
specifies how to draw the axis area.
XAXISOPTS=(axis-options)
specifies one or more X-axis options.
YAXISOPTS=(axis-options)
specifies one or more Y-axis options.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

 LAYOUT OVERLAYEQUATED Statement 171

COMMONAXISOPTS=(common-axis-options)
specifies one or more options to apply to both the X and Y equated axes.

Requirements Axis options must be enclosed in parentheses.

Each option must be specified as a name = value pair and must be

separated by a space.

See “Options That Apply to Both Equated Axes” on page 1197 for a list
of options.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.

FALSE
does not cycle the default visual attributes of multiple plots.

For example, if you overlay three series plots, then each series line has the
same default visual properties.

TRUE
attempts to use the GraphData1–GraphDataN style elements to assign
different visual properties to applicable plots (scatter plots and series plots
and others).

Some plots in the layout do not participate in the cycling (for example,
reference lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 210

“boolean ” on page xii for other Boolean values that you can use.

EQUATETYPE=FIT | SQUARE | SQUAREDATA | EQUATE

specifies how to draw the axis area.
172 Chapter 4 / Layout Statements

FIT
specifies that the X and Y axes have equal increments between tick values.

The data ranges of both axes are compared to establish a common increment
size. The axes might be of different lengths and have a different number of
tick marks. Each axis represents its own data range. One axis might be
extended to use available space in the plot area. If a TICKVALUELIST= or
TICKVALUESEQUENCE= axis option is used on COMMONAXISOPTS= , then
it is ignored.

SQUARE
specifies that both the X and Y axes have the same length and the same
major tick values.

The axis length and tick values are chosen so that the minimum and
maximum of both X and Y appear in the range of values appearing on both
axes.

SQUAREDATA
specifies that both the X and Y axes have the same data range, but they can
have different tick values.

A UNION of the data ranges does not occur in this case. For example, if the
X-axis values are 20 to 40 (range of 20) and the Y-axis values are 200 to
260 (range of 60), then both axes have a range of 60 units, but the X axis can
have tick values 0, 20, 40, and 60, and the Y axis can have tick values 200,
220, 240, and 260.

EQUATE
same as FIT except that neither axis is extended to use available space in
the plot area.

Default FIT

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.
 LAYOUT OVERLAYEQUATED Statement 173

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0
174 Chapter 4 / Layout Statements

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style- attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.

STANDARD
displays a filled wall.

The setting of the FrameBorder attribute of the GraphWalls style element

determines whether the wall outline is displayed.

ALL
displays a filled, outlined wall.

NONE
displays no wall and no wall outline.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall fill and outline for occupied cells
are set by the GraphWalls style element. For empty cells, the outline
 LAYOUT OVERLAYEQUATED Statement 175

color is set by the ContrastColor attribute of the GraphBorderLines

style element, the line style is solid, and the line thickness is 1 pixel.

XAXISOPTS=(axis-options)
specifies one or more X-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Options That Apply Separately to an X or Y Equated Axis” on

page 1203 for a list of options.

YAXISOPTS=(axis-options)
specifies one or more Y-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Options That Apply Separately to an X or Y Equated Axis” on

page 1203 for a list of options..

Details
The LAYOUT OVERLAYEQUATED statement is similar to the LAYOUT OVERLAY
statement: it builds a composite using one or more GTL-statements. Similar to a
LAYOUT OVERLAY, you can specify one or more 2-D plots within the layout,
provided all plots can share the same type of axes. (Additional restrictions are
discussed in a moment.) You can also specify one or more insets.

As an overlay-type layout, OVERLAYEQUATED has the same behavioral

characteristics as an OVERLAY layout. It uses the same general logic for rendering
the composite (see “LAYOUT OVERLAY Statement” on page 159 for details), and its
default axis characteristics are generally determined by the first specified plot,
unless you use PRIMARY=TRUE on an alternative plot statement (see “When Plots
Share Data and a Common Axis” on page 1046).

OVERLAYEQUATED differs from OVERLAY in several ways. With

OVERLAYEQUATED,
n The axis type for both X and Y axes is always linear. Thus, plot types that have
discrete or binned axes cannot be used within this layout. This includes
BOXPLOT, BOXPLOTPARM, BARCHARTPARM, HISTOGRAM, and
HISTOGRAMPARM. It also includes HEATMAP and HEATMAPPARM when one
or both axes are binned.
176 Chapter 4 / Layout Statements

n For equal data intervals on both axes, the display distance is the same. For
example, an interval of 2 on the X axis maps to the same display distance as an
interval of 2 on the Y axis.
n The aspect ratio of the plot display equals the aspect ratio of the plot data. In
other words, a 45-degree slope in data is represented by a 45-degree slope in
the display. The EQUATETYPE= option determines how the axes are drawn.

The following figure illustrates how a series plot might map differently when
specified in an OVERLAYEQUATED layout versus an OVERLAY layout:

From a data standpoint, this layout has only two independent axes, X and Y. If any
plots within the layout block use an XAXIS=X2 or YAXIS=Y2 option, then the
option is ignored and the data are mapped to the X or Y axis. To display X2 and Y2
axes, use the DISPLAYSECONDARY= suboption of the XAXISOPTS= and
YAXISOPTS= options.

If an OVERLAYEQUATED statement is nested in a LATTICE layout, then some of

the LATTICE’s alignment and external axis features are not supported on the
OVERLAYEQUATED layout.

Example: LAYOUT OVERLAYEQUATED Statement

The following graph was generated by the “Example Program” on page 177:
 LAYOUT OVERLAYEQUATED Statement 177

Example Program
proc template;
define statgraph layoutoverlayequated;
begingraph;
entrytitle "Gas Mileage for GMC Models";
layout overlayequated / equatetype=fit;
referenceline y=16.2 /
curvelabel="City Average for Trucks/SUVs"
curvelabellocation=inside
curvelabelattrs=GraphReference;
referenceline x=20.6 /
curvelabel="Highway Average for Trucks/SUVs"
curvelabellocation=inside
curvelabelattrs=GraphReference;
scatterplot x=mpg_highway y=mpg_city /
datalabel=model;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars
template=layoutoverlayequated;
where make="GMC";
run;
 178 Chapter 4 / Layout Statements

LAYOUT OVERLAY3D Statement

Builds a 3-D composite from one or more GTL-statements. The composite could be an entire graph.
Or, if this layout is nested in a GRIDDED or LATTICE layout, then the composite typically provides
contents for one cell in the parent layout.

Restriction: You can add one or more 3-D plots to the graph area that the LAYOUT OVERLAY3D
statement creates, but all of the graphs will share the same set of axes.
Tips: By default, the first plot specified in the LAYOUT OVERLAY3D block defines the
axis characteristics. If you want another plot to define the axis characteristics,
specify PRIMARY=TRUE in the statement for that plot.
Overlaid plots are stacked in the order in which they are specified. The last plot
specified is placed on top. When plots are stacked, a plots’s graphical data might
obscure the graphical data of the plots beneath it. In that case, you can change the
plot order or use transparency to make all of the graphical data visible.
The LAYOUT OVERLAY3D block is primarily used to specify 3-D plot statements,
but you can also specify other statements such as CONTINUOUSLEGEND.

Syntax
LAYOUT OVERLAY3D </options>;
GTL-statements;
ENDLAYOUT;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.
BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the layout.
CUBE=TRUE | FALSE
specifies whether the layout displays the lines that indicate the complete
bounding cube of the axis planes.
CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills
in nested plot statements automatically change from plot to plot.
OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or
transparent (FALSE).
 LAYOUT OVERLAY3D Statement 179

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout
border.
ROTATE=number
specifies the angle of rotation.
TILT=number
specifies the angle of tilt in degrees.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.
ZOOM=positive-number
specifies a zoom factor.

Axes options
XAXISOPTS=(axis-options)
specifies one or more X-axis options.
YAXISOPTS=(axis-options)
specifies one or more Y-axis options.
ZAXISOPTS=(axis-options)
specifies one or more Z-axis options.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.
180 Chapter 4 / Layout Statements

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

CUBE=TRUE | FALSE
specifies whether the layout displays the lines that indicate the complete
bounding cube of the axis planes.

Default TRUE

Note The cube lines are displayed independently of the wall borders and axis
lines. Because some cube lines coincide with wall borders and axis
lines, it might appear that turning off wall borders or axis lines has no
effect when CUBE=TRUE.

Tip The color, thickness, and pattern of the cube lines are determined by
the GraphAxisLines style element.

See “boolean ” on page xii for other Boolean values that you can use.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.

FALSE
does not cycle the default visual attributes of multiple plots.

For example, if you overlay three series plots, then each series line has the
same default visual properties.

TRUE
attempts to use the GraphData1–GraphDataN style elements to assign
different visual properties to applicable plots (scatter plots and series plots
and others).

Some plots in the layout do not participate in the cycling (for example,
reference lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 210

“boolean ” on page xii for other Boolean values that you can use.
 LAYOUT OVERLAY3D Statement 181

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.
182 Chapter 4 / Layout Statements

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

ROTATE=number
specifies the angle of rotation. Rotation is measured in a clockwise direction
about a virtual axis parallel to the Z axis (vertical) and passing through the
center of the bounding cube. A counterclockwise rotation can be specified with
a negative value.

Default 54

TILT=number
specifies the angle of tilt in degrees. Tilt is measured in a clockwise direction
about a virtual axis parallel to the X axis (vertical) and passing through the
center of the bounding cube. A counterclockwise rotation can be specified with
a negative value.

Default 20

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
 LAYOUT OVERLAY3D Statement 183

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style- attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.

STANDARD
displays a filled wall.

The setting of the FrameBorder attribute of the GraphWalls style element

determines whether the wall outline is displayed.

ALL
displays a filled, outlined wall.

NONE
displays no wall and no wall outline.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall fill and outline for occupied cells
are set by the GraphWalls style element. For empty cells, the outline
color is set by the ContrastColor attribute of the GraphBorderLines
style element, the line style is solid, and the line thickness is 1 pixel.

See the CUBE= option.

XAXISOPTS=(axis-options)
specifies one or more X-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.
184 Chapter 4 / Layout Statements

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY3D” on page 1119 for a list of
axis options.

YAXISOPTS=(axis-options)
specifies one or more Y-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY3D” on page 1119 for a list of
axis options.

ZAXISOPTS=(axis-options)
specifies one or more Z-axis options.

Requirements Axis options must be enclosed in parentheses and separated by

spaces.

Each option must be specified as a name = value pair, and each

pair must be separated by a space.

See “Axis Options for LAYOUT OVERLAY3D” on page 1119 for a list of
axis options..

ZOOM=positive-number
specifies a zoom factor. Factors greater than 1 move closer to the bounding
cube, less than 1 move farther away

Default 1

Details
The LAYOUT OVERLAY3D statement builds a 3-D composite using one or more
GTL-statements. You can specify one or more 3-D plots within the layout, provided
all plots can share the same type of axes. You can also specify “annotations” (for
example, with one or more ENTRY statements or LAYOUT GRIDDED statements).
However, annotations in the OVERLAY3D layout are more likely to collide with
other graphics features than are annotations in other overlay-type layouts.

As an overlay-type layout, OVERLAY3D has the same behavioral characteristics as

an OVERLAY layout. It uses the same general logic for rendering the composite
(see “LAYOUT OVERLAY Statement” on page 159 for details), and its default axis
characteristics are generally determined by the first specified plot, unless you use
PRIMARY=TRUE on another plot statement (see “When Plots Share Data and a
Common Axis” on page 1046).
 LAYOUT OVERLAY3D Statement 185

Within an OVERLAY3D layout, a graph’s bounding cube can be tilted, rotated, and
zoomed to provide a different viewpoint. By default, the outline of the bounding
cube is displayed and the viewing rotation angle is 57 degrees, the tilt angle is 20
degrees, and the zoom factor is 1. See the CUBE= , ROTATE= , TILT= , and ZOOM=
options for information about how to change the viewpoint.

Example: LAYOUT OVERLAY3D Statement

The following graph was generated by the “Example Program” on page 185:

Example Program
proc template;
define statgraph layoutoverlay3d;
begingraph;
entrytitle "Density Plot of Height and Weight";
layout overlay3d / tilt=10 rotate=54
walldisplay=none cube=false;
surfaceplotparm x=height y=weight z=density /
surfacecolorgradient=density;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.gridded template=layoutoverlay3d;
run;
 186 Chapter 4 / Layout Statements

LAYOUT PROTOTYPE Statement

Builds a composite from one or more plot-statements. The composite is used as a prototype or
"rubber stamp" that repeats in each cell of a parent DATALATTICE or DATAPANEL layout.

Restrictions: You can specify only one LAYOUT PROTOTYPE block in a LAYOUT DATAPANEL or
LAYOUT DATALATTICE block. If you specify more than one, then only the last
prototype block specified is honored. The remaining prototype blocks are ignored.
Only the following plots can be used in a LAYOUT PROTOTYPE block: BANDPLOT,
BARCHART, BARCHARTPARM, BLOCKPLOT, BOXPLOTPARM,
COUNTOURPLOTPARM, DROPLINE, ELLIPSEPARM, FRINGEPLOT,
HEATMAPPARM, HISTOGRAMPARM, LINECHART, LINEPARM, NEEDLEPLOT,
REFERENCELINE, SCATTERPLOT, SERIESPLOT, STEPPLOT, and VECTORPLOT.
SCATTERPLOTMATRIX plots, 3-D plots, and region plots such as PIECHART or
MOSAICPLOTPARM cannot be used in the LAYOUT PROTOTYPE block.
ENTRY, DISCRETELEGEND, and CONTINUOUSLEGEND statements cannot be
used in the LAYOUT PROTOTYPE block.
A plot statement cannot be used if it contains a column defined with an EVAL
expression.
You can add one or more two-dimensional plots and one-dimensional plots to the
graph area that the LAYOUT PROTOTYPE statement creates, provided all of the
graphs can share the same axis type.
If you include a plot statement with a CURVELABEL= option (such as
SERIESPLOT), then only CURVELABELLOCATION=INSIDE is supported.
If you include a plot statement that supports a CLIP= option (such as LINEPARM or
ELLIPSEPARM), then the CLIP value is always set to TRUE.
Requirement: The LAYOUT PROTOTYPE statement must be nested in a LAYOUT DATAPANEL or
LAYOUT DATALATTICE block.
Note: Nesting an INNERMARGIN block in the LAYOUT PROTOTYPE statement is valid in
SAS 9.4M1 and later releases.
See: “LAYOUT DATAPANEL Statement” on page 79
“LAYOUT DATALATTICE Statement” on page 51

Syntax
LAYOUT PROTOTYPE </option(s)>;
plot-statements;
<INNERMARGIN </options(s)>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;>
<… more-innermargin-blocks …> >
ENDLAYOUT;
 LAYOUT PROTOTYPE Statement 187

Optional Arguments
ASPECTRATIO=AUTO | positive-number
specifies the aspect ratio of the prototype cell. The ratio is expressed as a
positive decimal fraction representing wall-height divided by wall-width. For
example, 0.75 is a 3/4 aspect ratio and 1.0 is a square aspect ratio.

Default AUTO. The prototype cell is sized to the maximum area that can fill the
available space inside the layout cell.

Note If AUTO is not used for the aspect ratio, then the entire DATALATTICE
or DATAPANEL grid is affected and changes shape.

CYCLEATTRS=TRUE | FALSE
specifies whether the default visual attributes of markers, lines, and fills in
nested plot statements automatically change from plot to plot.

FALSE
does not cycle the default visual attributes of multiple plots.

For example, if you overlay three series plots, then each series line has the
same default visual properties.

TRUE
attempts to use the GraphData1–GraphDataN style elements to assign
different visual properties to applicable plots (scatter plots and series plots
and others).

Some plots in the layout do not participate in the cycling (for example,
reference lines and drop lines).

Default FALSE

See “Rotating Visual Attributes for Each Plot in an Overlay” on page 210

“boolean ” on page xii for other Boolean values that you can use.

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style- attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.
188 Chapter 4 / Layout Statements

STANDARD
displays a filled wall.

The setting of the FrameBorder attribute of the GraphWalls style element

determines whether the wall outline is displayed.

ALL
displays a filled, outlined wall.

NONE
displays no wall and no wall outline.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE displays the wall outline.

FILL displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall fill and outline for occupied cells
are set by the GraphWalls style element. For empty cells, the outline
color is set by the ContrastColor attribute of the GraphBorderLines
style element, the line style is solid, and the line thickness is 1 pixel.

When the wall outline is suppressed, adjacent lines such as axis lines
and cell-header borders are still displayed. To suppress the axis lines,
use the appropriate display option for the axes. The cell-header borders
cannot be suppressed.

Details
The LAYOUT PROTOTYPE statement defines a plot prototype or “rubber stamp”
that repeats automatically. It assembles the results of nested GTL statements into
a single axis area. The plots are drawn in the order in which they are specified. The
results of the last statement are placed on top.

The plot-statements determine the graphical content of the cells in the parent
layout, based on the subsetting of the specified classification variables. For an
example, see “LAYOUT DATALATTICE Statement” on page 51 or “LAYOUT
DATAPANEL Statement” on page 79.

A PROTOTYPE layout is essentially a restricted OVERLAY layout with the same

general rules for overlaying plots. The main difference is that there are no axis
options available on the LAYOUT PROTOTYPE statement. Axis properties are set
with the ROWAXISOPTS= and COLUMNAXISOPTS= options of the parent
DATAPANEL or DATALATTICE statement.
 LAYOUT REGION Statement 189

In SAS 9.4 and later releases, a PROTOTYPE layout can also contain an inner
margin, which is a nested region at the top or bottom of the PROTOTYPE container.
One or more inner margin plots can be specified, and each is specified within an
INNERMARGIN block. Within the INNERMARGIN block, only one-dimensional plot
statements such as BLOCKPLOT and AXISTABLE can be specified. See
“INNERMARGIN Statement” on page 193.

LAYOUT REGION Statement

Creates the drawing area for a plot that does not use axes.

Restrictions: A LAYOUT REGION block cannot contain more than one plot.
A LAYOUT REGION block can contain a PIECHART or MOSAICPLOTPARM plot
only.

Syntax
LAYOUT REGION </options>;
GTL-statements;
ENDLAYOUT;

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the layout background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style attribute named COLOR or CONTRASTCOLOR is used.

color
specifies a color by name.

See color on page xii

Default The GraphBackground:Color style reference.

Interaction OPAQUE=TRUE must be in effect in order for the color to be seen.

By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is drawn around the layout.

Default FALSE

Interaction If this option is set to FALSE, then the BORDERATTRS= option is

ignored.
190 Chapter 4 / Layout Statements

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the layout.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

OPAQUE=TRUE | FALSE
specifies whether the layout background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the layout border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the layout border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

 LAYOUT REGION Statement 191

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the layout border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 0

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

Details
The REGION layout provides a container for plots that do not use axes. Within the
LAYOUT REGION block, you can specify a single plot statement of a type that
never uses axes, such as a PIECHART or MOSAICPLOTPARM. If multiple plot
 192 Chapter 4 / Layout Statements

statements are specified, then only the first one is honored. You can also specify
one or more insets, such as nested layout statements (for example, LAYOUT
GRIDDED), ENTRY statements, and legend statements (CONTINUOUSLEGEND or
DISCRETELEGEND). For example, you could specify a PIECHART statement with a
DISCRETELEGEND statement and an ENTRY statement. You can also nest one or
more layout blocks within the REGION layout. For example, you could nest a
LAYOUT GRIDDED statement that creates a small table of text.

When nested within another layout type, such as a GRIDDED or LATTICE layout,
the REGION layout defines the graphical display for one cell of the parent layout. A
separate REGION layout is specified for each cell.

Example: LAYOUT REGION Statement

The following graph was generated by the “Example Program” on page 192:

Example Program
proc template;
define statgraph layoutregion;
begingraph;
entrytitle "Average Weight by Age";
layout region;
piechart category=age response=weight /
stat=mean name="p"
datalabelcontent=(response) datalabellocation=outside;
discretelegend "p" / title="Age" across=2
 INNERMARGIN Statement 193

border=true halign=right valign=top;

endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=layoutregion;

run;

INNERMARGIN Statement
Provides a nested region in a LAYOUT OVERLAY or LAYOUT PROTOTYPE container in which a block
plot or axis table can be placed.

Restriction: This statement is valid in LAYOUT OVERLAY and LAYOUT PROTOTYPE blocks
only.
Notes: Two or more INNERMARGIN blocks that have the same alignment are stacked.
Multiple statements within an INNERMARGIN block are stacked.
For an X axis, the offsets on each end of the Y axis are increased to make room for
the inner margin plots. For a Y axis, the offsets on each end of the X axis are
increased to make room for the inner margin plots.

Syntax
INNERMARGIN < /options>;
block-plot-statement(s); | axis-table statement(s);
ENDINNERMARGIN;

Optional Arguments
ALIGN=TOP | BOTTOM | LEFT | RIGHT
specifies the alignment of the inner margin.

Default BOTTOM
194 Chapter 4 / Layout Statements

Restrictions For a block plot, only TOP and BOTTOM are valid. LEFT and RIGHT
are ignored.

For an axis table, LEFT and RIGHT can be used for a Y or Y2 axis.

Multiple statements within an INNERMARGIN block are stacked.

Note For an inner margin with ALIGN=TOP or ALIGN=BOTTOM, the

offsets on each end of the Y axis are increased to reserve space for
the inner margin plots. For an inner margin with ALIGN=LEFT or
ALIGN=RIGHT, the offsets on each end of the X axis are increased
to reserve space for the inner margin plots.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inner margin background.

style-reference
specifies a style reference in the form style-element:style-attribute. Only the
style attributes named COLOR or CONTRASTCOLOR are used.

color
specifies a color by name.

See color on page xii

Default The graph wall color. (See WALLCOLOR=.)

Interaction For this option to have any effect, the OPAQUE= option must be set
to TRUE.

Note The inner margin background is set to the wall color even when
WALLDISPLAY= NONE.

GUTTER=dimension
specifies the gap between stacked items in the inner margin.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default 0

Requirement The inner margin must contain two or more stacked items for this
option to have any effect.

Note The default units for dimension are pixels.

See “dimension” on page xiii

OPAQUE=TRUE | FALSE
specifies whether the inner margin's background is opaque.

TRUE
specifies that the background is opaque.

FALSE
specifies that the background is transparent.
 INNERMARGIN Statement 195

Default FALSE

Interaction When this option is FALSE, the BACKGROUNDCOLOR= option is

ignored.

Tip To prevent axis color bars and grid lines from passing through the
axis table, set OPAQUE=TRUE.

See “boolean ” on page xii for other Boolean values that you can use.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the inner-margin border.

Note: This feature applies to SAS 9.4M2 and to later releases.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 0

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 0

TOP=dimension
specifies the amount of extra space added to the top.

Default 5 px for the first inner margin adjacent to the bottom of the plot
area. Otherwise, 0.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 5 px for the first inner margin adjacent to the top of the plot area.
Otherwise, 0.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

196 Chapter 4 / Layout Statements

See “dimension” on page xiii

SEPARATOR=TRUE | FALSE
specifies whether a separating line is drawn between the inner margin and the
rest of the layout content.

Note: This option is valid starting in SAS 9.4M1.

Default FALSE

Tip Use the SEPARATORATTRS= option to specify the attributes of the

separating line.

See “Example: Overlay with an Inner Margin Plot” on page 197

“boolean ” on page xii for other Boolean values that you can use.

SEPARATORATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the inner margin separating line.

Note: This option is valid starting in SAS 9.4M1.

Default The graphAxisLines style element

Interaction This option is ignored when SEPARATOR=FALSE.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

“Example: Overlay with an Inner Margin Plot” on page 197

Details
An inner margin is a nested region in an OVERLAY container. You can specify one or
more inner margin plots. Specify each plot within an INNERMARGIN block. Within
an INNERMARGIN block, you can specify only the BLOCKPLOT and AXISTABLE
statements. See “Example: Overlay with an Inner Margin Plot” on page 197.
 INNERMARGIN Statement 197

Example: Overlay with an Inner Margin Plot

Example Overview
This example shows how to place a plot in an inner margin of an OVERLAY layout.
It creates a graph that shows monthly total sales for a specific year. A LINECHART
statement is used to draw the plot. The months are shown along the category axis,
and the total sales values are shown along the response axis. The Sashelp.Prdsale
data set is used as the data source. The tick marks on the category axis are
positioned between the midpoints to align with the beginning of each month.

A BLOCKPLOT statement in an INNERMARGIN block is used to display the

quarters above the category axis. The INNERMARGIN statement uses the default
alignment, so the inner margin is positioned at the bottom of the layout container,
beneath the line chart. Alternate shading is used in the block plot to show the block
boundaries. Because the tick marks are positioned between the midpoints, they
align with the block boundaries. The SEPARATOR= and SEPARATORATTRS=
options are used in the INNERMARGIN statement to specify a dark-red, two-pixel-
wide separator line between the inner margin and the rest of the graph.

Note: The INNERMARGIN statement SEPARATOR= and SEPARATORATTRS=

options are valid in SAS 9.4M1 and later releases.

Example Output
Here is the output for this example.
 198 Chapter 4 / Layout Statements

Example Program
Here is the SAS code.
/* Create a format for the quarters */
proc format;
value quartername 1="Quarter 1" 2="Quarter 2"
3="Quarter 3" 4="Quarter 4";
run;

/* Define the graph template */

proc template;
define statgraph innermargin;
dynamic year;
begingraph / subpixel=on;
entrytitle "Total Sales in " year;
layout overlay /
xaxisopts=(type=discrete
discreteopts=(ticktype=inbetween));
innermargin /
separator=true
separatorattrs=(color=darkred thickness=2px);
blockplot x=month block=quarter /
filltype=alternate
fillattrs=(color=cxd7d7d7)
altfillattrs=(color=cxf7f7f7)
display=(fill values) valuehalign=center;
endinnermargin;
linechart category=month response=actual /
smoothconnect=true;
endlayout;
endgraph;
end;
run;

/* Generate the graph */

proc sgrender data=sashelp.prdsale template=innermargin;
format quarter quartername.;
where year=1994;
dynamic year=1994;
run;
 199

PART 4

Plot Statements

Chapter 5
Key Concepts for Using Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Chapter 6
Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
200
 201

5
Key Concepts for Using Plots

Minimum Requirements to Generate a Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

ODS Graphics Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Display Attributes for Non-Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
Display Attributes for Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Rotating Visual Attributes for Each Plot in an Overlay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Remapping Groups for Grouped Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Interactions between Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Location and Position of Curve Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Curve Label Location Relative to the Plot Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Curve Label Position Relative to the Curve Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214

Minimum Requirements to Generate a

Plot
ODS graphics are generated by template definitions that determine a graph’s layout
and appearance and specify the variable roles to be represented in the graph
display. A graph can be rendered from a compiled template by associating the
template with a data source at run time.

The following SAS program shows the basic structure needed to meet the minimum
requirements for generating a plot using GTL:
proc template;
define statgraph minimumreq;
begingraph;
layout overlay;
scatterplot x=weight y=height;
endlayout;
endgraph;
 202 Chapter 5 / Key Concepts for Using Plots

end;
run;

proc sgrender data=sashelp.class template=minimumreq;

run;
n The DEFINE STATGRAPH statement on PROC TEMPLATE is required to open a
definition block for defining and naming a graphics template. The END
statement closes the template definition.
n A BEGINGRAPH statement block is required to define the outermost container
for the graph. The ENDGRAPH statement closes the block.
n At least one layout statement block is required for specifying the elements that
compose the graph. To generate a plot, the layout block must contain at least
one plot statement. The ENDLAYOUT statement closes the layout block.
n The PROC TEMPLATE statement must be run to compile the template and save
it in the template store (Sasuser.Templat by default).
n The PROC SGRENDER statement is required to produce a graph from a
compiled template. The DATA= option specifies a run-time data source to use,
and the TEMPLATE= option specifies the template to use. The input data source
must satisfy any restrictions that are imposed by the template. For example, it
must contain any variables that have been specified on the template’s GTL
statements.

ODS Graphics Environment

The ODS GRAPHICS statement manages the settings of the ODS Graphics
environment and is a statement that you will probably use frequently in your SAS
sessions. For example, the ODS GRAPHICS statement provides options that
control the physical aspects of your graphs, such as the image size and the name of
the image file that is created for the graph.

The default image size of 640 pixels by 480 pixels (4:3 aspect ratio) for ODS
Graphics is set in the SAS Registry. You can change the image size using the
WIDTH= option, or the HEIGHT= option, or both in the ODS GRAPHICS statement.
To name the output image file, use the IMAGENAME= option.

The following ODS GRAPHICS statement sets a 320-pixel width for the graph and
names the output image file modelfit:

ods graphics / width=320px

imagename="modelfit" reset;

proc sgrender data=sashelp.class template=modelfit;

run;

ods graphics off;

 Display Attributes 203

n The WIDTH= option sets the image width to 320 pixels. Because no HEIGHT=
option is used, SAS uses the design aspect ratio of the graph to compute the
appropriate height. (The width of 320px is half the default width, so SAS sets
the height to 240px, which is half the default height.)
n The IMAGENAME= option sets the name of the output image file to modelfit.
The RESET option ensures that each time the graph is produced, the previous
version of the image file is replaced. Otherwise, image names are incremented
(modelfit1, modelfit2, and so on) every time the graph is produced.

In general, it is good practice to specify only one sizing option without the other—
just the WIDTH= option or just the HEIGHT= option. That way SAS can maintain
the design aspect ratio of the graph, which might be important for many graphs. For
example, a graph that has multiple columns or a statistics table on the side needs a
wide aspect ratio. Specifying both width and height in such cases might produce
unpredictable results.

Note: Size settings in the ODS GRAPHICS statement affect all of the graphs that
are rendered in the SAS session, unless they are changed by another ODS
GRAPHICS statement. The size for a graph produced by an individual template can
be set with the DESIGNWIDTH= and DESIGNHEIGHT= options in the
BEGINGRAPH statement. Size settings in the ODS GRAPHICS statement override
size settings in the BEGINGRAPH statement and remain in effect unless they are
changed in another ODS GRAPHICS statement or ODS GRAPHICS are turned off.

For more information about using the ODS GRAPHICS statement in GTL, see SAS
Graph Template Language: User’s Guide. For a more complete discussion of the ODS
GRAPHICS statement, see “ODS GRAPHICS Statement” on page 1561.

Display Attributes

Overview
The display attributes for the lines, colors, marker symbols, and text used in a graph
are derived from the ODS style that is in effect when the graph is produced. These
display attributes might also be influenced by grouped data. To override default
display attributes, all GTL plot statements provide options that manage the graph’s
visual appearance. For example, a BOXPLOT statement provides an
OUTLIERATTRS= option that manages the visual appearance of outliers.

Two ways are generally available for modifying a graph’s display attributes:
n Change the ODS style that is in effect for the graph. “ODS Styles” on page 18
provides an overview of the use of styles in a graph. SAS Graph Template
Language: User’s Guide discusses the use of styles in more detail.
 204 Chapter 5 / Key Concepts for Using Plots

n Override default style settings using GTL statement options. Some examples
are given in the sections that follow.

Display Attributes for Non-Grouped Data

Appendix 3, “Display Attributes,” on page 1599 documents the attribute settings
that can be specified for the lines, data markers, text, or area fills in a plot. The
defaults for these attributes are defined on style elements, but you can use
attribute options on the plot statement to change the defaults.

For example, the LINEPARM statement provides a LINEATTRS= option that

specifies the color, line pattern, or line thickness of the plot line. For non-grouped
data, if you do not set a line pattern in your template, then the default line pattern
for the plot is obtained from the GraphDataDefault:LineStyle style reference.

To change the default line pattern, a PATTERN= suboption on LINEATTRS= is

available. The following table shows the most common line patterns available for
the PATTERN= suboption.

Table 5.1 Commonly Used Line Patterns

Pattern
Number Pattern Name Example

1 Solid

2 ShortDash

4 MediumDash

5 LongDash

8 MediumDashShortDash

14 DashDashDot

15 DashDotDot

20 Dash

26 LongDashShortDash

34 Dot

35 ThinDot

41 ShortDashDot
 Display Attributes 205

Pattern
Number Pattern Name Example

42 MediumDashDotDot

“Available Line Patterns” on page 1607 provides the complete list of line patterns
that can be used with GTL.

In the following template definition, the LINEPARM statement’s LINEATTRS=

option overrides the default line pattern by specifying PATTERN=DASH:
proc template;
define statgraph patternchange;
begingraph;
layout overlay;
scatterplot y=height x=weight;
lineparm x=0 y=intercept slope=slope /
lineattrs=(pattern=dash);
endlayout;
endgraph;
end;
run;

Other display options can be managed the same way. For example, the
SCATTERPLOT statement provides a MARKERATTRS= option that specifies the
color, size, symbol, and weight of the plot data markers. For non-grouped data, if
you do not set a marker symbol in your template, then the default marker symbol is
obtained from the GraphDataDefault:MarkerSymbol style reference.

To change the default marker symbol, a SYMBOL= suboption on MARKERATTRS=

is available. The following table shows the marker symbols available for the
SYMBOL= suboption.

Table 5.2 Supported Marker Symbols

Symbol Name Plot Symbol Symbol Name Plot Symbol

ArrowDown StarFilled

Asterisk Tack

Circle Tilde

CircleFilled Triangle

Diamond TriangleFilled

DiamondFilled TriangleDown
 206 Chapter 5 / Key Concepts for Using Plots

Symbol Name Plot Symbol Symbol Name Plot Symbol

GreaterThan TriangleDownFilled

Hash TriangleLeft

HomeDown TriangleLeftFilled

HomeDownFilled TriangleRight

IBeam TriangleRightFilled

LessThan Union

Plus X

Square Y

SquareFilled Z

Star

In the following template definition, the SCATTERPLOT statement’s

MARKERATTRS= option overrides the default marker symbol by specifying
SYMBOL=CIRCLEFILLED, which uses a filled circle to represent the data points.
proc template;
define statgraph symbolchange;
begingraph;
layout overlay;
scatterplot y=height x=weight /
markerattrs=(symbol=circlefilled);
endlayout;
endgraph;
end;
run;

Display Attributes for Grouped Data

Appendix 3, “Display Attributes,” on page 1599 documents the attribute settings
that you can specify for the lines, data markers, text, or area fills in a plot. For a
grouped plot (that is, when you use the GROUP= option in the plot statement),
each distinct group value can be represented in the graph by a different
combination of line pattern, color, and marker symbol (depending on the graph
 Display Attributes 207

type). The defaults for these features are set by the LineStyle, Color, ContrastColor,
FillPattern, and MarkerSymbol attributes of the GraphData1–GraphDataN style
elements. The MarkerSize and LineThickness style attributes are not honored in the
case of grouped data.Table 5.1 on page 204 shows the common line patterns that
are available, and Table 5.2 on page 205 shows the marker symbols that are
available.

When missing group values are displayed, the default attributes of the missing
value are set by the GraphMissing style element unless the MISSING= system
option specifies a character other than "." or " ". In that case, missing group value
attributes are determined by the GraphData1–GraphDataN style elements.

For grouped plots, attributes such as colors, line patterns, and marker symbols are
used to distinguish the individual group values. The attributes are derived from the
GraphData1–GraphDataN style elements in the current style. Each group value is
assigned attributes from a GraphData1-GraphDataN style element sequentially (1
to N). When plotting a SAS data set, the group-value attributes are assigned in data
order. When plotting a CAS in-memory table, the group-value attributes are
assigned in ascending order of the group column character or unformatted numeric
values. For information about attribute rotation, see “Attribute Rotation Patterns”
in SAS Graph Template Language: User’s Guide.

As discussed in “Rotating Visual Attributes for Each Plot in an Overlay” on page

210, plot options might also influence the attribute rotation pattern. For example,
in the following template definition, the LINEPARM statement’s LINEATTRS=
option specifies PATTERN=DASH. This explicit setting overrides the default line
pattern for the plot lines and uses dashed lines for all of the plots, leaving color to
distinguish among group values.
proc template;
define statgraph dashedline;
begingraph;
layout overlay;
scatterplot y=height x=weight / group=gender;
lineparm x=0 y=intercept slope=slope / group=gender
lineattrs=(pattern=dash);
endlayout;
endgraph;
end;
run;

Rather than setting the same line pattern on all group values, you can change the
default sequence of line patterns that is used for grouped values. To do so, set the
LineStyle attribute in some of the style elements GraphData1–GraphDataN.

In the following example, a style is defined to change the line pattern for
GraphData1 and GraphData2. In this example, the style is derived from the
DEFAULT style. Values are set for the LineStyle attributes in the GraphData1 and
GraphData2 style elements. The first default line in the sequence has long dashes
(style value 6) and the second line has short dashes (style value 4). The LineStyle
settings for the remaining GraphData elements are not set, so are derived from the
parent style (DEFAULT). This new line sequence is used as the default line
sequence for any plot that uses the MyDefault style. To apply the style to a graph,
the STYLE= option is used in the ODS HTML statement to specify the style name.

Here is the code for this example.

208 Chapter 5 / Key Concepts for Using Plots

/* Specify a path for the ODS output */

filename odsout "output-path";

/* Sort the SASHELP.CLASS data by sex and age. */

proc sort data=sashelp.class(keep=height weight sex age)
out=class;
by sex age;
run;

/* Generate slope and intercept data for plot reference lines. */

proc robustreg data=class method=m
plots=none
outest=stats(rename=(weight=slope));
by sex;
model height=weight;
run;

data class;
merge class stats(keep=intercept slope sex);
run;

proc template;
/* Create custom style MYDEFAULT from the STYLES.DEFAULT style. */
define style MyDefault;
parent=Styles.Default;
style GraphData1 from GraphData1 /
LineStyle=6;
style GraphData2 from GraphData2 /
LineStyle=4;
end;

/* Create the plot template. */

define statgraph testPattern;
begingraph;
layout overlay;
scatterplot y=height x=weight / group=sex;
lineparm x=0 y=intercept slope=slope / group=sex
name="lines";
discretelegend "lines";
endlayout;
endgraph;
end;
run;

/* Generate the plot. */

ods _all_ close;
ods html path=odsout file="mydefaultstyle.html"
style=MyDefault; /* Apply style MyDefault to the graph. */

proc sgrender data=class template=testPattern;

run;

ods html close;

ods html; /* Not required in SAS Studio */

Here is the output.

 Display Attributes 209

Similarly, for grouped data, you can set the MarkerSymbol attribute in each of the
style elements GraphData1–GraphDataN. In the following example, a style is
defined to change the MarkerSymbol attribute for GraphData1–GraphData3. This
new sequence is used as the default marker symbol sequence for any grouped plot
that uses the MyDefault style.

Here is the code for this example.

Note: The data that was generated in the previous example is used again in this
example.

/* Specify a path for the ODS output */

filename odsout "output-path";

proc template;
/* Create custom style MYDEFAULT from STYLES.DEFAULT. */
define style MyDefault;
parent=Styles.Default;
style GraphData1 from GraphData1 /
MarkerSymbol="DIAMOND";
style GraphData2 from GraphData2 /
MarkerSymbol="CROSS";
style GraphData3 from GraphData3 /
MarkerSymbol="CIRCLE";
end;

/* Create the plot template. */

define statgraph testSymbols;
begingraph;
layout Overlay;
scatterPlot y=height x=weight / group=age name="symbols";
discretelegend "symbols" / title="Age";
endlayout;
endgraph;
end;
run;

/* Generate the plot. */

 210 Chapter 5 / Key Concepts for Using Plots

ods _all_ close;

ods html path=odsout file="mydefaultstyle.html"
style=MyDefault; /* Apply style MyDefault to the graph. */

proc sgrender data=class template=testSymbols;

run;

Rotating Visual Attributes for Each Plot in an

Overlay
Overlay-type layouts provide the CYCLEATTRS= option, which specifies whether
the default visual attributes of lines, marker symbols, and area fills in nested plot
statements automatically change from plot to plot. When CYCLEATTRS=TRUE, all
applicable plot statements (SCATTERPLOT, SERIESPLOT, and others) are
sequentially assigned the next unused GraphDataN style element. (The sequence is
overridden for plot statements that have an explicit setting, either through a style
element assignment or option settings.) No plot retains its default (implicit) style
element.

In the following example, assuming ungrouped data and the default attribute
rotation pattern, the series plots are assigned line properties based on the
GraphData1, GraphData2, and GraphData3 style elements. The reference line uses
GraphReference, not GraphData4.
layout overlay / cycleattrs=true;
seriesplot x=date y=var1;
seriesplot x=date y=var2;
seriesplot x=date y=var3;
referenceline x=cutoff / lineattrs=GraphReference;
endlayout;

If one of the plots in this example uses grouped data, then the grouped plots also
participate in the default cycles. For example, if the second plot has three groups,
 Display Attributes 211

then it generates three plots, which are assigned line properties based on the
GraphData2, GraphData3, and GraphData4 style elements.

If the plot statement that uses grouped data also uses the INDEX= option to
manage the group values (see “Remapping Groups for Grouped Data” on page 211),
then the INDEX= option overrides the default behavior. In that case, the grouped
plots do not participate in the default cycling.

When one or more of the plots within the layout override the default cycling
behavior, the arrangement of the plots within the layout might affect the default
mapping of the GraphDataN elements to those statements that participate in the
default cycling.

Remapping Groups for Grouped Data

Indexing can be used to collapse the number of groups that are represented in a
graph. For example, if 10 groups are in the data, then indexes 1 and 2 can be
assigned to the first two groups, and index 3 can be assigned to all other groups.
The third through tenth data groups are treated as a single group in the graph.

Indexing can control the order in which colors, area fills, marker symbols, and line
styles are mapped to group values in a graph. This ordering method is needed only
for coordinating the data display of multiple graphs when the default mapping
would cause group values to be mismatched between graphs.

For example, consider two studies of three drugs, A, B, and C. If Study 1 uses all
three drugs, then the first combination of color and marker symbol is mapped to
Drug A. The second combination of color and marker symbol is mapped to Drug B,
and the third is mapped to Drug C. If Study 2 omits Drug A, then the first
combination of color and marker symbol is mapped to Drug B, and the second is
mapped to Drug C. If the two graphs are viewed together, then this default mapping
causes the group values to be mismatched. The visual attributes that represent
Drug A in the first graph represent Drug B in the second graph. Those that represent
Drug B in the first graph represent Drug C in the second group.

The GROUP= option mappings can be made consistent between the two graphs by
creating an index column for each study. For these example studies, the GROUP
and INDEX columns are the following:

Table 5.3 Study 1

Drug1 Index1

A 1

B 2

B 2

C 3
 212 Chapter 5 / Key Concepts for Using Plots

Table 5.4 Study 2

Drug2 Index2

B 2

C 3

C 3

If the graph for Study 1 specifies INDEX=INDEX1 and the graph for Study 2 specifies
INDEX=INDEX2, then the second combination of color and marker symbol is
mapped to Drug B in both graphs. The third combination of color and marker
symbol is mapped to Drug C in both graphs.

Interactions between Options

When you use GTL statement options to manage the graph display, interactions
between options might cause some option settings to be ignored. For example, an
ENTRYTITLE statement provides BORDER= and BORDERATTRS= options for
managing a border line around the graph title. Border attributes that are set on the
BORDERATTRS= option have no effect on the graph title unless the title border
line is displayed by setting BORDER=TRUE.

Similarly, if a BOXPLOT statement’s DISPLAY= option suppresses the display of

outliers in a box plot, then using the OUTLIERATTRS= option to set outlier
attributes has no effect. The OUTLIERATTRS= settings take effect only if
DISPLAY= enables the display of outliers.

The option interactions are not limited to options that simply manage visual
elements. For example, on a BOXPLOT, if the EXTREME= option extends the box
whiskers beyond the fences, then outliers are suppressed in the plot and options
that affect the outliers are ignored, if set.

The documentation for each GTL statement identifies the option interactions that
might occur on that statement.
 Location and Position of Curve Labels 213

Location and Position of Curve Labels

Overview
On plots that generate a curve line (a series plot or a density plot, for example), you
can specify a label for the curve line. You can also determine the label’s location in
the graph. For example, the SERIESPLOT statement provides the following options
for managing a curve label:
CURVELABEL
specifies a label for the curve line.
CURVELABELLOCATION
specifies the location of the curve line label relative to the plot area.
CURVELABELPOSITION
specifies the position of the label relative to the curve line.

Curve Label Location Relative to the Plot Area

By default, the label for a curve line is displayed inside the plot area. The following
figure shows the default location of the label for a series plot labeled “Curve Label”:

Depending on the shape of the curve line, its distribution of values, and the other
plot elements that must be displayed within the plot area, GTL might have to add
an offset (see “Adjusting Axis Offsets” on page 1053) to one of the plot’s axis lines to
provide enough room for the curve label. To prevent the offset of the axis line, you
 214 Chapter 5 / Key Concepts for Using Plots

can move the curve label outside of the plot area by specifying
CURVELABELLOCATION=OUTSIDE on the plot statement:

Regardless of whether the curve label is displayed inside or outside of the plot area,
you can use the CURVELABELPOSITION= option to adjust the label’s position
relative to the curve line.

Curve Label Position Relative to the Curve Line

Given a curve label’s location inside or outside of the plot area, a plot statement’s
CURVELABELPOSITION= option can adjust the label’s position relative to the
curve line. For example, the following positions are available for a series plot (for
some plots, START and END are not available):
AUTO
positions the curve label automatically near the end series line along unused
axes whenever possible (typically Y2 or X2) to avoid collision with tick values.
This position is used only when CURVELABELLOCATION=OUTSIDE.
MAX
forces the curve label to appear near maximum series values (typically, to the
right).
MIN
forces the curve label to appear near minimum series values (typically, to the
left).
START
forces the curve label to appear near the beginning of the curve. This position is
particularly useful when the curve line has a spiral shape. It is used only when
CURVELABELLOCATION=INSIDE.
END
forces the curve label to appear near the end of the curve. This position is
particularly useful when the curve line has a spiral shape. It is used only when
CURVELABELLOCATION=INSIDE.
 Location and Position of Curve Labels 215

When CURVELABELLOCATION=INSIDE, you can choose whether to position the

curve label near the START or END of the curve, or near the minimum data values
(MIN) or maximum data values (MAX). START and END use a different algorithm
than MIN and MAX. They are particularly useful for spiral-shaped curves whose
end points do not correlate with the minimum and maximum data values. In those
cases, START or END provides “better” label locations than MIN and MAX.

When CURVELABELLOCATION=OUTSIDE and CURVELABELPOSITION=AUTO, a

“good” position is automatically chosen to avoid collision with the axis information.

The following figure shows the different combinations of label locations and
positions:

n The minimum or maximum axis tick marks can be adjusted (see “Adjusting Axis
Offsets” on page 1053) so that the label can be placed inside the plot area.
Increasing label length decreases the area available for displaying plots.
n When CURVLABELLOCATION=OUTSIDE, you can set the
CURVELABELPOSITION to MIN or MAX, but the label might collide with the
axis ticks and tick values, unless you are aware of where the axes are
positioned.
216 Chapter 5 / Key Concepts for Using Plots
 217

6
Plot Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
AXISTABLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
BANDPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
BARCHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
BARCHARTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
BIHISTOGRAM3DPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
BLOCKPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
BOXPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
BOXPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
BUBBLEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
CONTOURPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
DENDROGRAM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
DENSITYPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
DROPLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496
ELLIPSE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
ELLIPSEPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514
FRINGEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
HEATMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
HEATMAPPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
HIGHLOWPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
HISTOGRAM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 588
HISTOGRAMPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
LINECHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622
LINEPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 646
LOESSPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
MODELBAND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673
MOSAICPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685
NEEDLEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 699
PBSPLINEPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
PIECHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732
POLYGONPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
REFERENCELINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 778
REGRESSIONPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
SCATTERPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 806
SCATTERPLOTMATRIX Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 850
SERIESPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 878
STEPPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 925
 218 Chapter 6 / Plot Statements

SURFACEPLOTPARM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 958

TEXTPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 967
VECTORPLOT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 997
WATERFALLCHART Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016

Dictionary

AXISTABLE Statement
Creates an event plot of input data along an axis of an X-Y plot.

Syntax
AXISTABLE X=column | expression VALUE=column </options>;

AXISTABLE Y=column | expression VALUE=column </options>;

Summary of Optional Arguments

Appearance options
CLASS=column | expression
creates a separate row or column for each unique class value.
CLASSORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the order in which the class values are displayed.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
COLORGROUP=column | expression | discrete-attr-var
specifies a column that is used to discretely map the color of the value
text.
DATATRANSPARENCY=number
specifies the degree of the transparency of the header, label, and values.
DISPLAY=STANDARD | ALL | (display-options)
specifies which features to display.
DROPONMISSING=TRUE | FALSE
specifies whether the entire axis table is dropped when all of the VALUE=
column values are missing.
GUTTER=dimension
specifies the gap between rows when a class variable is used.
 AXISTABLE Statement 219

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing class values are represented in the table.
INDENT=dimension
specifies a value to be used with the INDENTWEIGHT= option to
determine the indention for each text value in a Y-axis table.
INDENTWEIGHT=numeric-column | expression
specifies the indention weight (multiplier) for each observation in a Y-
axis table.
PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the table border.
POSITION=number
positions the plot along the axis orthogonal to the axis used for the
values.
SHOWMISSING=TRUE | FALSE
specifies whether missing values are represented in the table.
TEXTGROUP=discrete-attr-var
specifies the discrete attribute variable for a discrete attribute map that
maps text attributes to values for each observation.
VALUEATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the text values.
VALUEFORMAT=format
specifies a SAS format or a user-defined format for the table values.
VALUEHALIGN=AUTO | LEFT | CENTER | RIGHT
in a Y-axis table, specifies the horizontal alignment of the column values
relative to the axis-table width.
VALUEJUSTIFY=AUTO | LEFT | CENTER | RIGHT
specifies the justification of the values in the axis table relative to the
axis-table width.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Header options
HEADERLABEL="string"
specifies the text for the table header.
HEADERLABELATTRS=style-element | style-element(text-options) | (text-
options)
specifies the color and font attributes of the table header.

Label options
LABEL="string"
specifies the text for the table label.
LABELATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the column label.
220 Chapter 6 / Plot Statements

LABELHALIGN=AUTO | LEFT | CENTER | RIGHT

specifies the horizontal alignment of the column label when it is
displayed.
LABELJUSTIFY=LEFT | CENTER | RIGHT
specifies the justification of the column label when it is displayed.
LABELPOSITION=MIN | MAX
specifies the end of the axis on which the label is displayed.

Midpoint options
CLASSDISPLAY=STACK | CLUSTER
specifies how the class values are displayed.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
STAT=AUTO | SUM | MEAN
specifies the statistic that is to be computed for the VALUE= column
when the column is numeric.

Title options
TITLE="string"
specifies the text for the table title.
TITLEATTRS=style-element | style-element(text-options) | (text-options)
specifies the color and font attributes of the table title.
TITLEHALIGN=AUTO | CENTER | LEFT | RIGHT
specifies the horizontal alignment of the title bounding box of the axis
table, relative to the axis table width.
TITLEJUSTIFY=LEFT | CENTER | RIGHT
specifies the justification of the lines of a multiline title in the title
bounding box. The justification is relative to the axis table width.

Required Arguments
Either the X= or the Y= argument must be specified in the AXISTABLE statement.
Specifying X= places an axis table along the X axis of a plot. Specifying Y= places an axis
table along the Y axis of a plot.
X=column | expression
specifies the column for the X axis.

Requirement If not specified, then Y= must be specified.

Y=column | expression
specifies the column for the Y axis.

Requirement If not specified, then X= must be specified.

VALUE=column
specifies the column that contains the axis table values.
 AXISTABLE Statement 221

Optional Arguments
CLASS=column | expression
creates a separate row or column for each unique class value. Each row or
column is labeled by the class value.

Interaction The DISPLAY= option that is in effect must include LABEL for any
labels to appear.

Note If you specify a column in a SAS data set, the visual attributes for
each class value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each class value
are assigned in ascending order of the class column character values
or of unformatted numeric values.

Tip Use the LABELATTRS= option to modify the label text attributes.

CLASSDISPLAY=STACK | CLUSTER
specifies how the class values are displayed.

Note: This feature applied to SAS 9.4M2 and to later releases.

STACK
displays the class values vertically at each midpoint value on the X axis or
horizontally on the Y axis.

CLUSTER
displays the class values horizontally at each midpoint value on the X axis or
vertically on the Y axis.

Restriction CLUSTER applies only when the axis table is on a discrete axis.

Tip The CLUSTERWIDTH= option controls the cluster width.

Default STACK

Interaction The CLASS= option must be specified for this option to have any
effect.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

CLASSORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the order in which the class values are displayed.

Note: This feature applies to SAS 9.4M2 and to later releases.

DATA
displays the class values in the order in which they occur in the data.
222 Chapter 6 / Plot Statements

REVERSEDATA
displays the class values in the reverse order from which they occur in the
data.

Tip This option is useful when the plot axis is reversed.

ASCENDING
DESCENDING
displays the class values in ascending or descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric class values or on formatted
values for character class values.

Interaction The CLASS= option must be specified for this option to have any
effect.

Note In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.

Note: This feature applies to SAS 9.4M2 and to later releases.

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Requirement For this option to take effect, the CLASS= option must also be
specified, and the CLASSDISPLAY= option must be set to
CLUSTER.

COLORGROUP=column | expression | discrete-attr-var

specifies a column that is used to discretely map the color of the value text.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.
 AXISTABLE Statement 223

Each unique value of this column is mapped to the Color attribute of the
GraphData1–GraphDataN style elements that are in effect. If a discrete attribute
variable is specified, the color mapping from its associated
DISCRETEATTRMAP statement is used.

Interaction This option is ignored when the TEXTGROUP= option is specified.

Note If you specify a column in a SAS data set, colors are mapped to
group values in data order. If you specify a column in a CAS in-
memory table, colors are mapped to group values in ascending order
of the group column character values or unformatted numeric
values.

See “DISCRETEATTRVAR Statement” on page 1516

“DISCRETEATTRMAP Statement” on page 1505

DATATRANSPARENCY=number
specifies the degree of the transparency of the header, label, and values.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISPLAY=STANDARD | ALL | (display-options)

specifies which features to display.

STANDARD
displays the table values and, if provided, the table label.

ALL
displays the same features as STANDARD.

(display-options)
specifies a space-separated list of display options, enclosed in parentheses.

The following options are supported:

LABEL
displays the table label.

The label can be the VALUE= column label or name, the LABEL= value, or
the CLASS= value for the table, depending on the options that you
specify.

VALUES
displays the column values.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip The column values are always displayed, even if DISPLAY=(LABEL) is

specified. To hide the table label, specify DISPLAY=(VALUES).

Default STANDARD
224 Chapter 6 / Plot Statements

Note If a table title is specified, it is always displayed.

DROPONMISSING=TRUE | FALSE
specifies whether the entire axis table is dropped when all of the VALUE=
column values are missing.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default FALSE

Tip The SHOWMISSING= option controls whether missing values are

shown in the table.

See VALUE= on page 220

“boolean ” on page xii for other Boolean values that you can use.

GUTTER=dimension
specifies the gap between rows when a class variable is used.

Note: This feature applies to SAS 9.4M2 and to later releases.

Defaults Y-axis table: 8 px

X-axis table: 0 px

Interaction The CLASS= option must be specified for this option to have any
effect.

See “dimension” on page xiii

HEADERLABEL="string"
specifies the text for the table header.

Note: Starting with SAS 9.4M2, the HEADERLABEL= option is deprecated and
is replaced with the TITLE= option. The syntax and functionality are the same.
The HEADERLABEL= option is still honored, but the TITLE= option is preferred.

Default No table header is displayed

Tip Use the HEADERLABELATTRS= option to control the appearance of

the table header.

HEADERLABELATTRS=style-element | style-element(text-options) | (text-

options)
specifies the color and font attributes of the table header.

Note: Starting with SAS 9.4M2, the HEADERLABELATTRS= option is

deprecated and is replaced with the TITLEATTRS= option. The syntax and
 AXISTABLE Statement 225

functionality are the same. The HEADERLABELATTRS= option is still honored,

but the TITLEATTRS= option is preferred.

See TITLEATTRS= on page 232

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing class values are represented in the table.

Note: This feature applies to SAS 9.4M2 and to later releases.

Missing class values are included by default. When the data contains missing
class values, the label for those values is either blank for missing character
values or a dot for missing numeric values.

The following figure shows block plots for classes Class 1, Class 2, and any
missing class values.

The following figure shows an X-axis axis table that displays values for classes
Class 1, Class 2, and any missing class values.

Notice that the label for the missing class values is blank. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing class values. If
you want to keep the missing class values, then you can create a format that
specifies a more meaningful label for the missing class. For example, here is a
format that specifies a label for missing character and numeric class values.
proc format;
value $missingClass " " = "(Missing)";
value missingClass . = "(Missing)";
run;

A single space enclosed in quotation marks specifies a missing character value

and a dot specifies a missing numeric value. Although it might seem appropriate
to use empty quotation marks ('' or "") to specify a missing character value,
doing so produces unexpected results. To specify a missing character value,
enclose a single space in quotation marks (' ' or " "). You can use this format for
the class columns in the PROC SGRENDER statement. In that case, if the class
columns contain missing values, then the labels specified in the format
statement are used for the missing classes.

The following figure shows the previous example when format $missingClass is
applied to the class variable.
226 Chapter 6 / Plot Statements

Note: In SAS 9.4M2 and in earlier releases, ODS Graphics does not support
Unicode values in user-defined formats. Starting with SAS 9.4M3, ODS Graphics
supports Unicode values in user-defined formats only if they are preceded by
the (*ESC*) escape sequence. Example: "(*ESC*){unicode beta}". ODS
Graphics does not support an escape character that is defined in an ODS
ESCAPECHAR statement in user-defined formats.

Default TRUE

Interaction The CLASS= option must be specified for this option to have any
effect.

See “boolean ” on page xii for other Boolean values that you can use.

INDENT=dimension
specifies a value to be used with the INDENTWEIGHT= option to determine the
indention for each text value in a Y-axis table.

Default 1/8 inch

Interaction The INDENTWEIGHT= option must be specified for this option to

have any effect.

Note This option is ignored for X-axis tables.

See “dimension” on page xiii

INDENTWEIGHT=numeric-column | expression
specifies the indention weight (multiplier) for each observation in a Y-axis table.

Interaction For each observation, the INDENT= option value is multiplied by the
value of the column specified by this option to determine the
indention for that observation’s value.

Notes This option is ignored for X-axis tables.

If the axis-table values are numeric and a user-defined format is

applied to the values, indentation might not work well, especially if
the format results in long strings.

Tip When using the INDENTWEIGHT= option, specify

VALUEHALIGN=LEFT to ensure correct indentation. Otherwise, the
indentation might not be as expected.

LABEL="string"
specifies the text for the table label.
 AXISTABLE Statement 227

Note: This feature applies to SAS 9.4M2 and to later releases.

Default The class values are used when the CLASS= option is set and
CLASSDISPLAY=STACK is in effect. Otherwise, the VALUE= column
label or name is used.

Interaction This option is ignored when the CLASS= option is in effect.

Note If the length of the label exceeds the available space, the label is
split on blank space as needed to make it fit.

See CLASS= on page 221

LABELATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the column label.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, the label color changes to match the group color
derived from the ContrastColor attribute of the GraphData1–
GraphDataN style elements.

Restriction Group behavior occurs only when the CLASS= and COLORGROUP=
option values are the same.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the properties that are not specified are derived from the
GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELHALIGN=AUTO | LEFT | CENTER | RIGHT

specifies the horizontal alignment of the column label when it is displayed.

Note: This feature applies to SAS 9.4M2 and to later releases.

The alignment is relative to the axis-table width. The axis table width is set by
the system to accommodate the maximum width of the values and, when
displayed, the label.

AUTO
uses the effective value of the LABELJUSTIFY= option.

LEFT
CENTER
RIGHT
horizontally justifies the label left, center, or right.

The following figure shows these positions for the axis-table label Size.
228 Chapter 6 / Plot Statements

Default AUTO

Restriction This option applies only to Y-axis tables.

Interaction The DISPLAY= option must include LABEL for this option to have
any effect.

LABELJUSTIFY=LEFT | CENTER | RIGHT

specifies the justification of the column label when it is displayed.

Note: This feature applies to SAS 9.4M2 and to later releases.

The justification is relative to the axis-table width. The axis table width is set by
the system to accommodate the maximum width of the values and, when
displayed, the label. The following figure shows the justifications LEFT,
CENTER, and RIGHT for axis-table label Nominal Size.

Default CENTER

Restriction This option applies only to Y-axis tables.

Interaction The DISPLAY= option must include LABEL for this option to have
any effect.

LABELPOSITION=MIN | MAX
specifies the end of the axis on which the label is displayed. The label is aligned
with the tick values on the axis. The following figure shows the MIN and MAX
positions for Y axis and X-axis tables.
 AXISTABLE Statement 229

Default MIN

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

PAD=dimension | (pad-options)
specifies the amount of extra space that is added inside the table border.

Note: This feature applies to SAS 9.4M2 and to later releases.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the table border.

(pad-options)
specifies a space-separated list of one or more name-value-pair pad options,
enclosed in parentheses.

The list can contain one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

Default 4 px

Restriction This option applies only to Y-axis tables.

RIGHT=dimension
specifies the amount of extra space added to the right side.

Default 4 px
230 Chapter 6 / Plot Statements

Restriction This option applies only to Y-axis tables.

TOP=dimension
specifies the amount of extra space added to the top.

Default 0 px

Restriction This option applies only to X-axis tables.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Default 0 px

Restriction This option applies only to X-axis tables.

Note Sides that are not assigned padding are padded with the default
amount of space.

Tip Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

POSITION=number
positions the plot along the axis orthogonal to the axis used for the values. This
option enables you to position the plot when the AXISTABLE statement is not
placed in an INNERMARGIN block.

number
specifies the position on the orthogonal axis as a fraction of the axis range.

When using the VALUEHALIGN option to align the values, in some cases, the
axis-table values might not align properly within the space allocated for the axis
table. For example, right-justifying the values might cause the values to overlap
other graphical elements. In that case, rather than allocating more space for the
axis table, you can use the POSITION= option to move the axis table to the left
along the X axis as shown in the following figure.

Default Determined by the system.

Range 0 (bottom)–1 (top)

 AXISTABLE Statement 231

Interaction This option is ignored when the AXISTABLE statement is placed in

an INNERMARGIN block. It is also ignored when the AXISTABLE
statement is placed in a LAYOUT OVERLAY block by itself.

Tip To reserve space for the plot at either end of the orthogonal axis,
specify the OFFSETMIN= or OFFSETMAX= option for the
orthogonal axis.

SHOWMISSING=TRUE | FALSE
specifies whether missing values are represented in the table.

Note: This feature applies to SAS 9.4M2 and to later releases.

The values are evaluated before the column format is applied. When this option
is set to FALSE, missing numeric and character values are hidden.

Default TRUE

See “boolean ” on page xii for other Boolean values that you can use.

STAT=AUTO | SUM | MEAN

specifies the statistic that is to be computed for the VALUE= column when the
column is numeric.

AUTO
computes the SUM statistic when the VALUE= column is numeric.

When the column is character, it uses the first column value as the statistic
value.

SUM
MEAN
computes the SUM or MEAN statistic when the VALUE= column is numeric.

When the column is character, it uses the first column value as the statistic
value.

Default AUTO

Interaction When the VALUE= column is character, the STAT= option always
uses the first column value as the statistic value. In that case, SUM
and MEAN are ignored.

TEXTGROUP=discrete-attr-var
specifies the discrete attribute variable for a discrete attribute map that maps
text attributes to values for each observation. The discrete attribute variable is
defined in a DISCRETEATTRVAR statement.

Restrictions A discrete attribute variable specification must be a direct

reference to the attribute variable. It cannot be set by a dynamic
variable.

The SIZE= specification in the discrete attribute map TEXTATTRS=

option is ignored.
232 Chapter 6 / Plot Statements

Interaction When this option is specified, the COLORGROUP= option is

ignored .

TITLE="string"
specifies the text for the table title.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default No table title is displayed

Note When an axis table is specified in the prototype of a data-driven layout,

if the table is on the X axis, then the table title appears only in the first
column of each row. If the table is on the Y axis, then the table title
appears only in the first row of each column.

Tip Use the TITLEATTRS= option to control the appearance of the table
title.

TITLEATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the table title.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the properties that are not specified are derived from the
GraphLabelText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Text Options” on page 1606 for available text-options.

TITLEHALIGN=AUTO | CENTER | LEFT | RIGHT

specifies the horizontal alignment of the title bounding box of the axis table,
relative to the axis table width.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
aligns the title bounding box based on the table type:
n For a Y-axis table, aligns the title bounding box according to the effective
TITLEJUSTIFY= option value.
n For an X-axis table, aligns the title bounding box LEFT.
 AXISTABLE Statement 233

CENTER
LEFT
RIGHT
horizontally aligns the table title bounding box center, left, or right.

Default AUTO

Restriction In SAS 9.4M2, this option applies only to Y-axis tables. Starting with
SAS 9.4M3, this option applies to Y-axis tables and to X-axis tables.

Interaction The TITLE= option must be specified for this option to have any
effect.

Tip For a multiline title in a Y-axis table, use TITLEJUSTIFY= to justify

the title text lines in the title bounding box.

TITLEJUSTIFY=LEFT | CENTER | RIGHT

specifies the justification of the lines of a multiline title in the title bounding
box. The justification is relative to the axis table width.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default LEFT

Restrictions This option applies only to Y-axis tables.

The title must wrap to multiple lines for this option to have any
effect.

Interaction The TITLE= option must be specified for this option to have any
effect.

Tip Use TITLEHALIGN= to align the title bounding box.

VALUEATTRS=style-element | style-element(text-options) | (text-options)

specifies the color and font attributes of the text values.

Default The GraphDataText style element.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the properties that are not specified are derived from the
GraphLabelText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALUEFORMAT=format
specifies a SAS format or a user-defined format for the table values.

Note: This feature applies to SAS 9.4M2 and to later releases.

234 Chapter 6 / Plot Statements

Default The format that is in effect for the column specified in the VALUE=
option. If no format is in effect, BEST6 is used for numeric columns.

Note Not all of the SAS formats are supported. See Appendix 4, “Format
Support in ODS Graphics,” on page 1621.

VALUEHALIGN=AUTO | LEFT | CENTER | RIGHT

in a Y-axis table, specifies the horizontal alignment of the column values
relative to the axis-table width. The axis table width is set by the system to
accommodate the maximum width of the values and, when displayed, the label.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
uses the effective value of the VALUEJUSTIFY= option.

LEFT
CENTER
RIGHT
aligns the values left, center, or right relative to the axis-table width.

The following figure shows an example of each of these alignments.

Default AUTO

Restriction This option applies only to Y-axis tables.

Tip If the values do not align properly within the space allocated for the
axis table, use the POSITION= on page 230 option to move the axis
table along the X axis as needed.

VALUEJUSTIFY=AUTO | LEFT | CENTER | RIGHT

specifies the justification of the values in the axis table relative to the axis-table
width. The axis table width is set by the system to accommodate the maximum
width of the values and, when displayed, the label.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
uses LEFT for text values or RIGHT for numeric values.
 AXISTABLE Statement 235

CENTER
LEFT
RIGHT
horizontally aligns the table values center, left, or right, relative to the axis-
table width.

The following figure shows an example of each of these alignments.

Default AUTO

Restriction This option applies only to Y-axis tables.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The AXISTABLE statement enables you to place text values along the X or Y axis at
specific values on the axis. It offers more flexibility than the BLOCKPLOT
statement, which is used to denote changes in block values along the axis. The X
and Y data does not need to be sorted.
236 Chapter 6 / Plot Statements

Example: AXISTABLE Statement

This example shows how to add a table of average sales data by division below a
bar chart of total sales by product and country. Here is the output that is generated
by this example.

An inner margin is created at the bottom of the layout container to reserve space
for the table. An AXISTABLE statement is used in the INNERMARGIN block to
show the average sales by division for each product. Here is the SAS code for this
example.
proc template;
define statgraph axistable;
begingraph;
entrytitle "Average Product Sales By Division and Country";
layout overlay / cycleattrs=true
yaxisopts=(offsetmax=0.15 label="Sales By Country");
innermargin / align=bottom opaque=true
backgroundcolor=cxf5f5f5;
axistable x=product value=actual /
name="division" stat=mean display=(label)
headerlabel="Sales By Division"
headerlabelattrs=GraphLabelText
valueattrs=(size=9pt weight=bold)
colorgroup=division class=division;
endinnermargin;
barchart category=product y=actual / name="country"
barlabel=true barlabelformat=dollar5.0
stat=mean group=country groupdisplay=cluster;
discretelegend "country" / title="Country:" location=inside
valign=top;
endlayout;
 BANDPLOT Statement 237

endgraph;
end;
run;

proc sgrender data=sashelp.prdsale template=axistable;

run;

See Also
“Creating an Axis-Aligned Inset with an Axis Table” in SAS Graph Template
Language: User’s Guide

BANDPLOT Statement
Creates a band plot that typically shows confidence or prediction limits.

Requirements: You must specify either an X= argument or a Y= argument. You cannot specify both.
When you specify the X argument, you must also specify LIMITLOWER and
LIMITUPPER arguments for Y values.
When you specify the Y argument, you must also specify LIMITLOWER and
LIMITUPPER arguments for X values.
The plot data should be sorted by the X or Y variable that is used in the BANDPLOT
statement. Otherwise, specify CONNECTORDER= AXIS in the BANDPLOT
statement.

Syntax
BANDPLOT X=column | expression
LIMITLOWER=number | numeric-column | expression
LIMITUPPER=number | numeric-column | expression < /options >;

BANDPLOT Y=numeric-column | expression

LIMITLOWER=number | numeric-column | expression
LIMITUPPER=number | numeric-column | expression < /option(s) >;

Summary of Optional Arguments

Appearance options
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
CONNECTORDER=VALUES | AXIS
specifies how to connect the data points to form the band lines.
238 Chapter 6 / Plot Statements

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the band labels.
DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and band outline.
DISPLAY=(display-options) | STANDARD | ALL
specifies which band features to display.
EXTEND=TRUE | FALSE
specifies whether the constant or "step" band is to be drawn to the area
bounded by the axes.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled band area.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
INDEX=positive-integer-column | expression
specifies indices for mapping band attributes (fill and outline) to one of
the GraphData1–GranphDataN style elements.
JUSTIFY=LEFT | CENTER | RIGHT
specifies the location of the data point relative to the step when
TYPE=STEP.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the band outlines.
TYPE=SERIES | STEP
specifies how the data points for lower and upper band boundaries are
interpolated.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over
the band plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
 BANDPLOT Statement 239

CURVELABELLOWER="string" | column
specifies a label for the lower band limit.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the band labels relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the band labels at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the band labels can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the band label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the band label
blocks.
CURVELABELUPPER="string" | column
specifies a label for the upper band limit.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate band plot for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

MODELNAME="plot-name"
specifies the name of the plot from which to derive the interpolation for
the band.
NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
You must specify either an X= or Y= argument. You cannot specify both. In addition, the
LIMITLOWER= and LIMITUPPER= arguments must be used to specify the lower and
upper lines for the band.
X=column | expression
specifies X values. Numeric or character values can be used.

Requirement You must also specify the LIMITLOWER= and the LIMITUPPER=
arguments for the Y values.

Y=column | expression
specifies Y values. Numeric or character values can be used.
240 Chapter 6 / Plot Statements

Requirement You must also specify the LIMITLOWER= and the LIMITUPPER=
arguments for the X values.

LIMITLOWER=number | numeric-column | expression

specifies a constant or column representing the values of the lower band line.

Interactions When this option is used with the X= option, it specifies the Y value
or values.

When this option is used with the Y= option, it represents the X

value or values.

Note If a constant is specified, then a straight line is drawn.

LIMITUPPER=number | numeric-column | expression

specifies a constant or column representing the values of the lower band line.

Interactions When this option is used with the X= option, it specifies the Y value
or values.

When this option is used with the Y= option, it represents the X

value or values.

Note If a constant is specified, then a straight line is drawn.

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the
ODS GRAPHICS statement.

OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS GRAPHICS
statement.

CONNECTORDER=VALUES | AXIS
specifies how to connect the data points to form the band lines.

VALUES
connects data points in the order read from the X column (or Y column).

Note In SAS 9.4M6 and earlier releases, unexpected results might occur
when VALUES is specified and data stored in a CAS in-memory table
is used. In that case, use AXIS instead. Starting with SAS Viya 3.5 and
 BANDPLOT Statement 241

SAS Studio 5.2, VALUES is ignored when data stored in a CAS in-
memory table is used.

AXIS
connects data points as they occur left-to-right along the X axis (or bottom-
to-top along the Y axis).

Tip You can use this value to ensure the expected connect order for certain
types of series lines (for example, time series) when the input data
might not be sorted by the X column (or Y column).

Defaults VALUES if the data is a SAS data set

AXIS if the data is a CAS in-memory table

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the band labels.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOWER="string" | column
specifies a label for the lower band limit.

Default No curve label is displayed for the lower band

Requirements For non-grouped data, use "string".

For grouped data, use a column to define the lower band labels
for each group value. All of the labels for a specific group value
must be the same. Otherwise, the results are unpredictable.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELUPPER="string" | column
specifies a label for the upper band limit.

Default No curve label is displayed for the upper band

Requirements For non-grouped data, use "string".

242 Chapter 6 / Plot Statements

For grouped data, use a column to define the upper band labels
for each group value. All of the labels for a specific group value
must be the same. Otherwise, the results are unpredictable.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the BANDPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE where
axes might be external to the grid.

Interactions This option is used with the CURVELABELPOSITION= option to

determine where the band labels appear. For more information, see
“Location and Position of Curve Labels” on page 213.

The EXTEND= option is ignored when

CURVELABELLOCATION=INSIDE is in effect.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the band labels relative to the curve line.

AUTO
positions the band labels automatically near the band boundary along
unused axes whenever possible (typically Y2 and X2).

Restriction This option is used only when

CURVELABELLOCATION=OUTSIDE.

MAX
forces the band labels to appear near maximum band values (maximum X-
values for horizontal curves, and maximum Y-values for vertical curves).

MIN
forces the band label to appear near minimum band values (minimum X-
values for horizontal curves, and minimum Y-values for vertical curves)

START
forces band labels to appear near the beginning of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.
 BANDPLOT Statement 243

END
forces band labels to appear near the end of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CURVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restrictions The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

The BANDPLOT statement does not support the BEGINGRAPH

statement option LABELPLACEMENT=SA on page 36.

Interaction This option is used with the CURVELABELLOCATION= option to

determine where the band labels appear. For more information, see
“Location and Position of Curve Labels” on page 213.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the band label might fall outside of the graph area. In
that case, the band label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the band labels at the specified split characters.
When a band label is split, the label is split on each occurrence of the specified
split characters.

Default FALSE. The band label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the band labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the band label. In that case, all of the specified split characters
together are treated as a single split character.
244 Chapter 6 / Plot Statements

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the band

labels are split unconditionally at each occurrence of any of the specified split
characters. If the band label does not contain any of the specified characters,
then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the band labels or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the band label blocks.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the band label text.

TRUE
drops the split characters from the band label text.

FALSE
includes the split characters in the band label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a band label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE
 BANDPLOT Statement 245

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the band for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the band labels.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the band label blocks.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the band for reference.
246 Chapter 6 / Plot Statements

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and band outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the band curve labels.

Tip The FILLATTRS= option can be used to set transparency for just the
band area. You can combine this option with FILLATTRS= to set one
transparency for the band outline but a different transparency for the
band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which band features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the band fill color

FILLPATTERN
displays the band fill pattern

Note This feature applies to SAS 9.4M5 and to later releases.

Tip For this option to have any effect, the active ODS style must
specify a fill pattern or a fill pattern must be specified with the
FILLPATTERNATTRS= option.

OUTLINE
displays the band outline

STANDARD
specifies FILL only
 BANDPLOT Statement 247

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Default The GraphBand:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the band outline and fill color. Starting with SAS 9.4M5,
use the FILLPATTERNATTRS= option to control the appearance of the
band fill pattern.

EXTEND=TRUE | FALSE
specifies whether the constant or "step" band is to be drawn to the area
bounded by the axes.

Default FALSE

Requirement When this option is used for a constant band, constants must be
specified for the upper and lower band limits. This requirement
does not apply to "step" bands.

Interactions This option is ignored when band labels are placed inside the plot
area (CURVELABELLOCATION=INSIDE). To extend the bands in
that case, use the CURVELABELLOCATION=OUTSIDE option.

If the X or Y value is character, then the EXTEND= option is

honored only when the upper and lower limits specify a number.

Tip If this option is not specified, then there can be a small gap
between the line and the axis. The gap is controlled by the axis
offset. If the axis offset is set to 0, then there is no gap.

See “boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled band area.

Defaults For non-grouped data, the GraphConfidence:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or by the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for both

the band fill and band outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the band
outline and a different transparency for the band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Fill Color Options” on page 1600 for available fill-options values.

248 Chapter 6 / Plot Statements

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

GROUP=column | discrete-attr-var | expression

creates a separate band plot for each unique group value of the specified
column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct band is created for each group value by varying the visual attributes
of the band display features. The display features are controlled by the current
ODS style or by the DISPLAY= option. The default group appearance for each
display feature is shown in the following table.

Display
Feature1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
 BANDPLOT Statement 249

Display
Feature1 Style Attributes That Control Default Group Appearance

The ContrastColor attribute controls the color of the fill pattern.

If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.

Outline ContrastColor and LineStyle attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.

1 The default display features are determined by the DisplayOpts attribute of the GraphBand
style element.
2 Fill patterns are valid for band plots starting with SAS 9.4M5.

Restrictions This option can be used only when a numeric column is specified
for the upper limit or the lower limit of the band plot. The other
limit could be a constant, if desired.

Starting with SAS 9.4M5, the number of vertices for a patterned

line cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000.
If the limit is exceeded, the plot is not drawn and a note is written
to the SAS log stating that the limit has been exceeded. In that
case, increase LINEPATTERNOBSMAX= to the value suggested in
the note or change the line pattern to SOLID.

Interactions To label grouped band plots, you must specify

CURVELABELLOWER= column and CURVELABELUPPER= column

The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of colors and
line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the band lines, but the
PATTERN= suboption of the OUTLINEATTRS= option could be
used to assign the same line pattern to all band outlines.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.
250 Chapter 6 / Plot Statements

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping band attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

JUSTIFY=LEFT | CENTER | RIGHT

specifies the location of the data point relative to the step when TYPE=STEP.
 BANDPLOT Statement 251

Default LEFT

Requirement TYPE= STEP must also be specified for this option to have any
effect.

Interaction If the MODELNAME= option is specified, then this option is

ignored.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MODELNAME="plot-name"
specifies the name of the plot from which to derive the interpolation for the
band. When this option is used, the band plot forms prediction or confidence
limits for the plot that supplies the fitted model.

Requirement plot-name must be the name that has been assigned on the
associated plot’s NAME= option.

Interaction This option overrides the JUSTIFY= and TYPE= options.

Tip If this option is not specified, then the interpolation is set by the
TYPE= option.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the band outlines.
252 Chapter 6 / Plot Statements

Defaults For non-grouped data, the GraphConfidence style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned line
cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000. If
the limit is exceeded, the plot is not drawn and a note is written to
the SAS log stating that the limit has been exceeded. In that case,
increase LINEPATTERNOBSMAX= to the value suggested in the
note or change the line pattern to SOLID.

Interaction For this option to have any effect, the outline must be enabled by
the ODS style or by the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Line Options” on page 1602 for available line-options values.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, LIMITUPPER, LIMITLOWER, GROUP,
CURVELABELUPPER, and CURVELABELLOWER.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the band
plot. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the band plot
can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique BANDPLOT and user-defined
roles. BANDPLOT roles include X , Y , LIMITUPPER , LIMITLOWER , GROUP ,
INDEX , CURVELABELUPPER , and CURVELABELLOWER .

User-defined roles are defined with the ROLENAME= option.

Note CURVELABELUPPER and CURVELABELLOWER are considered

roles only when they are assigned a column of values. They are not
 BANDPLOT Statement 253

considered roles and do not display data tips when assigned a

string.

Example This example displays data tips for the columns assigned to the
roles X, LIMITUPPER, and LIMITLOWER as well as the column
Obs, which is not assigned to any pre-defined BANDPLOT role. The
Obs column must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X LIMITUPPER LIMITLOWER)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X , Y , LIMITUPPER , LIMITLOWER , and
GROUP .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.
254 Chapter 6 / Plot Statements

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TYPE=SERIES | STEP
specifies how the data points for lower and upper band boundaries are
interpolated.

SERIES
connects the data points using line segments (as in a SeriesPlot).

STEP
connects the data points (as in a StepPlot).

Default SERIES

Interactions TYPE=STEP must be specified to enable the JUSTIFY= option.

If the MODELNAME= option is specified, then this option is

ignored.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interactions This option is ignored if the Y= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.
 BANDPLOT Statement 255

Details
A band plot can specify an X column with Y upper and lower limits, or a Y column
with X upper and lower limits. If you specify the X argument, then you must specify
LIMITLOWER and LIMITUPPER arguments for the Y values to apply the limits to
the Y axis. If you specify the Y argument, then you must specify LIMITLOWER and
LIMITUPPER arguments for the X values to apply the limits to the X axis.

When you use a BANDPLOT statement to display prediction or confidence limits,

the band plot can be used with another plot that specifies a fitted model. For
example, it can be used with a series or step plot. In these cases, use the
BANDPLOT option MODELNAME= or TYPE= to identify the interpolation for the
band.

You can use the BANDPLOT statement in displays that are independent of other
plots. For example, a band plot can be used to define yellow and green areas in an
OVERLAY LAYOUT statement that also contains a scatter plot. This use implies
concern for any of the scatter plot values that fall in the yellow area and comfort
for any values that fall in the green area. For this use, the upper and lower limits
would be specified by a constant.

Note: The BANDPLOT statement is optimized to work as a Confidence or

Prediction band. If the band is self-intersecting (not sorted for X or for Y), then the
resulting band is unpredictable. With unsorted data, the band that is generated for
an output Raster Image might not match the band that is generated for an output
Vector Graphic.

Example: BANDPLOT Statement

The following graph was generated by the “Example Program” on page 256:
 256 Chapter 6 / Plot Statements

Example Program
Here is the code for this example.
proc template;
define statgraph bandplot;
begingraph;
entrytitle "Fit Plot for Weight";
layout overlay;
bandplot x=height limitupper=uppermean
limitlower=lowermean /
name="band" modelname="fit"
legendlabel="95% Confidence Limits";
scatterplot x=height y=weight / primary=true;
seriesplot x=height y=predict / name="fit"
legendlabel="Fit Line";
discretelegend "fit" "band";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.classfit template=bandplot;

run;
 BARCHART Statement 257

BARCHART Statement
Creates a bar chart computed from input data.

Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In SAS
9.4M2 and in earlier releases, specify SUBPIXEL=ON in the BEGINGRAPH
statement to enable subpixel rendering. See SUBPIXEL= on page 39. Starting with
SAS 9.4M3, subpixel rendering is enabled by default.
To disable subpixel rendering in SAS 9.4M3 and in later releases, specify
SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 39. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” on page 1561.

Syntax
BARCHART CATEGORY=column | expression </option(s)>;

BARCHART CATEGORY=column | expression

RESPONSE=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
COLORBYFREQ=TRUE | FALSE
specifies whether the bar colors are based on statistical values when the
COLORRESPONSE= option is not specified.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option or the
COLORBYFREQ= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar
colors to a continuous color gradient.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
258 Chapter 6 / Plot Statements

specifies the appearance of the bar connect lines.

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline,
connect line, and bar labels, if displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
DISPLAYBASELINE=ON | OFF | AUTO
specifies whether the baseline is displayed.
DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bar area.
FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT
specifies the bar fill type.
INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the
interval width.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.
TARGET=numeric-column | expression
specifies the target value for each bar.

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
 BARCHART Statement 259

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
bar.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the
bar.
BARLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the text properties of the bar label text.
BARLABELFITPOLICY=AUTO | NONE | INSIDEPREFERRED
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
BARLABELFORMAT=format
specifies the text format used to display the bar label.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
SEGMENTLABEL=TRUE | FALSE
specifies whether a label is displayed inside each bar segment.
SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the text properties of the text for the bar segment label.
SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN
specifies a policy for fitting the bar segment labels within the bar
segments.
SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
GROUP100=NONE | MAGNITUDE | POSITIVE
displays the computed response values (FREQ, SUM, or MEAN),
normalized to 100%.
GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
260 Chapter 6 / Plot Statements

specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when the bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
COLORSTAT=FREQ | PCT | SUM | MEAN | PROPORTION
specifies the statistic to be calculated for the data range of the bar-color
gradient.
STAT=FREQ | PCT | SUM | MEAN | PROPORTION
specifies the statistic to be computed for the Y axis.

Required Arguments
Specifying only CATEGORY= creates a bar chart with bars that, by default, represent
frequency counts or percents of CATEGORY. Specifying both CATEGORY= and
RESPONSE= creates a bar chart with bars representing summarized values of RESPONSE
categorized by CATEGORY.
CATEGORY=column | expression
specifies the column or expression for the category values.

Notes You can use X= as an alternative to CATEGORY=. If you use X=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options recognize X
as the category role and not as CATEGORY.

For interval category values, if a user-defined format is applied to the

category column, the format should map each category value to only one
unique formatted value. Otherwise, unexpected results might occur.

RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Notes You can use Y= as an alternative to RESPONSE=. If you use Y=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will
recognize Y as the response role and not RESPONSE in that case.

This option is required only when you want summarized values of

RESPONSE that are categorized by CATEGORY.

Optional Arguments
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar. For
grouped clustered bars, each bar is labeled with the summarized value of the
 BARCHART Statement 261

bar. For grouped stacked bars, the segmented bar is labeled with the
accumulated, summarized value of all the bar segments.

Default FALSE

Note By default, the bar-label format is derived from the format that is
assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response-
column format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Tip The font and color attributes for the label are specified by the
BARLABELATTRS= option. The text format is specified by the
BARLABELFORMAT= option.

See “boolean ” on page xii for other Boolean values that you can use.

BARLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the bar label text.

Default The GraphDataText style element.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

BARLABELFITPOLICY=AUTO | NONE | INSIDEPREFERRED

specifies a policy for avoiding collisions among the bar labels when labels are
displayed.

AUTO
for a vertical bar chart, rotates the bar labels if the labels exceed the
midpoint spacing.

For a horizontal bar chart, always draws the labels horizontally. The
following figure shows an example.
262 Chapter 6 / Plot Statements

See the BARWIDTH= option for more information about the bar spacing.

NONE
does not rotate the bar labels.

Labels that are too long overlap.

INSIDEPREFERRED
attempts to place the bar labels in a horizontal bar chart inside the bars.

The following figure shows an example.

Note: This feature applies to SAS 9.4M5 and to later releases.

 BARCHART Statement 263

Restrictions This option applies to non-grouped or cluster-grouped

horizontal bar charts only.

Label splitting is not supported when a label is placed inside a

bar.

Requirement For a grouped horizontal bar chart, the

GROUPDISPLAY=CLUSTER option must be in effect.
Otherwise, the INSIDEPREFERRED policy is ignored.

Interaction If SEGMENTLABEL=TRUE is in effect, the INSIDEPREFERRED

policy is ignored, and the labels are placed outside.

Note If any bar label cannot fit inside the length of its bar, that label
is placed outside of its bar and is fit using the AUTO fit policy.

Labels can collide along their length and along their height. In some cases, if one
or more labels collide when the specified fit policy is used, then all of the labels
are dropped from the display. When that occurs, the following warning message
is written to the SAS log:
WARNING: The bar labels are suppressed. Use BARLABELFITPOLICY=NONE to
force the labels to be displayed.

TIP If the labels collide along their height, then using the
BARLABELATTRS= option to reduce the label font size might eliminate
the collision.

Default AUTO

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARLABELFORMAT=format
specifies the text format used to display the bar label.

Default The bar-label format is derived from the format that is applied to
the response column or from BEST6 if no format is assigned. The
derived format retains the precision of the response-column
format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

Note When a bar-label format is specified with this option, the bar
labels are formatted as specified by format. The specified format is
not automatically expanded to accommodate values that are too
wide.

Tip If you want the bar-label format to expand automatically for

summarized values on the response axis, specify the format for the
response column rather than in this option.
264 Chapter 6 / Plot Statements

BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interaction Starting with SAS 9.4M3, the INTERVALBARWIDTH= option

overrides this option for an interval bar chart.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number

of bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The ContrastColor, LineThickness, and LineStyle attributes of the

GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s Color,

LineStyle, and LineThickness attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. Prior to SAS 9.4M5, the
baseline is always displayed in the chart, whether for a specified value or for the
default value. Starting with SAS 9.4M5, the baseline display is controlled by the
DISPLAYBASELINE= option, which is ON by default. When the
BASELINEINTERCEPT= option is used, the axis range is adjusted to include the
baseline, and the baseline is placed at the specified value on the response axis.
 BARCHART Statement 265

Default 0

Interactions If GROUPDISPLAY=STACK is specified, then this option is ignored

and the baseline is not displayed.

This option is ignored when the GROUP100= option is used.

If necessary, the response axis data range is extended to include

the baseline intercept. When a logarithmic response axis is
requested and BASELINEINTERCEPT= specifies 0 or a negative
value, the response axis reverts to a linear axis. To restore the log
axis in that case, set BASELINEINTERCEPT= to a positive value.

When DISPLAYBASELINE=AUTO is in effect, the baseline is not

displayed if the baseline intercept is equal to the minimum or
maximum value of the response-axis range.

Note Label positions are automatically adjusted to prevent the labels

from overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline prior to SAS 9.4M5, use the

BASELINEATTRS= option to set the line thickness to 0. Starting
with SAS 9.4M5, specify DISPLAYBASELINE=OFF.

The baseline does not add a tick or a tick value to the axis. To label
the baseline, use a REFERENCELINE statement to overlay a line
with the same X or Y value and include the CURVELABEL= option
to specify the label text.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.

Default 0.85
266 Chapter 6 / Plot Statements

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Interaction When GROUPDISPLAY=CLUSTER, the default BARWIDTH is 1.0.

COLORBYFREQ=TRUE | FALSE
specifies whether the bar colors are based on statistical values when the
COLORRESPONSE= option is not specified. Setting this option to TRUE enables
you to color the bars based on frequency counts, percentages, or proportions.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default FALSE

Requirement The COLORSTAT= option must be FREQ, PCT, or PROPORTION

for this option to have any effect.

Interactions This option is ignored when the COLORRESPONSE= option is

specified.

When the GROUP= option is specified with the COLORBYFREQ=

option, the color attributes are controlled by the COLORBYFREQ=
option.

The COLOR= suboption of the FILLATTRS=,

FILLPATTERNATTRS=, and OUTLINEATTRS= options overrides
this option for the associated Color attribute.

Note This option is independent of the STAT= and RESPONSE= options.

Tips Use the COLORSTAT= option to specify whether frequency

counts, percentages, or proportions are computed for the
COLORRESPONSE= column.

Use the FILLTYPE= option to specify whether each bar is filled

with a solid color or with a gradient color.

See “Example 4: Bar Chart with Bar Colors Controlled by a Statistic” on

page 298

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option or the
COLORBYFREQ= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

 BARCHART Statement 267

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Defaults For outline-only bars, the ThreeColorAltRamp style element

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option or the
COLORBYFREQ=TRUE option must also be specified.

Tip Use the DISPLAY= option to specify whether outlines and fills are
displayed.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar colors to a
continuous color gradient.

Note: This feature applies to SAS 9.4M3 and to later releases.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. Each bar is colored using one color from the gradient
range. When a range attribute map variable is specified, the colors that are
defined in the associated range attribute map are used instead.
268 Chapter 6 / Plot Statements

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions The COLORBYFREQ= option is ignored when this option is

specified.

When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill, fill pattern, or both are displayed, this option overrides
suboption COLOR= in the FILLATTRS= option and in the
FILLPATTERNATTRS= option and varies the color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides

suboption COLOR= in the OUTLINEATTRS= option and varies the
outline color according to the color gradient or the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the COLORSTAT= option to specify the statistic to compute

for the COLORRESPONSE= column.

Use the FILLTYPE= option to specify whether each bar is filled

with a solid color or with a gradient color.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.

COLORSTAT=FREQ | PCT | SUM | MEAN | PROPORTION

specifies the statistic to be calculated for the data range of the bar-color
gradient.

Note: This feature applies to SAS 9.4M3 and to later releases.

The statistics that are available depend on the COLORRESPONSE= and

COLORBYFREQ= option specifications. When the COLORRESPONSE= option is
specified, the following values are valid:

SUM
MEAN

When the COLORRESPONSE= option is not specified and

COLORBYFREQ=TRUE is in effect, the following values are valid:

FREQ frequency count

PCT percentages between 0 and 100
PROPORTION proportions between 0 and 1
 BARCHART Statement 269

Defaults FREQ when the COLORRESPONSE= option is not specified and

COLORBYFREQ=TRUE is in effect.

SUM when the COLORRESPONSE= option is specified.

Interactions This option is ignored when the COLORRESPONSE= option is not

specified and COLORBYFREQ=FALSE is in effect.

This option might affect existing SAS programs. For programs

written before SAS 9.4M3, if STAT= and COLORRESPONSE= are
specified in a BARCHART statement, then the bar-chart colors and
color statistic might change from those of the previous SAS
releases. To restore the original colors and color statistic in that
case, set COLORSTAT= in the BARCHART statement to the same
statistic that is specified in STAT=.

Note This option is independent of the STAT= and RESPONSE= options.

See COLORBYFREQ= on page 266

COLORRESPONSE= on page 267

STAT= on page 287

“Example 4: Bar Chart with Bar Colors Controlled by a Statistic” on

page 298

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar connect lines.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement DISPLAY= must include CONNECT for this option to have any
effect.

Interaction This option is ignored when the GROUP= option is in effect.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
bars with each of the skins applied.
270 Chapter 6 / Plot Statements

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, connect line,
and bar labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
filled bar area. You can combine this option with FILLATTRS= to set
one transparency for the bar outlines and connect lines but a different
transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
 BARCHART Statement 271

DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.

Default 0 (no offset, all bars are centered on the category midpoints)

Range -0.5 to +0.5, where 0.5 represents half the distance between category
ticks. Normally, a positive offset is to the right when
ORIENT=VERTICAL, and up when ORIENT=HORIZONTAL. (If the
layout's axis options set REVERSE=TRUE, then the offset direction is
also reversed.)

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 292

“About the DISCRETEOFFSET= Option” on page 292 for information

about the REVERSE=, OFFSETMIN=, and OFFSETMAX= axis options

ORIENT=

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

CONNECT
displays the connect lines, which are line segments that connect adjacent
midpoints at the end of each bar.

FILL
displays the bar fill color

FILLPATTERN
displays the bar fill pattern

Tip For this option to have any effect, the active ODS style must specify
a fill pattern or a fill pattern must be specified with the
FILLPATTERNATTRS= option.

OUTLINE
displays the bar outline

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: CONNECT, FILL, FILLPATTERN, and OUTLINE

Default The GraphBar:DisplayOpts style reference.

272 Chapter 6 / Plot Statements

Interaction Connect lines are not drawn when the GROUP= option is in effect.

Note The connect lines are drawn in axis order starting with SAS 9.4M3.
They are drawn in data order in prior releases.

Tips Use the OUTLINEATTRS=, FILLATTRS=, and FILLPATTERNATTRS=

options to control the appearance of the bars.

Use CONNECTATTRS= to control the appearance of the connect

lines.

You can specify both FILL and FILLPATTERN to combine solid fills
and pattern fills in the bars.

DISPLAYBASELINE=ON | OFF | AUTO

specifies whether the baseline is displayed.

Note: This feature applies to SAS 9.4M5 and to later releases.

ON
always displays the baseline.

OFF
does not display the baseline.

AUTO
displays the baseline if the baseline intercept is within the response-axis
range, excluding the minimum and maximum axis values.

This is the typical case when the bar chart includes both positive and
negative bars, and the default baseline intercept of 0 is used. If the baseline
intercept is equal to the minimum or maximum value of the response-axis
range, the baseline is not displayed.

Default ON

Tip By default, the baseline intercept is 0. Use the BASELINEINTERCEPT=

option to change the baseline intercept.

DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.

Note: This feature applies to SAS 9.4M3 and to later releases.

A zero-length bar is displayed as a line spanning the normal bar width at the
bar-chart baseline on the response axis. When this option is set to TRUE, zero-
length bars are displayed. Otherwise, they are suppressed. The following figure
shows a simple example of each outcome. In the figure, the plot wall outline,
category axis line, and bar-chart baseline are suppressed for clarity.
 BARCHART Statement 273

Default TRUE

Interaction This option is ignored when the GROUP= and

GROUPDISPLAY=STACK options are in effect. In that case, zero-
length bar segments are drawn.

Note When this option is set to FALSE, the bar is not drawn, but other
elements associated with the bar such as the target bar, the error
bar, the bar label, and the data label, are drawn.

Tip This option is useful when the bar-chart baseline is suppressed.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bar area.

Defaults For non-grouped data, the GraphDataDefault:Color style reference

For grouped data, the Color attribute of the GraphData1–

GraphDataN style elements.

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is ignored,
and the bar fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

bar fills, bar outlines, and connect lines. You can combine this option
with DATATRANSPARENCY= to set one transparency for the bar
outlines and connect lines but a different transparency for the bar
fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphWall style element Color attribute.

Requirement FILLTYPE=COLORGRADIENT must be in effect. Otherwise, this

option is ignored.
274 Chapter 6 / Plot Statements

See “color ” on page xii

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Restriction In SAS 9.4M4 and in earlier releases, the only styles that are
delivered by SAS that support fill patterns are JOURNAL2,
JOURNAL3, and MONOCHROMEPRINTER. If style-element is
specified and the active ODS style does not support fill patterns,
this option is ignored.

Interactions For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILLPATTERN display, the FILLPATTERNATTRS=
suboption COLOR= is ignored, and the fill-pattern colors vary
according to the gradient.

Tip In order to change the fill pattern for each group value in ODS
styles such as HTMLBlue that use color-priority attribute rotation,
specify ATTRPRIORITY=NONE in an ODS GRAPHICS statement or
in the BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT

specifies the bar fill type.

Note: This feature applies to SAS 9.4M2 and to later releases.

SOLID
fills each bar with the color that is assigned to that bar.

ALPHAGRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This feature applies to SAS 9.4M5 and to later releases.

 BARCHART Statement 275

Tips The fill color is determined by a style element or by the FILLATTRS=

option COLOR= suboption. The initial transparency is determined by
the DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete attribute

map to set the initial transparency in the gradients for specific values.

COLORGRADIENT
fills each bar with a color gradient that starts with the fill color that is
assigned to that bar and ends with the color that is specified in the
FILLENDCOLOR= option.

Note: This feature applies to SAS 9.4M5 and to later releases.

Alias GRADIENT

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tips Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

See “FILLENDCOLOR=color” on page 273

GRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This option applies to SAS 9.4M2 through SAS 9.4M4. Starting with
SAS 9.4M5, ALPHAGRADIENT replaces GRADIENT, and GRADIENT is
changed to an alias of COLORGRADIENT.

Interactions The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

In SAS 9.4M2, FILLTYPE=GRADIENT is ignored when

GROUPDISPLAY=STACK is in effect. Starting with SAS 9.4M3,
FILLTYPE=COLORGRADIENT is honored in that case.

Tips The initial fill color is determined by a style element or by the

FILLATTRS= option COLOR= suboption. The initial transparency
276 Chapter 6 / Plot Statements

is determined by the DATATRANSPARENCY= option or by the

FILLATTRS= option TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

Note The output for SAS programs written prior to SAS 9.4M5 that
specify FILLTYPE=GRADIENT might change starting with SAS
9.4M5. To restore the original appearance, specify
FILLTYPE=ALPHAGRADIENT.

See DATASKIN= on page 269

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the
specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The bar display depends on the setting for the GROUPDISPLAY= option. For
example, for a vertical bar chart with GROUPDISPLAY=STACK, bar segments
are stacked to form the bar. The height of each segment represents the
corresponding group value’s proportional contribution to the response value.

A distinct bar or bar segment is created for each group value by varying the
visual attributes of the bar display features. The display features are controlled
by the current ODS style or by the DISPLAY= option. The default group
appearance for each display feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.
If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.
 BARCHART Statement 277

Display
Feature 1 Style Attributes That Control Default Group Appearance

Outline n if the outline is enabled with fill

color, fill pattern, or both,
ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is the only display
feature enabled, ContrastColor and
LineStyle attributes of a
GraphData1–GraphDataN style
element or the GraphMissing style
element

1 The default display features are determined by the DisplayOpts attribute of the GraphBar style
element.

Interactions Connect lines are not drawn for grouped data.

By default, the group values are mapped in the order of the data.
Use the GROUPORDER= option to control the sorting order of the
grouped bar segments. Use the INDEX= option to alter the default
sequence of colors and line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Notes The bar display depends on the setting for the GROUPDISPLAY=
option.

If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the bar outlines, but you
can use the PATTERN= setting on the OUTLINEATTRS= option to
assign the same line pattern to all bar outlines and connect lines.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.
278 Chapter 6 / Plot Statements

See “Example 2: Grouped Vertical Bar Chart” on page 295

“DISCRETEATTRVAR Statement” on page 1516

GROUP100=NONE | MAGNITUDE | POSITIVE

displays the computed response values (FREQ, SUM, or MEAN), normalized to
100%.

Note: This feature applies to SAS 9.4M3 and to later releases.

NONE
displays the summarized data.

MAGNITUDE
normalizes both the negative and positive values by magnitude to 100% of
the midpoint value, and displays the group values, preserving the sign.

The positive values are displayed above the bars for a vertical bar chart and
on the right end for a horizontal bar chart. The negative values are displayed
enclosed in parentheses below the bars for a vertical bar chart and on the
left end for a horizontal bar chart.

The following figure illustrates the effect of MAGNITUDE on stacked bars in

a vertical bar chart.

Notice that the group values for each midpoint value add up to 100%.

POSITIVE
drops the negative values and normalizes only the positive values to 100% of
the midpoint value.

The following figure demonstrates the effect of POSITIVE on clustered bars

in a vertical bar chart. This chart uses the same data as the chart in the
previous figure.
 BARCHART Statement 279

Notice that the negative values are dropped from the chart and the group
values for each midpoint value add up to 100%.

Default NONE

Requirement The GROUP= option must be specified for this option to have any
effect.

Interactions When this option is used, the BASELINEINTERCEPT= and

TARGET= options are ignored.

When MAGNITUDE or POSITIVE is specified and

BARLABEL=TRUE is in effect, the BARLABELFORMAT= option is
ignored and the bar labels are displayed as percentages with two-
digit precision.

Note You can use this option with any value for the GROUPDISPLAY=
option.

Tip To display the values, specify BARLABEL=TRUE.

See “Example 2: Grouped Vertical Bar Chart” on page 295

GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.

STACK
displays group values as stacked segments within the category bar.
280 Chapter 6 / Plot Statements

CLUSTER
displays group values as separate adjacent bars that replace the single
category bar.

Each cluster of group values is centered at the category midpoint on the axis.
This example illustrates the clusters and also how groups are displayed
when they have an unequal number of unique values.

Default STACK

Interaction When you use the BARLABEL= option and the GROUP= option, the
BARLABEL values are displayed for each bar when
GROUPDISPLAY=CLUSTER. When GROUPDISPLAY=STACK, the
whole bar is labeled at the top.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

Tip For a linear response axis, when STAT=MEAN or STAT=PCT, the axis
tick values might be displayed as integer values when
GROUPDISPLAY=STACK. Changing GROUPDISPLAY= to CLUSTER
in that case might cause the axis values to change to decimal values.
 BARCHART Statement 281

To keep the integer axis values in both cases, you can specify the
INTEGER=TRUE option for the response axis. See INTEGER= on
page 1085.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.
282 Chapter 6 / Plot Statements

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval
width.

Note: This feature applies to SAS 9.4M3 and to later releases.

 BARCHART Statement 283

Default The width specified by the BARWIDTH= option.

Restriction This option applies only to an interval category axis. When the
category axis is discrete, this option is ignored.

Interaction When the category data is interval, this option overrides the
BARWIDTH= option.

Tips To make the category axis type linear or time, include

TYPE=LINEAR or TYPE=TIME in the category axis options or assign
the role of primary plot to a plot that makes the category axis linear
or time.

The bar width that you specify with this option is honored even if
the bars overlap. If the bars overlap, reduce the interval bar width or
use the BARWIDTH= option instead.

See “dimension” on page xiii

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the

response-variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

Notes When this option is set to HORIZONTAL, the category variable appears
on the Y (or Y2) axis and the response variable appears on the X (or
X2) axis. To set the axis properties for this chart, you should use the
appropriate axis options of the layout container.
284 Chapter 6 / Plot Statements

When this option is set to VERTICAL, the category variable appears on

the X (or X2) axis and the response variable appears on the Y (or Y2)
axis. To set the axis properties for this chart, you should use the
appropriate axis options of the layout container.

If you change the orientation of the bar chart, then you should adjust
the layout container’s axis options appropriately.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled bars, the ContrastColor attribute of the
GraphData1–GraphDataN style elements, and the LineThickness
and LineStyle attributes of the GraphOutlines style element.

For grouped data and unfilled bars, the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements, and the
LineThickness attribute of the GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.
 BARCHART Statement 285

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

SEGMENTLABEL=TRUE | FALSE
specifies whether a label is displayed inside each bar segment.

Note: This feature applies to SAS 9.4M2 and to later releases.

For an ungrouped bar chart or for a grouped bar chart with

GROUPDISPLAY=CLUSTER, AUTO displays a bar label inside each bar. The
label displays the statistic for that bar. For a grouped bar chart with
GROUPDISPLAY=STACK, AUTO displays a label inside each bar segment. Each
segment label displays the statistic for that bar segment, as shown in the
following figure.

When this value is set to FALSE, no labels are displayed inside the bars.

Default FALSE

Note By default, the segment-label format is derived from the format that is
assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response-
column format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Tips For a grouped bar chart with GROUPDISPLAY=STACK, specify both

SEGMENTLABEL=TRUE and BARLABEL=TRUE to display a label for
each bar segment and a label for the entire bar.

Use the SEGMENTLABELATTRS= option to modify the appearance of

the label text.

Use the SEGMENTLABELFITPOLICY= option to specify a policy for

fitting the labels inside the bars.
286 Chapter 6 / Plot Statements

Use the SEGMENTLABELFORMAT= option to modify the format of the

segment labels.

See “boolean ” on page xii for other Boolean values that you can use.

SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the text properties of the text for the bar segment label.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default The GraphDataText style element.

Interaction This option is ignored when SEGMENTLABEL=FALSE.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Text Options” on page 1606 for available text-options.

SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN

specifies a policy for fitting the bar segment labels within the bar segments.

Note: This feature applies to SAS 9.4M2 and to later releases.

NONE
no attempt is made to fit each segment label within its bar. Long bar
segment labels might overlap other graphical elements.

The segment labels are not considered when the axis ranges are computed.
As a result, segment labels that extend beyond the plot area are clipped.

NOCLIP
does not clip bar segment labels that extend beyond the plot area.

Labels that do not fit within the plot area extend into the graph axis area and
might overlap axis elements.

THIN
drops any bar segment label that does not fit within its segment.

For a vertical bar chart, the label width must not exceed the bar width, and
the text height must not exceed the segment height. For a horizontal bar
chart, the label text height must not exceed the bar width, and the label
length must not exceed the segment length.

Default THIN

Interaction This option is ignored when SEGMENTLABEL=FALSE.

SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.
 BARCHART Statement 287

Note: This feature applies to SAS 9.4M2 and to later releases.

Default The segment-label format is derived from the format that is

assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response-
column format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Interaction This option is ignored when SEGMENTLABEL=FALSE.

Note When a segment-label format is specified with this option, the

segment labels are formatted as specified by format. The specified
format is not automatically expanded to accommodate values that
are too wide.

Tip If you want the segment-label format to expand automatically for

the summarized values on the response axis, specify the format for
the response column rather than in this option.

STAT=FREQ | PCT | SUM | MEAN | PROPORTION

specifies the statistic to be computed for the Y axis. For bar charts with no
RESPONSE= column:

FREQ frequency count

PCT percentages between 0 and 100
PROPORTION proportions between 0 and 1

Note: Prior to SAS 9.4, PCT displayed proportions between 0 and 1. To restore
the original PCT results in SAS 9.4 and later releases, specify PROPORTION
instead.

For bar charts with a RESPONSE= column:

SUM
MEAN

Defaults For bar charts with no RESPONSE= column, the default is FREQ.

For bar charts with a RESPONSE= column, the default is SUM.

Interaction When the GROUP100=MAGNITUDE or GROUP100=POSTIVE

option is in effect, the group values for each midpoint value are
normalized to 100% of the midpoint value. See GROUP100= on page
324.

Note When this option is used with the GROUP=group option, the
specified statistic is computed for each subgroup value of each
group value. See “Example 2: Grouped Vertical Bar Chart” on page
295.
288 Chapter 6 / Plot Statements

Tip If this option is used with COLORRESPONSE= in SAS programs that

were written before SAS 9.4M3, the bar-chart colors and color
statistic might change from those of the previous SAS releases. To
restore the original colors and color statistic, set COLORSTAT= in
the BARCHART statement to the same statistic that is specified in
STAT=.

TARGET=numeric-column | expression
specifies the target value for each bar. The visual representation is a triangle
with a line at the target value.
layout overlay;
barchart category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
name='bar';
discretelegend 'bar';
endlayout;

layout overlay;
barchartparm category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
name='bar';
discretelegend 'bar';
endlayout;

Default No targets are displayed.

Interactions For this option to take effect, the RESPONSE= argument must also
be used.

If the GROUP= option is used and GROUPDISPLAY= STACK, then

this option is ignored.

This option is ignored when the GROUP100= option is used.

Tips The statistic indicated by the STAT= option applies to the

TARGET=column. If a constant value is desired for each target, then
specify it only once for repeated category (X) values (or category
 BARCHART Statement 289

and GROUP combinations), and leave the other target values

missing.

The target color is that of the bar outline.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If
this option is used, then the information specified replaces all of the information
that is displayed by default.

(role-list)
an ordered, space-separated list of unique BARCHART roles. BARCHART
roles include CATEGORY or X, RESPONSE or Y, COLORRESPONSE, INDEX,
GROUP, and TARGET.

Notes For the category and response roles, the TIP= option recognizes
only the category and response arguments that you use in the
BARCHART statement. If you use the CATEGORY= and
RESPONSE= arguments, then you must specify roles CATEGORY
and RESPONSE. Conversely, if you use the X= and Y= arguments,
then you must specify roles X and Y.

The COLORRESPONSE role is valid starting with SAS 9.4M3.

Example The following example displays data tips for the columns assigned
only to the roles CATEGORY and RESPONSE:

TIP=(CATEGORY RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: CATEGORY or X, RESPONSE or Y,
COLORRESPONSE, and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.
290 Chapter 6 / Plot Statements

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when the bar is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each bar that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable bars, you must include an ODS GRAPHICS

ON statement that specifies the IMAGEMAP option, and you must
write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Notes For non-grouped data, the values of the column are expected to be
same for each unique category value. If they are not, then the
results might be unpredictable.

For grouped data, the values of the column are expected to be the
same for each unique category and GROUP combination.
 BARCHART Statement 291

Tips The URL value can be blank for some category values, meaning
that no action is taken when the bars for those category values are
selected.

The URL value can be the same for different category values,
meaning that the same action is taken when the bars for those
category values are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interactions This option is ignored if the RESPONSE= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Description
The BARCHART statement creates a bar chart with bars that represent
summarized response values. The response values are categorized by the unique
category values or, starting with SAS 9.4M3, by the bins in binned category data.
The BARCHART statement takes nonsummarized data as input and calculates the
appropriate summarization statistics (sum, mean, and so on) for each unique
category value or category bin. Prior to SAS 9.4M3, the category axis for a bar chart
must be discrete. Starting with SAS 9.4M3, the category axis can be discrete, linear,
or time. The response axis in all cases is interval.

When the chart is oriented vertically, the X (or X2) axis is used for CATEGORY and
the Y (or Y2) axis is used for RESPONSE. When it is oriented horizontally, the X (or
292 Chapter 6 / Plot Statements

X2) axis is used for RESPONSE and the Y (or Y2) axis is used for CATEGORY. (See
ORIENT= on page 283.)

By default, if the category column is character, then the bars in the chart appear in
the order in which the category values are present in the input data. If the category
column is numeric, then the values are presented in ascending order. For non-
grouped data, duplicated category values are summarized into a unique value. For
grouped data, the category values are summarized as needed. (See the GROUP=
option.)

Starting with SAS 9.4M3, for numeric category values, an interval bar chart is
generated only when the category axis type is linear or time. To specify a category
axis type of linear or time, include the TYPE= option in the category axis options, or
assign the role of primary plot to a plot that sets the category axis type to linear or
time automatically. By default, a bar is drawn for each unique category value, which
can result in a large number of bars for numeric category data.

When binning is used, for each bin, a summarization statistic is computed, and a bar
is drawn that represents that statistic. The width of each bar spans the width of the
bin that it represents. The left-most edge of the bar represents the start of the bin,
and the right-most edge represents the end. See “Example 1: Horizontal Bar Chart”
on page 294.

TIP Prior to SAS 9.4M3, use the HISTOGRAM statement to create a bar
chart that represents response values along an interval axis.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables
side by side on a common axis. By default within an overlay-type layout, if multiple
BARCHART statements are used with different response variables, then the bars
for matching category values are centered on the midpoints and the bars are
superimposed. To make it easier to distinguish among superimposed bars, you can
assign a different BARWIDTH= setting to each BARCHART statement in the
overlay:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A" barwidth=0.8 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B" barwidth=0.6 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C" barwidth=0.4 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
 BARCHART Statement 293

To place the different response values side by side, you can assign a different offset
to each BARCHART statement. The BARWIDTH= option can be used with
DISCRETEOFFSET= to create narrower bars that require less width within the plot
area:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A"
discreteoffset=-0.3 barwidth=0.3 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B"
discreteoffset=0 barwidth=0.3 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C"
discreteoffset=+0.3 barwidth=0.3 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
294 Chapter 6 / Plot Statements

Different combinations of DISCRETEOFFSET and BARWIDTH can be used to get

the effect that you want. Gaps can be created between bars by providing a
narrower bar width. Or, bars can be overlapped if the bar widths are increased in
proportion to the discrete offset.
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchart category=year response=A_revenue / stat=sum name="A"

legendlabel="A" datatransparency=0.2
discreteoffset=-0.2 barwidth=0.5 ;
barchart category=year response=B_revenue / stat=sum name="B"
legendlabel="B" datatransparency=0.2
discreteoffset=0 barwidth=0.5 ;
barchart category=year response=C_revenue / stat=sum name="C"
legendlabel="C" datatransparency=0.2
discreteoffset=+0.2 barwidth=0.5 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;

Examples

Example 1: Horizontal Bar Chart

The following graph was generated by the “Example Program” on page 295:
 BARCHART Statement 295

Example Program
proc template;
define statgraph barchart;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
layout overlay;
barchart category=type response=mpg_highway /
stat=mean orient=horizontal;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.cars template=barchart;

run;

Example 2: Grouped Vertical Bar Chart

The following graph was generated by the “Example Program” on page 296:
 296 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph barchart;
begingraph / attrpriority=none;
entrytitle "Percentage of Total Models by Type and Origin";
layout overlay;
barchart x=type / name="carmodels"
stat=pct display=all
group=origin groupdisplay=cluster group100=positive
barlabel=true;
discretelegend "carmodels";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=barchart;

run;

Details
In the BARCHART statement, the STAT= option specifies the PCT statistic. The
GROUP= specifies grouping by variable Origin. The GROUP100= option specifies
normalization of the frequency statistic to 100% of the midpoint value. For each
midpoint, the sum of the percentages is 100%. The BARLABEL=TRUE option
enables the display of bar labels.
 BARCHART Statement 297

Example 3: Interval Bar Chart

Interval bar charts are available starting with SAS 9.4M3. In SAS 9.4M2 and in
earlier releases, use the HISTOGRAM statement to generate an interval bar chart.
The following graph was generated by the “Example Program” on page 297:

Example Program
Here is the SAS code.
proc template;
define statgraph cylinders;
begingraph;
entrytitle "Interval Bar Chart of Vehicle Engine Cylinders";
layout overlay /
xaxisopts=(label="Engine Cylinders" type=linear
linearopts=(tickvaluelist=(3 4 5 6 8 10 12)))
yaxisopts=(label="Percentage of Vehicles Manufactured"
griddisplay=on linearopts=(tickvalueformat=percent7.1));
barchart category=cylinders / stat=proportion
barlabel=true barlabelformat=percent7.1;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=cylinders;

run;
 298 Chapter 6 / Plot Statements

Details
An interval bar chart can be generated only when the category axis type is LINEAR
or TIME. In this example, the TYPE=LINEAR option is included in the XAXISOPTS=
options. With numeric category data, a bar is drawn for each unique category value.
In some cases, that can generate too many bars in the resulting chart. In this
example, there are only seven unique values. The TICKVALUELIST= option is used
in the XAXISOPTS= option to display all of the values on the category axis.

Example 4: Bar Chart with Bar Colors Controlled by a Statistic

The ability to use a computed statistic to control the bar colors in a bar chart is
available starting with SAS 9.4M3. This example uses the COLORBYFREQ=TRUE
option to enable a computed statistic to control the bar colors and the
COLOSTAT=PCT to specify percentage as the controlling statistic. Here is the
output from “Example Program” on page 298.

Example Program
proc template;
define statgraph barchart;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
layout overlay;
barchart category=type response=mpg_highway / name="bar"
stat=mean orient=horizontal
 BARCHART Statement 299

colorbyfreq=true colorstat=pct;
continuouslegend "bar" /
title="Percent of Total Models Manufactured";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=barchart;

run;

Example 5: Bar-Line Chart

This example uses the BARCHART and LINECHART statements to overlay a bar
chart and line chart to create a bar-line chart. The following figure was generated
by the “Example Program” on page 299.

Example Program
proc template;
define statgraph barline;
begingraph;
entrytitle "Volume and Closing Price in 1H2001";
layout overlay /
xaxisopts=(label="Month"
timeopts=(tickvalueformat=monname3.))
yaxisopts=(label="Volume (Millions)" offsetmin=0
linearopts=(tickvaluesequence=(start=0 end=120
increment=20)
tickvaluepriority=true))
y2axisopts=(label="Close" offsetmin=0
 300 Chapter 6 / Plot Statements

linearopts=(viewmin=50 viewmax=75));
barchart category=date response=eval(volume/1000000) /
name="bar"
legendlabel="Volume" datatransparency=0.6;
linechart category=date response=close /
legendlabel="Closing Price" vertexlabel=true
vertexlabelattrs=(color=darkblue weight=bold)
name="line" stat=sum yaxis=y2 display=(line markers)
markerattrs=(symbol=circlefilled color=darkblue)
lineattrs=(color=darkblue);
discretelegend "bar" "line";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.stocks template=barline;

where stock eq "Microsoft" and year(date) eq 2001 and
month(date) le 6;
format date monname3. close dollar6.0;
run;

BARCHARTPARM Statement
Creates a bar chart specified by pre-summarized data.

Requirement: The input data must be pre-summarized, with appropriate summarization statistics
(sum, mean, and so on) computed for the RESPONSE column.
Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In SAS
9.4M2 and in earlier releases, specify SUBPIXEL=ON in the BEGINGRAPH
statement to enable subpixel rendering. See SUBPIXEL= on page 39. Starting with
SAS 9.4M3, subpixel rendering is enabled by default.
To disable subpixel rendering in SAS 9.4M3 and in later releases, specify
SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 39. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” on page 1561.

Syntax
BARCHARTPARM CATEGORY=column | expression
RESPONSE=numeric-column | expression </options>;
 BARCHARTPARM Statement 301

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar
colors to a continuous color gradient.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar connect lines.
CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, error
bars, connect line, and data labels, if displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
DISPLAYBASELINE=ON | OFF | AUTO
specifies whether the baseline is displayed.
DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars that are associated with the data
points.
ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the error-bar caps.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.
ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bar area.
FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.
302 Chapter 6 / Plot Statements

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT
specifies the bar fill type.
INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.
INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the
interval width.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.
TARGET=numeric-column | expression
specifies the target value for each bar.

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
bar.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies the label that appears at the end of each bar.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the labels that are specified in
the DATALABEL= option.
 BARCHARTPARM Statement 303

DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS |

INSIDEPREFERRED
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed data
labels.
DATALABELTYPE=AUTO | COLUMN
specifies whether the data labels display the RESPONSE values or the
values of the column that is specified by the DATALABEL= option.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the text properties of the text for the bar segment label.
SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN
specifies a policy for fitting the bar segment labels within the bar
segments.
SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.
SEGMENTLABELTYPE=NONE | AUTO
specifies whether a label is displayed inside each bar segment.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
GROUP100=NONE | MAGNITUDE | POSITIVE
displays the response values, normalized to 100%.
GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when the bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
304 Chapter 6 / Plot Statements

Required Arguments
CATEGORY=column | expression
specifies the column for the unique category values. All values are treated as
discrete.

Note You can use X= as an alternative to CATEGORY=. If you use X=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will
recognize X as the category role and not CATEGORY in that case.

RESPONSE=numeric-column | expression
specifies the column for the response values.

Note You can use Y= as an alternative to RESPONSE=. If you use Y=, then be
aware that the TIP=, TIPFORMAT=, and TIPLABEL= options will
recognize Y as the response role and not RESPONSE in that case.

Optional Arguments
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interaction Starting with SAS 9.4M3, the INTERVALBARWIDTH= option

overrides this option for an interval bar chart.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number

of bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The ContrastColor, LineThickness, and LineStyle attributes of the

GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s Color,

LineStyle, and LineThickness attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

 BARCHARTPARM Statement 305

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. Prior to SAS 9.4M5, the
baseline is always displayed in the chart, whether for a specified value or for the
default value. Starting with SAS 9.4M5, the baseline display is controlled by the
DISPLAYBASELINE= option, which is ON by default. When the
BASELINEINTERCEPT= option is used, the axis range is adjusted to include the
baseline, and the baseline is placed at the specified value on the response axis.

Default 0

Interactions If GROUPDISPLAY=STACK is specified, then this option is ignored

and the baseline is not displayed.

If necessary, the response axis data range is extended to include

the baseline intercept. When a logarithmic response axis is
requested and BASELINEINTERCEPT= specifies 0 or a negative
value, the response axis reverts to a linear axis. To restore the log
axis in that case, set BASELINEINTERCEPT= to a positive value.

When DISPLAYBASELINE=AUTO is in effect, the baseline is not

displayed if the baseline intercept is equal to the minimum or
maximum value of the response-axis range.

Note Label positions are automatically adjusted to prevent the labels

from overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline prior to SAS 9.4M5, use the

BASELINEATTRS= option to set the line thickness to 0. Starting
with SAS 9.4M5, specify DISPLAYBASELINE=OFF.

The baseline does not add a tick or a tick value to the axis. To label
the baseline, use a REFERENCELINE statement to overlay a line
with the same X or Y value and include the CURVELABEL= option
to specify the label text.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.
306 Chapter 6 / Plot Statements

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Interaction When GROUPDISPLAY=CLUSTER, the default BARWIDTH is 1.0.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

 BARCHARTPARM Statement 307

See “color ” on page xii

Defaults For outline-only bars, the ThreeColorAltRamp style element

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip Use the DISPLAY= option to specify whether outlines and fills are
displayed.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar colors to a
continuous color gradient.

Note: This feature applies to SAS 9.4M3 and to later releases.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. Each bar is colored using one color from the gradient
range. When a range attribute map variable is specified, the colors that are
defined in the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill, fill pattern, or both are displayed, this option overrides
suboption COLOR= in the FILLATTRS= option and in the
FILLPATTERNATTRS= option and varies the color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides

suboption COLOR= in the OUTLINEATTRS= option and varies the
outline color according to the color gradient or the attribute map.
308 Chapter 6 / Plot Statements

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the FILLTYPE= option to specify whether each bar is filled

with a solid color or with a gradient color.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar connect lines.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement DISPLAY= must include CONNECT for this option to have any
effect.

Interaction This option is ignored when the GROUP= option is in effect.

DATALABEL=column | expression
specifies the label that appears at the end of each bar.

Restriction This option is ignored when one or both of the ERRORLOWER= and
ERRORUPPER= options are in effect.

Interactions Starting with SAS 9.4M2, this option is ignored when

DATALABELTYPE=AUTO.

When the GROUP= option is in effect, the data label values are
displayed only when GROUPDISPLAY=CLUSTER.

If the GROUP= option is in effect and there are multiple input

observations per bar for the GROUP= column, then the value for
the DATALABEL= column should be the same for each observation
that is on the same bar.

Tip The data-label format is the format that is assigned to column or is

derived from BEST6 if no format is assigned. To change the data-
label format, apply the new format to column.
 BARCHARTPARM Statement 309

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the labels that are specified in the
DATALABEL= option.

Default The GraphDataText style element.

Interaction For this option to take effect, the DATALABEL= option must also be
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS |

INSIDEPREFERRED
specifies a policy for avoiding collisions among the bar labels when labels are
displayed.

AUTO
selects a collision avoidance policy based on the chart orientation and data
type.

For a numeric column with ORIENT=VERTICAL, AUTO rotates the labels if

they do not fit the midpoint spacing. For a character column, AUTO splits the
labels if they do not fit the midpoint spacing.

Note When ORIENT=HORIZONTAL, AUTO always draws the labels

horizontally.

Tip If character labels do not fit after splitting, then try using ROTATE
instead of AUTO.

See ORIENT= on page 330 for information about chart orientation.

BARWIDTH= for information about bar spacing.

NONE
does not attempt to fit bar labels that collide.

ROTATE
rotates the bar labels for vertical bars if the labels collide in the available
width.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

SPLIT
splits the label for vertical bars at a split character only if a split is needed at
that character in order to make the label fit the available space.

No split occurs at split characters that occur where a split is not needed. If
the label does not contain any of the specified split characters, then a split
does not occur. In that case, if the label does not fit the available space, then
it might collide with the adjoining labels.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

310 Chapter 6 / Plot Statements

See the DATALABELSPLITCHAR= option for information about

specifying the split characters

SPLITALWAYS
splits the label for vertical bars at every occurrence of a split character.

If the label does not contain any of the specified split characters, then a split
does not occur.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See the DATALABELSPLITCHAR= option for information about

specifying the split characters

INSIDEPREFERRED
attempts to place the data labels in a horizontal bar chart inside the bars.

Note: This feature applies to SAS 9.4M5 and to later releases.

Restrictions This option applies to non-grouped or cluster-grouped

horizontal bar charts only.

Label splitting is not supported when a label is placed inside a

bar.

Requirement For a grouped horizontal bar chart, the

GROUPDISPLAY=CLUSTER option must be in effect.
Otherwise, the INSIDEPREFERRED policy is ignored.

Interaction If SEGMENTLABEL=TRUE is in effect, the INSIDEPREFERRED

policy is ignored, and the labels are placed outside.

Notes If any bar label does not fit inside its bar because the font
height is too high, all of the bar labels are placed outside.

If any bar label cannot fit inside the length of its bar, that label
is placed outside of its bar and is fit using the AUTO fit policy.

If the outside labels collide because the font height is greater

than the midpoint space, then all labels are dropped.

Here is an example of a vertical bar chart where DATALABELFITPOLICY=AUTO

and a numeric column is used as the data labels.
 BARCHARTPARM Statement 311

In this case, AUTO rotates the numeric labels to avoid collision.

In some cases, if one or more labels collide when the specified fit policy is used,
then all of the labels are dropped from the display. When that occurs, the
following warning message is written to the SAS log:
WARNING: The bar labels are suppressed. Use DATALABELFITPOLICY=NONE to
force the labels to be displayed.

Here is an example of a horizontal bar chart where

DATALABELFITPOLICY=INSIDEPREFERRED and a numeric column is used as
the data labels.

Default AUTO

Requirement The DATALABEL= option must also be specified.

312 Chapter 6 / Plot Statements

Interaction When DATALABELTYPE=AUTO is in effect, for a vertical bar chart,

only AUTO, NONE, and ROTATE are valid. All other values revert
to AUTO. For a horizontal bar chart, only AUTO and NONE are
valid. All other values revert to AUTO.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split. When
multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in
the data label. In that case, all of the specified split characters together are
treated as a single split character.

When DATALABELFITPOLICY=SPLIT and a data label collision is detected, the

data label is split on a specified split character only if a split is needed at that
point in order to make the label fit. In that case, a split might not occur on every
split character. When DATALABELFITPOLICY=SPLITALWAYS, the data label is
split unconditionally on every occurrence of a split character. If the data label
does not contain any of the specified split characters, then the label is not split.

"character-list"
specifies one or more characters with no space between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The DATALABELFITPOLICY= option specifies the policy that is

used to manage the split behavior of the data label.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the displayed data label or are
dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed data labels.

TRUE
drops a split character from the display when a split occurs at that character.

Split characters at which a split does not occur are left in place. The
DATALABELFITPOLICY= option determines where the labels are split. When
 BARCHARTPARM Statement 313

DATALABELFITPOLICY=SPLIT, each label is split at a split character only

where a split is needed in order to make the label fit the available space. At
each split point, the split character is dropped, and the characters that
follow the split character, up to but not including the split character at the
next split point, are wrapped to the following line.

When DATALABELFITPOLICY=SPLITALWAYS, each label is split at every

instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next
split character, are wrapped to the next line.

The following figure shows how label Product*Group*1 is split when the
DATALABELSPLITCHARDROP=TRUE and DATALABELSPLITCHAR="*"
options are specified with the SPLIT and SPLITALWAYS fit policies.

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is
needed at the second asterisk. The first asterisk is dropped, and Group*1
wraps to the next line. Notice that the second asterisk is not dropped in this
case. When DATALABELFITPOLICY=SPLITALWAYS, the label is split at
every occurrence of the asterisk. In this case, both asterisks are dropped, and
the characters that follow each asterisk wrap to the next line.

FALSE
includes the split characters in the data label display.

The DATALABELFITPOLICY= option determines how the split characters are

displayed. When DATALABELFITPOLICY=SPLIT, each data label is split at a
split character only where a split is needed in order to make the label fit the
available space. A split might not occur at every split character in the label.
At each split point, the split character remains as the last character in the
current line. The characters that follow the split character, up to and
including the split character at the next split point, are then wrapped to the
following line. This process repeats until the entire data label is displayed.

When DATALABELFITPOLICY=SPLITALWAYS, each data label is split at

every instance of a split character in the label regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line.

The following figure shows how label Product*Group*1 is split when the
DATALABELSPLITCHARDROP=FALSE and DATALABELSPLITCHAR="*"
options are specified with the SPLIT and SPLITALWAYS fit policies.
314 Chapter 6 / Plot Statements

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is
needed at the second asterisk. The characters that follow the first asterisk
wrap to the next line. When DATALABELFITPOLICY=SPLITALWAYS, the
label is split at every occurrence of the asterisk. Each asterisk remains as the
last character in the current line, and the characters that follow are wrapped
to the next line.

Default TRUE. A split character is dropped from the data-label display

when a split occurs at that character.

Requirements The DATALABEL= option must also be specified.

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELTYPE=AUTO | COLUMN
specifies whether the data labels display the RESPONSE values or the values of
the column that is specified by the DATALABEL= option.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
the system determines how the labels are displayed.

The system displays the labels as follows:

n For an ungrouped bar chart, the summarized value for each bar is placed
above the bar.
n For a grouped bar chart with stacked bars, the total of the summarized
segment values for each bar is placed above the segmented bar.
n For a grouped bar chart with clustered bars, the summarized value for
each bar in the cluster is placed above the bar.

Interactions AUTO overrides the DATALABEL= option.

When AUTO is in effect, some data-label fit policies are

unavailable. See DATALABELFITPOLICY=.

COLUMN
the data labels display the DATALABEL= column values.
 BARCHARTPARM Statement 315

Interaction The DATALABEL= option must be specified for COLUMN to have

any effect. If the DATALABEL= option is not specified, AUTO is
used instead.

Default COLUMN

Restriction The DATALABELTYPE= option and the DATALABEL= on page 308

option are ignored when one or both of the ERRORLOWER= on page
319 and ERRORUPPER= on page 320 options are in effect.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
bars with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.
316 Chapter 6 / Plot Statements

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, error bars,
connect line, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
bar fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines, error bars, and connect lines but a
different transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all bars from the category midpoints.

Default 0 (no offset, all bars are centered on the category midpoints)

Range -0.5 to +0.5, where 0.5 represents half the distance between category
ticks. Normally, a positive offset is to the right when
ORIENT=VERTICAL, and up when ORIENT=HORIZONTAL. (If the
layout's axis options set REVERSE=TRUE, then the offset direction is
also reversed.)

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete offset
can cause clipping at each end of the axis. In those cases, use the
OFFSETMIN= and OFFSETMAX= axis options to increase the axis
minimum and maximum offsets to accommodate the discrete offset.

See “About the DISCRETEOFFSET= Option” on page 338

“About the DISCRETEOFFSET= Option” on page 338 for information

about the OFFSETMIN= and OFFSETMAX= axis options

ORIENT=

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

CONNECT
displays the connect lines, which are line segments that connect adjacent
midpoints at the end of each bar

FILL
displays the bar fill color
 BARCHARTPARM Statement 317

FILLPATTERN
displays the bar fill pattern

Tip For this option to have any effect, the active ODS style must specify
a fill pattern or a fill pattern must be specified with the
FILLPATTERNATTRS= on page 321 option.

OUTLINE
displays the bar outline

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: CONNECT, FILL, FILLPATTERN, and OUTLINE

Default The GraphBar:DisplayOpts style reference.

Interaction Connect lines are not drawn when the GROUP= option is in effect.

Note Error bars are automatically displayed whenever the ERRORUPPER=

or ERRORLOWER= options are specified.

Tips Use the OUTLINEATTRS= , FILLATTRS= , and

FILLPATTERNATTRS= options to control the appearance of the
bars. Use CONNECTATTRS= to control the appearance of the
connect lines.

Both FILL and FILLPATTERN can be specified to combine solid fills

and pattern fills in the bars.

DISPLAYBASELINE=ON | OFF | AUTO

specifies whether the baseline is displayed.

Note: This feature applies to SAS 9.4M5 and to later releases.

ON
always displays the baseline.

OFF
does not display the baseline.

AUTO
displays the baseline if the baseline intercept is within the response-axis
range, excluding the minimum and maximum axis values.

This is the typical case when the bar chart includes both positive and
negative bars, and the default baseline intercept of 0 is used. If the baseline
intercept is equal to the minimum or maximum value of the response-axis
range, the baseline is not displayed.

Default ON

Tip By default, the baseline intercept is 0. Use the BASELINEINTERCEPT=

option to change the baseline intercept.
318 Chapter 6 / Plot Statements

DISPLAYZEROLENGTHBAR=TRUE | FALSE
specifies whether zero-length bars are drawn.

Note: This feature applies to SAS 9.4M3 and to later releases.

A zero-length bar is displayed as a line spanning the normal bar width at the
bar-chart baseline on the response axis. When this option is set to TRUE, zero-
length bars are displayed. Otherwise, they are suppressed. The following figure
shows a simple example of each outcome. In the figure, the plot wall outline,
category axis line, and bar-chart baseline are suppressed for clarity.

Default TRUE

Interaction This option is ignored when the GROUP= and

GROUPDISPLAY=STACK options are in effect. In that case, zero-
length bar segments are drawn.

Note When this option is set to FALSE, the bar is not drawn, but other
elements associated with the bar such as the target bar, the error
bar, the bar label, and the data label, are drawn.

Tip This option is useful when the bar-chart baseline is suppressed.

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data
points.

Default The GraphError style element ContrastColor, LineStyle, and

LineThickness attributes.

Interaction For this option to take effect, error bars must be displayed by the
ERRORLOWER= , ERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the error-bar caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

 BARCHARTPARM Statement 319

Default 1

Interaction This option is ignored when ERRORBARCAPSHAPE=NONE is in

effect.

Tips Use the ERRORBARCAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the ERRORBARATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the error-bar caps:

errorbarcapscale=2

Reduce the size of the error-bar caps by 50%:

errorbarcapscale=0.5

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.

Default The GraphError:CapStyle style reference. If attribute CapStyle is not

defined in the active style, then SERIF is the default value.

Interaction Starting with SAS 9.4M5, when ERRORBARCAPSCALE= is in effect,

the display width for SERIF caps is the cap's default width multiplied
by the value specified in ERRORBARCAPSCALE=.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.

Default The lower segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interactions If the GROUP= option is specified with GROUPDISPLAY=STACK or

with GROUP100=POSITIVE or MAGNITUDE, then this option is
ignored.

When this option is specified, the DATALABEL= on page 308 and

DATALABELTYPE= on page 314 options are ignored, and data
labels are not displayed.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.
320 Chapter 6 / Plot Statements

ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.

Default The upper segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interactions If the GROUP= option is specified with GROUPDISPLAY=STACK or

with GROUP100=POSITIVE or MAGNITUDE, then this option is
ignored.

When this option is specified, the DATALABEL= on page 308 and

DATALABELTYPE= on page 314 options are ignored, and data
labels are not displayed.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bar area.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the GraphData1:Color–GraphDataN:Color style

references.

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is ignored,
and the bar fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for bar

fills, bar outlines, error bars, and connect lines. You can combine this
option with DATATRANSPARENCY= to set one transparency for the
bar outlines, error bars, and connect lines but a different
transparency for the bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element .

“Fill Color Options” on page 1600 for available fill-options.

FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphWall style element Color attribute.

Requirement FILLTYPE=COLORGRADIENT must be in effect. Otherwise, this

option is ignored.

See “color ” on page xii

 BARCHARTPARM Statement 321

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Restriction In SAS 9.4M4 and in earlier releases, the only styles that are
delivered by SAS that support fill patterns are JOURNAL2,
JOURNAL3, and MONOCHROMEPRINTER. If style-element is
specified and the active ODS style does not support fill patterns,
this option is ignored.

Interactions For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILLPATTERN display, the FILLPATTERNATTRS=
suboption COLOR= is ignored, and the fill-pattern colors vary
according to the gradient.

Tip In order to change the fill pattern for each group value in ODS
styles such as HTMLBlue that use color-priority attribute rotation,
specify ATTRPRIORITY=NONE in an ODS GRAPHICS statement or
in the BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT

specifies the bar fill type.

Note: This feature applies to SAS 9.4M2 and to later releases.

SOLID
fills each bar with the color that is assigned to that bar.

ALPHAGRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This feature applies to SAS 9.4M5 and to later releases.

Tips The fill color is determined by a style element or by the FILLATTRS=

option COLOR= suboption. The initial transparency is determined by
322 Chapter 6 / Plot Statements

the DATATRANSPARENCY= option or by the FILLATTRS= option

TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete attribute

map to set the initial transparency in the gradients for specific values.

COLORGRADIENT
fills each bar with a color gradient that starts with the fill color that is
assigned to that bar and ends with the color that is specified in the
FILLENDCOLOR= option.

Note: This feature applies to SAS 9.4M5 and to later releases.

Alias GRADIENT

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tips Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

See “FILLENDCOLOR=color” on page 320

GRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This option applies to SAS 9.4M2 through SAS 9.4M4. Starting with
SAS 9.4M5, ALPHAGRADIENT replaces GRADIENT, and GRADIENT is
changed to an alias of COLORGRADIENT.

Interactions The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

In SAS 9.4M2, FILLTYPE=GRADIENT is ignored when

GROUPDISPLAY=STACK is in effect. Starting with SAS 9.4M3,
FILLTYPE=COLORGRADIENT is honored in that case.

Tips The initial fill color is determined by a style element or by the

FILLATTRS= option COLOR= suboption. The initial transparency
is determined by the DATATRANSPARENCY= option or by the
FILLATTRS= option TRANSPARENCY= suboption.
 BARCHARTPARM Statement 323

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

Note The output for SAS programs written prior to SAS 9.4M5 that
specify FILLTYPE=GRADIENT might change starting with SAS
9.4M5. To restore the original appearance, specify
FILLTYPE=ALPHAGRADIENT.

See DATASKIN= on page 315

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the
specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The bar display depends on the setting for the GROUPDISPLAY= option. For
example, for a vertical bar chart with GROUPDISPLAY=STACK, bar segments
are stacked to form the bar. The height of each segment represents the
corresponding group value’s proportional contribution to the response value.

A distinct bar or bar segment is created for each group value by varying the
visual attributes of the bar display features. The display features are controlled
by the current ODS style or by the DISPLAY= option. The default group
appearance for each display feature is shown in the following table.

Display
Feature1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.
If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.

Outline n if the outline is enabled with fill

color, fill pattern, or both,
324 Chapter 6 / Plot Statements

Display
Feature1 Style Attributes That Control Default Group Appearance

ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is the only display
feature enabled, ContrastColor and
LineStyle attributes of a
GraphData1–GraphDataN style
element or the GraphMissing style
element

1 The default display features are determined by the DisplayOpts attribute of the GraphBar style
element.

Interactions Connect lines are not drawn for grouped data.

By default, the group values are mapped in the order of the data.
Use the GROUPORDER= option to control the sorting order of the
group values.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the bar outlines, but you
could use the PATTERN= setting on the OUTLINEATTRS= option to
assign the same line pattern to all bar outlines and connect lines.

Use the INDEX= option to alter the default sequence of colors, fill
patterns, and line patterns.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

GROUP100=NONE | MAGNITUDE | POSITIVE

displays the response values, normalized to 100%.
 BARCHARTPARM Statement 325

Note: This feature applies to SAS 9.4M3 and to later releases.

NONE
displays the data as received.

MAGNITUDE
normalizes both the negative and positive values by magnitude to 100% of
the midpoint value, and displays the group values, preserving the sign.

The positive values are displayed above the bars for a vertical bar chart and
on the right end for a horizontal bar chart. The negative values are displayed
enclosed in parentheses below the bars for a vertical bar chart and on the
left end for a horizontal bar chart.

The following figure illustrates the effect of MAGNITUDE on stacked bars in

a vertical bar chart.

Notice that the group values for each midpoint value add up to 100%.

POSITIVE
drops the negative values and normalizes only the positive values to 100% of
the midpoint value.

The following figure demonstrates the effect of POSITIVE on clustered bars

in a vertical bar chart. This chart uses the same data as the chart in the
previous figure.
326 Chapter 6 / Plot Statements

Notice that the negative values are dropped from the chart and the group
values for each midpoint value add up to 100%.

Default NONE

Requirement The GROUP= option must be specified for this option to have any
effect.

Interaction Error bars are not drawn when GROUP100=POSITIVE or

MAGNITUDE. See ERRORLOWER= and ERRORUPPER=.

Note You can use this option with any value for the GROUPDISPLAY=
option.

Tip To display the values, specify DATALABELTYPE=AUTO.

GROUPDISPLAY=STACK | CLUSTER
specifies how to display grouped bars.

STACK
displays group values as stacked segments within the category bar.
 BARCHARTPARM Statement 327

CLUSTER
displays group values as separate adjacent bars that replace the single
category bar.

Each cluster of group values is centered at the category midpoint on the axis.
This example illustrates the clusters and also how groups are displayed
when they have an unequal number of unique values.

Default STACK

Interactions When you use the DATALABEL= option and the GROUP= option,
the DATALABEL values are displayed for each bar when
GROUPDISPLAY=CLUSTER. When GROUPDISPLAY=STACK, the
whole bar is labeled at the top.

Error bars are not drawn when GROUPDISPLAY=STACK.

When the TARGET= and GROUP= options are in effect, the target
values are not displayed when GROUPDISPLAY=STACK. In that
case, you must specify GROUPDISPLAY=CLUSTER to display the
target values.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they
328 Chapter 6 / Plot Statements

are drawn in ascending order of the group column character values

or unformatted numeric values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.
 BARCHARTPARM Statement 329

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping bar attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

INTERVALBARWIDTH=dimension
specifies the width of the bars in an interval bar chart as a ratio of the interval
width.

Note: This feature applies to SAS 9.4M3 and to later releases.

330 Chapter 6 / Plot Statements

Default The width specified by the BARWIDTH= option.

Restriction This option applies only to an interval category axis. When the
category axis is discrete, this option is ignored.

Interaction When the category data is interval, this option overrides the
BARWIDTH= option.

Tips To make the category axis type linear or time, include

TYPE=LINEAR or TYPE=TIME in the category axis options or assign
the role of primary plot to a plot that makes the category axis linear
or time.

The bar width that you specify with this option is honored even if
the bars overlap. If the bars overlap, reduce the interval bar width or
use the BARWIDTH= option instead.

See “dimension” on page xiii

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the

response-variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

Notes When this option is set to VERTICAL, the category variable appears on
the X (or X2) axis and the response variable appears on the Y (or Y2)
axis. To set the axis properties for this chart, you should use the
appropriate axis options of the layout container.
 BARCHARTPARM Statement 331

When this option is set to HORIZONTAL, the category variable appears

on the Y (or Y2) axis and the response variable appears on the X (or
X2) axis. To set the axis properties for this chart, you should use the
appropriate axis options of the layout container.

If you change the orientation of the bar chart, then you should adjust
the layout container’s axis options appropriately.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled bars, the ContrastColor attribute of the
GraphData1–GraphDataN style elements, and the LineThickness
and LineStyle attributes of the GraphOutlines style element.

For grouped data and unfilled bars, the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements, and the
LineThickness attribute of the GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.
332 Chapter 6 / Plot Statements

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY or X, RESPONSE or Y,
ERRORLOWER, ERRORUPPER, GROUP, and INDEX.

SEGMENTLABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the text properties of the text for the bar segment label.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default The GraphDataText style element.

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Text Options” on page 1606 for available text-options.

SEGMENTLABELFITPOLICY=NONE | NOCLIP | THIN

specifies a policy for fitting the bar segment labels within the bar segments.

Note: This feature applies to SAS 9.4M2 and to later releases.

NONE
no attempt is made to fit each segment label within its bar. Long bar
segment labels might overlap other graphical elements.

The segment labels are not considered when the axis ranges are computed.
As a result, segment labels that extend beyond the plot area are clipped.

NOCLIP
does not clip bar segment labels that extend beyond the plot area.
 BARCHARTPARM Statement 333

Labels that do not fit within the plot area extend into the graph axis area and
might overlap axis elements.

THIN
drops any bar segment label that does not fit within its segment.

For a vertical bar chart, the label width must not exceed the bar width, and
the text height must not exceed the segment height. For a horizontal bar
chart, the label text height must not exceed the bar width, and the label
length must not exceed the segment length.

Default THIN

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

SEGMENTLABELFORMAT=format
specifies the text format for the bar segment labels.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default If DATALABELTYPE=AUTO is in effect, the segment-label format is

derived from the response column format. If
DATALABELTYPE=COLUMN is in effect, the segment-label format
is derived from the DATALABEL= column format. If no format is
applied to the column, the label format is derived from BEST6. The
derived format retains the precision of the column format and, if
necessary, increases the format width to accommodate the column
values.

Interaction This option is ignored when SEGMENTLABELTYPE=NONE.

Note When a segment-label format is specified with this option, the

segment labels are formatted as specified by format. The specified
format is not automatically expanded to accommodate values that
are too wide.

Tip If you want the segment-label format to expand automatically for

the column values, specify the format for the response column or
DATALABEL= column rather than in this option.

SEGMENTLABELTYPE=NONE | AUTO
specifies whether a label is displayed inside each bar segment.

Note: This feature applies to SAS 9.4M2 and to later releases.

For an ungrouped bar chart or for a grouped bar chart with

GROUPDISPLAY=CLUSTER, AUTO displays a bar label inside each bar. The
label for each bar displays the value for that bar. For a grouped bar chart with
GROUPDISPLAY=STACK, AUTO displays a label inside each bar segment. The
label for each bar segment displays the value for that segment.
334 Chapter 6 / Plot Statements

When this value is set to NONE, no labels are displayed inside the bars.

Default NONE

Interaction The DATALABELTYPE= option determines whether the segment

labels display the RESPONSE column values or the values of the
column that is specified by the DATALABEL= option.

Note By default, the segment-label format is derived from the format that
is assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response
column format and, if necessary, increases the format width to
accommodate the column values.

Tips Use the SEGMENTLABELATTRS= option to modify the appearance

of the label text.

Use the SEGMENTLABELFITPOLICY= option to specify a policy for

fitting the labels inside the bars.

Use the SEGMENTLABELFORMAT= option to modify the format of

the segment labels.

See “boolean ” on page xii for other Boolean values that you can use.

TARGET=numeric-column | expression
specifies the target value for each bar.
layout overlay;
barchart category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
name='bar';
discretelegend 'bar';
endlayout;

The visual representation is a triangle with a line at the target value.

layout overlay;
barchartparm category=type response=mpg_highway / barwidth=.8
target=mpg_city group=origin groupdisplay=cluster
 BARCHARTPARM Statement 335

name='bar';
discretelegend 'bar';
endlayout;

Default No targets are displayed.

Interactions For this option to take effect, the RESPONSE= argument must also
be used.

If the GROUP= option is used and GROUPDISPLAY= STACK, then

this option is ignored.

Tip The target color is that of the bar outline.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If
this option is used, then the information specified replaces all of the information
that is displayed by default. You can specify roles for columns that do not
contribute to the bar chart along with roles that do.

(role-list)
an ordered, space-separated list of unique BARCHARTPARM roles and user-
defined roles.

BARCHARTPARM roles include CATEGORY or X, RESPONSE or Y,

COLORRESPONSE, ERRORUPPER, ERRORLOWER, INDEX, GROUP, and
DATALABEL.

Notes For the category and response roles, the TIP= option recognizes only
the category and response arguments that you use in the
BARCHARTPARM statement. If you use the CATEGORY= and
RESPONSE= arguments, then you must specify roles CATEGORY and
RESPONSE. Conversely, if you use the X= and Y= arguments, then
you must specify roles X and Y.

The COLORRESPONSE role is valid starting with SAS 9.4M3.

Tip Use the ROLENAME= option to define user-defined roles.

336 Chapter 6 / Plot Statements

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: CATEGORY or X, RESPONSE or Y,
COLORRESPONSE, ERRORLOWER, ERRORUPPER, and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

Example The following example displays data tips for the columns assigned
to the roles CATEGORY and RESPONSE as well as for the column
Pct. The Pct column is not assigned to any pre-defined bar chart
role, so it must first be assigned a role.

ROLENAME=(TIP1=PCT)
TIP=(TIP1 CATEGORY RESPONSE)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
 BARCHARTPARM Statement 337

TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when the bar is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each bar that is to have an active link.

Example http://www.sas.com

Requirements To generate selectable bars, you must include an ODS GRAPHICS

ON statement that specifies the IMAGEMAP option, and you
must write the output to the ODS HTML destination.

For non-grouped data, the values of the column are expected to

be same for each unique category value. If they are not, then only
the first URL value for a given category value is used.

For grouped data, the values of the column are expected to be the
same for each unique category and GROUP combination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some category values, meaning
that no action is taken when the bars for those category values
are selected.

The URL value can be the same for different category values,
meaning that the same action is taken when the bars for those
category values are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET=
on page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.
338 Chapter 6 / Plot Statements

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Description
The input data for the BARCHARTPARM statement is expected to be pre-
summarized, with appropriate summarization statistics (sum, mean, and so on)
computed for the response column. When the category values are not unique, the
display is not meaningful. Prior to SAS 9.4M3, the category axis must be discrete.
Starting with SAS 9.4M3, the category axis can be discrete, linear, or time. The
response axis in all cases is interval.

When the chart is oriented vertically, the X (or X2) axis is used for category and the
Y (or Y2) axis is used for response. When the chart is oriented horizontally, the X
(or X2) axis is used for response and the Y (or Y2) axis is used for category. (See
ORIENT= on page 330.) If the chart is the primary chart, then any charts or plots
that are overlaid with it must have similar axis types.

By default, if the CATEGORY= column is character, then the bars in the chart
appear in the order in which the category values are present in the input data. If the
CATEGORY= column is numeric, then the values are presented in ascending order.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables
side by side on a common axis. By default within an overlay-type layout, if multiple
BARCHART statements are used with different response variables, then the bars
for matching category values are centered on the midpoints and the bars are
superimposed. To make it easier to distinguish among superimposed bars, you can
assign a different BARWIDTH= setting to each BARCHARTPARM statement in the
overlay:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

 BARCHARTPARM Statement 339

legendlabel="A" barwidth=0.8 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B" barwidth=0.6 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C" barwidth=0.4 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;

To place the different response values side by side, you can assign a different offset
to each BARCHARTPARM statement. The BARWIDTH= option can be used with
DISCRETEOFFSET= to create narrower bars that require less width within the plot
area:
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

legendlabel="A"
discreteoffset=-0.3 barwidth=0.3 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B"
discreteoffset=0 barwidth=0.3 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C"
discreteoffset=+0.3 barwidth=0.3 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
340 Chapter 6 / Plot Statements

Different combinations of DISCRETEOFFSET and BARWIDTH can be used to get

the effect that you want. Gaps can be created between bars by providing a
narrower bar width. Or, bars can be overlapped if the bar widths are increased in
proportion to the discrete offset.
layout overlay / cycleattrs=true
xaxisopts=(display=(tickvalues))
yaxisopts=(label="Revenue" offsetmax=0.2);

barchartparm category=year response=A_revenue / name="A"

legendlabel="A" datatransparency=0.2
discreteoffset=-0.2 barwidth=0.5 ;
barchartparm category=year response=B_revenue / name="B"
legendlabel="B" datatransparency=0.2
discreteoffset=0 barwidth=0.5 ;
barchartparm category=year response=C_revenue / name="C"
legendlabel="C" datatransparency=0.2
discreteoffset=+0.2 barwidth=0.5 ;

discretelegend "A" "B" "C" / title="Product:"

location=inside halign=right valign=top;
endlayout;
 BARCHARTPARM Statement 341

Example: BARCHARTPARM Statement

The following graph was generated by the “Example Program” on page 341:

Example Program
proc template;
define statgraph barchartparm;
begingraph;
entrytitle "Average Mileage by Vehicle Type";
entryfootnote halign=left
"Error bars show +/- 1 Standard Error";
layout overlay;
barchartparm category=type response=mean /
errorlower=eval(mean-stderr)
errorupper=eval(mean+stderr) ;
endlayout;
endgraph;
end;
run;

/* create summarized data for barchartparm */

proc summary data=sashelp.cars nway;
class type;
var mpg_highway;
output out=mileage mean=mean stderr=stderr ;
run;
 342 Chapter 6 / Plot Statements

proc sgrender data=mileage template=barchartparm;

run;

BIHISTOGRAM3DPARM Statement
Creates a three-dimensional bivariate histogram of three variables X, Y, and Z, where the values of X
and Y have been gridded. The Z variable represents a response value for the frequency, percentage
counts, or densities of each bin combination.

Restriction: BIHISTOGRAM3DPARM does not support the data tips that are enabled by the
IMAGEMAP= option in the ODS GRAPHICS statement.
Requirements: The input data must be binned by both X and Y. That is, the values for X column
and Y column must form a complete rectangular grid of bins. Input data with non-
binned columns should be preprocessed with the KDE procedure (SAS/STAT),
which enables you to set the number of bins for X and Y, or with a technique similar
to that used in “Example: BIHISTOGRAM3DPARM Statement” on page 346.
The BIHISTOGRAM3DPARM statement must be specified within a LAYOUT
OVERLAY3D statement and cannot be nested under an OVERLAY,
OVERLAYEQUATED, or PROTOTYPE layout.
The input data for Z= column must be nonnegative.
Notes: In the plot display, the direction of the Z axis is upward rather than outward.
If the plot data is a CAS in-memory table, the data is sorted automatically by X and
Y before it is plotted.
See: The KDE Procedure in SAS/STAT User’s Guide

Syntax
BIHISTOGRAM3DPARM X=numeric-column | expression
Y=numeric-column | expression
Z=non-negative-numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the bins.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bin features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bins.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bin outlines.
 BIHISTOGRAM3DPARM Statement 343

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower
endpoints, or upper endpoints of the bins.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input Y values represent midpoints, lower
endpoints, or upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.

Label options
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the bin location of the numeric X values.

Y=numeric-column | expression
specifies the bin location of the numeric Y values.

Z=nonnegative-numeric-column | expression
specifies the response values, such as the frequency counts, percentages, or
densities.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.

TRUE
specifies that the ENDLABELS= option determines how the axis ticks and
value labels are displayed.

FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are
set by the ENDLABELS= option are ignored.

Default TRUE
344 Chapter 6 / Plot Statements

Interactions For this option to take effect, this plot must be the primary plot in
the parent OVERLAY3D layout. For more information, see the
PRIMARY= option.

When this option is set to TRUE, some X-axis options that are set
on the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “ENDLABELS=TRUE | FALSE” on page 344

“boolean ” on page xii for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bins.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISPLAY=(display-options) | STANDARD | ALL

specifies which bin features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the bin fill color

OUTLINE
displays the bin outline

STANDARD
specifies FILL

ALL
specifies FILL and OUTLINE

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the bins.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the
parent layout. For more information, see the PRIMARY= option.

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.
 BIHISTOGRAM3DPARM Statement 345

See “boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bins.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bin outlines.

Default The GraphOutlines style element.

Restriction This option uses only the color specification in the style element or
line options. The line pattern and line thickness specifications are
ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

COLOR= on page 1602

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE
346 Chapter 6 / Plot Statements

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Default MIDPOINTS

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input Y values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Default MIDPOINTS

Details
Because the BIHISTOGRAM3DPARM statement does not perform a binning
computation on the input columns, you must pre-bin the data. In addition, input
data for the statement must be binned by both X and Y. That is, the values for the X
column and the Y column must form a complete, rectangular grid of bins. Input data
with non-binned columns should be preprocessed with the KDE procedure
(SAS/STAT), which enables you to set the number of bins for X and Y. Alternatively,
the data can be preprocessed with a technique similar to the example, where a pre-
defined bin width is used.

The bounding cube can be titled, rotated, and zoomed to provide a different
viewpoint. By default, the outline of the bounding cube is displayed and the viewing
rotation angle is 57 degrees, the tilt angle is 20 degrees and the zoom factor is 1.
See the CUBE=, ROTATE=, TILT=, and ZOOM= options of the LAYOUT OVERLAY3D
statement for information about how to change the viewpoint.

The X axis, Y axis, and Z axis are linear by default. You can change axis properties
with the XAXISOPTS=, YAXISOPTS=, and ZAXISOPTS= options of the LAYOUT
OVERLAY3D statement.

Note: When BINAXIS=TRUE, some axis options for the X- and Y-axes might not
apply.

Example: BIHISTOGRAM3DPARM Statement

The following graph was generated by the “Example Program” on page 347:
 BIHISTOGRAM3DPARM Statement 347

Example Program
proc template;
define statgraph bihistogram;
begingraph;
entrytitle "Distribution of Height and Weight";
entryfootnote halign=right "SASHELP.HEART";
layout overlay3d / cube=false zaxisopts=(griddisplay=on);
bihistogram3dparm x=height y=weight z=count / display=all;
endlayout;
endgraph;
end;
run;

data heart;
set sashelp.heart(keep=height weight);
if height ne . and weight ne .;
height=round(height,5);
weight=round(weight,25);
run;

proc summary data=heart nway completetypes;

class height weight;
var height;
output out=stats(keep=height weight count) N=Count;
run;

proc sgrender data=stats template=bihistogram;

run;
 348 Chapter 6 / Plot Statements

BLOCKPLOT Statement
Creates one or more strips of rectangular blocks containing text values. The width of each block
corresponds to specified numeric intervals.

Syntax
BLOCKPLOT X=column | expression
BLOCK=column | expression </options>;

Summary of Optional Arguments

Appearance options
ALTFILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of alternate fills.
BLOCKINDEX=positive-integer-column | expression
specifies indices for remapping the assignment of the color of the block
fills.
CLASS=column | expression
creates a separate block plot for each unique value of the specified
column or expression.
DATATRANSPARENCY=number
specifies the degree of the transparency of the block fill and outline.
DISPLAY=(display-options) | STANDARD | ALL
specifies which block features to display.
EXTENDBLOCKONMISSING=TRUE | FALSE
specifies whether a missing value in the BLOCK column starts a new
block or reverts to the previous nonmissing value.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the block fills.
FILLTYPE=MULTICOLOR | ALTERNATE
specifies how the blocks are filled.
INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing values in the class column are included in the
plot.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the block outlines.
REPEATEDVALUES=TRUE | FALSE
specifies whether contiguous block values that are identical create
separate blocks.

Axes options
 BLOCKPLOT Statement 349

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.

Label options
BLOCKLABEL=column | expression
specifies alternative text to display for the internal block text values.
LABEL="string"
specifies an external label for a single block plot.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the external block label(s).
LABELPOSITION=LEFT | RIGHT | TOP | BOTTOM
specifies the alignment of BLOCK label.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Text options
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the appearance of the internal block text values or the
alternative values that are specified by the BLOCKLABEL= option.
VALUEFITPOLICY=NONE | SHRINK | SPLIT | SPLITALWAYS | TRUNCATE
specifies how text values are adjusted to fit within the containing block.
VALUEHALIGN=LEFT | CENTER | RIGHT | START
specifies the horizontal alignment of the value text within the blocks.
VALUESPLITCHAR="character-list"
specifies one or more characters on which the values can be split, if
needed.
VALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed
values.
VALUEVALIGN=TOP | CENTER | BOTTOM
specifies the vertical alignment of the value text within the blocks.

Required Arguments
X=column | expression
specifies numeric X-axis positions. If the X axis is discrete and the specified
column is numeric, values are treated as numeric-discrete. If the X axis is
interval and the specified X column is numeric, the following guidelines apply:
n If the data is a SAS data set, you must sort the data in ascending order by the
specified X column before you plot it.
n If the data is a CAS in-memory table, you do not need to sort the data. The
data is automatically sorted in ascending order by the X column before it is
plotted.
350 Chapter 6 / Plot Statements

BLOCK=column | expression
specifies a value for each X position. Numeric values are converted to text
strings according to an assigned format or according to BEST6.

Optional Arguments
ALTFILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of alternate fills. This option in conjunction with the
FILLATTRS= option controls fill appearance when FILLTYPE=ALTERNATE.

Default The GraphAltBlock style element.

Requirement FILLTYPE= ALTERNATE must set for this option to have any
effect.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Note If transparency is in effect with fill and the template includes a

DISCRETELEGEND statement with the AUTOALIGN= option, the
block plot bands might appear darker than normal. In that case,
replace the AUTOALIGN= option in the DISCRETELEGEND
statement with HALIGN= or VALIGN=.

Tips The FILLATTRS= option controls the fill color.

To make all block fill areas the same color, set the FILLATTRS=
and ALTFILLATTRS= options to the same value.

The DATATRANSPARENCY= option sets the transparency for the

block fills and the outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
but a different transparency for the alternate block fills. Example:
datatransparency=0.2 altfillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

BLOCKINDEX=positive-integer-column | expression
specifies indices for remapping the assignment of the color of the block fills.

Requirements FILLTYPE= MULTICOLOR must be set for this option to have any
effect.

The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indices for a specific block value must be the same.
Otherwise, the results are unpredictable.
 BLOCKPLOT Statement 351

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style element to use.

If this option is not used, then the color values are mapped in the
order of GraphData1–GraphDataN.

BLOCKLABEL=column | expression
specifies alternative text to display for the internal block text values.

Default The BLOCK= values

Interaction The DISPLAY= option must include VALUES for the alternative text
to appear.

Note The text for each block segment must be the same. Otherwise, the
results are unpredictable.

Tips This option is particularly useful for showing regimes in forecasting

and time series plots.

The font and color attributes for the alternative text are specified by
the VALUEATTRS= option.

CLASS=column | expression
creates a separate block plot for each unique value of the specified column or
expression. Each block plot is labeled externally by the class value.

Interactions The DISPLAY= option must include LABEL for any external labels
to appear.

This option overrides the LABEL= option.

Note If you specify a column in a SAS data set, the visual attributes for
each class value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each class value
are assigned in ascending order of the class column character
values or of unformatted numeric values.

Tip The font and color attributes for the external labels are specified by
the LABELATTRS= option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the block fill and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Notes This option does not affect the block values or labels.

If transparency is in effect with fill and the template includes a

DISCRETELEGEND statement with the AUTOALIGN= option, the block
plot bands might appear darker than normal. In that case, replace the
352 Chapter 6 / Plot Statements

AUTOALIGN= option in the DISCRETELEGEND statement with

HALIGN= or VALIGN=.

Tip The ALTFILLATTRS= option can be used to set transparency for just
the alternate block fills. The FILLATTRS= option can be used to specify
transparency for the block fills. You can combine this option with
ALTFILLATTRS= and with FILLATTRS= to set one transparency for the
outlines but a different transparency for the block fills. Example:
datatransparency=0.2
altfillattrs=(transparency=0.6) fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which block features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the block fill color

LABEL
displays the external block label or labels

OUTLINE
displays the block outline

VALUES
displays internal block values or the alternative block values that are
specified by the BLOCKLABEL= option

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: FILL, LABEL, OUTLINE, and VALUES

Default STANDARD

Tips Use the OUTLINEATTRS=, FILLATTRS=, ALTFILLATTRS=, and

BLOCKINDEX= options to control the appearance of the blocks.

Use the VALUEATTRS= and LABELATTRS= options to control the text

appearance.

EXTENDBLOCKONMISSING=TRUE | FALSE
specifies whether a missing value in the BLOCK column starts a new block or
reverts to the previous nonmissing value.

Default FALSE

Tip When EXTENDBLOCKONMISSING=TRUE, you can set up the input

data for the BLOCK= column with nonmissing values where you expect
the blocks to change and leave the remaining block values missing. For
 BLOCKPLOT Statement 353

an example, see “Example 1: BlockPlot Overlaid with SeriesPlot” on

page 361.

See “boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the block fills. This option in conjunction with the
ALTFILLATTRS= option controls fill appearance when FILLTYPE= ALTERNATE.

Default The GraphBlock style element.

Requirement FILLTYPE= ALTERNATE must set for this option to have any
effect.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Note If transparency is in effect with fill and the template includes a

DISCRETELEGEND statement with the AUTOALIGN= option, the
block plot bands might appear darker than normal. In that case,
replace the AUTOALIGN= option in the DISCRETELEGEND
statement with HALIGN= or VALIGN=.

Tips The ALTFILLATTRS= option controls the alternating fill color.

To make all block fill areas the same color, set the FILLATTRS=
and ALTFILLATTRS= options to the same value.

The DATATRANSPARENCY= option sets the transparency for the

block fills and the outlines. You can combine this option with
TRANSPARENCY= in (fill-options) to set one transparency for the
outlines and a different transparency for the block fills. For
example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLTYPE=MULTICOLOR | ALTERNATE
specifies how the blocks are filled.

MULTICOLOR
fills the blocks with the color specified by the Color attribute of the style
elements GraphData1–GraphDataN and GraphMissing (for missing BLOCK=
values), or the style elements indicated by the BLOCKINDEX= column.

See “BLOCKINDEX=positive-integer-column | expression ” on page 350

ALTERNATE
alternates the block fill between the colors specified by the FILLATTRS= and
ALTFILLATTRS= options.
354 Chapter 6 / Plot Statements

See “FILLATTRS=style-element | style-element (fill-options) | (fill-options) ”

on page 353 and “ALTFILLATTRS=style-element | style-element (fill-
options) | (fill-options) ” on page 350

Default MULTICOLOR

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

When this option is set to ALTERNATE, the block plot does not
support a DISCRETELEGEND entry.

INCLUDEMISSINGCLASS=TRUE | FALSE
specifies whether missing values in the class column are included in the plot.
Missing class values are included by default. When the data contains missing
class values, the label for those values is either blank for missing character
values or a dot for missing numeric values.

The following figure shows block plots for classes Class 1, Class 2, and any
missing class values.

The following figure shows an X-axis axis table that displays values for classes
Class 1, Class 2, and any missing class values.

Notice that the label for the missing class values is blank. You can use the
INCLUDEMISSINGCLASS=FALSE option to exclude the missing class values. If
you want to keep the missing class values, then you can create a format that
specifies a more meaningful label for the missing class. For example, here is a
format that specifies a label for missing character and numeric class values.
proc format;
value $missingClass " " = "(Missing)";
value missingClass . = "(Missing)";
run;

A single space enclosed in quotation marks specifies a missing character value

and a dot specifies a missing numeric value. Although it might seem appropriate
to use empty quotation marks ('' or "") to specify a missing character value,
doing so produces unexpected results. To specify a missing character value,
enclose a single space in quotation marks (' ' or " "). You can use this format for
the class columns in the PROC SGRENDER statement. In that case, if the class
columns contain missing values, then the labels specified in the format
statement are used for the missing classes.

The following figure shows the previous example when format $missingClass is
applied to the class variable.
 BLOCKPLOT Statement 355

Note: In SAS 9.4M2 and in earlier releases, ODS Graphics does not support
Unicode values in user-defined formats. Starting with SAS 9.4M3, ODS Graphics
supports Unicode values in user-defined formats only if they are preceded by
the (*ESC*) escape sequence. Example: "(*ESC*){unicode beta}". ODS
Graphics does not support an escape character that is defined in an ODS
ESCAPECHAR statement in user-defined formats.

Default TRUE

Interaction The CLASS= option must be specified for this option to have any
effect.

See “boolean ” on page xii for other Boolean values that you can use.

LABEL="string"
specifies an external label for a single block plot.

Defaults The label of the BLOCK= column.

The name of the BLOCK= column, if there is no column label.

Interactions If the CLASS= option is specified, then this option is ignored.

The DISPLAY= option must include LABEL for any external label(s)
to appear.

Tip The font and color attributes for the external label are specified by
the LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the external block label(s).

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all
of the font properties such as color, family, size, weight, style, then
the non-specified properties are derived from the GraphLabelText
style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELPOSITION=LEFT | RIGHT | TOP | BOTTOM

specifies the alignment of BLOCK label.
356 Chapter 6 / Plot Statements

LEFT
places the label to the left of the strip of block values.

RIGHT
places the label to the right of the strip of block values.

TOP
places the label above the strip of block values.

BOTTOM
places the label below the strip of block values.

Default LEFT

NAME="string"
assigns a name to this plot statement for reference in other template
statements. This option is used mostly in the DISCRETELEGEND statement in
order to coordinate the use of colors and line patterns between the plot and the
legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the block outlines.

Default The GraphOutlines style element.

Interactions For this option to have any effect, the outlines must be enabled by
the ODS style or the DISPLAY= option.

If labels are displayed in the TOP or BOTTOM position, then they

are also outlined.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

REPEATEDVALUES=TRUE | FALSE
specifies whether contiguous block values that are identical create separate
blocks.

FALSE
creates only one block when two or more identical block values appear
consecutively.

TRUE
creates one block for each identical value when two or more identical values
appear consecutively.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.
 BLOCKPLOT Statement 357

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the appearance of the internal block text values or the alternative
values that are specified by the BLOCKLABEL= option.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
of the font properties such as color, family, size, weight, style, then
the non-specified properties are derived from the GraphValueText
style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALUEFITPOLICY=NONE | SHRINK | SPLIT | SPLITALWAYS | TRUNCATE

specifies how text values are adjusted to fit within the containing block.

NONE
makes no attempt to fit values that collide.

SHRINK
reduces the font size of the values until they all fit.

SPLIT
splits a value that does not fit within the containing block at a split character
in order to make the value fit the available space.

No split occurs at split characters that occur where a split is not needed. If
the value does not contain any of the specified split characters, then a split
does not occur.

Interaction The VALUESPLITCHAR= option specifies the split characters.

Tip Use the VALUEHALIGN= option to control the alignment of the

split lines.

SPLITALWAYS
splits each value at every occurrence of a split character.

If the value does not contain any of the specified split characters, then a
split does not occur.

Interaction The VALUESPLITCHAR= option specifies the split characters.

Tip Use the VALUEHALIGN= option to control the alignment of the

split lines.

TRUNCATE
truncates any value that does not fit.

For a numeric column, an asterisk (*) is substituted for the entire value
whenever truncation occurs. For a character column, the truncated portion
of the text is replaced by an ellipsis (...).

Default TRUNCATE
358 Chapter 6 / Plot Statements

Interaction The SPLIT and SPLITALWAYS policies are ignored when the
BLOCKPLOT statement is placed in a DATALATTICE or DATAPANEL
layout and the BLOCKPLOT CLASS= option is set. In that case, the
TRUNCATE fit policy is used instead.

VALUEHALIGN=LEFT | CENTER | RIGHT | START

specifies the horizontal alignment of the value text within the blocks.

LEFT
left-aligned within the block

CENTER
center-aligned within the block

RIGHT
right-aligned within the block

START
center-aligned at the starting value of the block

Default LEFT

Restriction When the BLOCKPLOT statement is placed inside an

INNERMARGIN block, only VALUEALIGN=CENTER is honored.

Requirement For this option to have any effect, the DISPLAY= option must
include VALUE.

Interaction When REPEATEDVALUES= TRUE and X values are numeric, only

CENTER and START can be used for a discrete axis. In addition,
only LEFT and START can be used for a linear or log axis. For
example, if REPEATEDVALUES=TRUE and the axis is discrete,
then a setting of RIGHT for this option is ignored and LEFT is used
instead.

VALUESPLITCHAR="character-list"
specifies one or more characters on which the values can be split, if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the value. In that case, all of the specified split characters
together are treated as a single split character.

When VALUESPLITPOLICY=SPLIT, if a value collision is detected, then the

value is split at each occurrence of any of the specified split characters. When
VALUESPLITPOLICY=SPLITALWAYS, the value is split unconditionally on each
of the specified split characters. If the value does not contain any of the
specified split characters, then the value is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

 BLOCKPLOT Statement 359

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
valuesplitchar="abc"

The VALUEFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The VALUESPLITCHARDROP= option specifies whether the split

characters are included in the displayed value or are dropped.

The VALUEFITPOLICY= option sets the policy that is used to

manage the split behavior of the value.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

VALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed values.

TRUE
drops a split character from the value when a split occurs at that character.

Split characters at which a split does not occur are left in place. The
VALUEFITPOLICY= option determines where the values are split. When
VALUEFITPOLICY=SPLIT, each value is split at a split character only where
a split is needed in order to make the value fit the available space. At each
split point, the split character is dropped, and the characters that follow the
split character, up to but not including the split character at the next split
point, are wrapped to the following line.

When VALUEFITPOLICY=SPLITALWAYS, each value is split at every

instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next
split character, are wrapped to the next line.

The following figure shows how value Product*Group*1 is split when the
VALUESPLITCHARDROP=TRUE, VALUESPLITCHAR="*", and
VALUEHALIGN=CENTER options are specified with the SPLIT and
SPLITALWAYS fit policies.

In this example, when VALUEFITPOLICY=SPLIT, the value is split at the first

occurrence of the asterisk in order to make the value fit. No split is needed at
the second asterisk. The first asterisk is dropped, and Group*1 wraps to the
next line. Notice that the second asterisk is not dropped in this case. When
360 Chapter 6 / Plot Statements

VALUEFITPOLICY=SPLITALWAYS, the value is split at every occurrence of

the asterisk. In this case, both asterisks are dropped, and the characters that
follow each asterisk wrap to the next line.

FALSE
includes the split characters in the value display.

The VALUEFITPOLICY= option determines how the split characters are

displayed. When VALUEFITPOLICY=SPLIT, each value is split at a split
character only where a split is needed in order to make the value fit the
available space. A split might not occur at every split character in the value.
At each split point, the split character remains as the last character in the
current line. The characters that follow the split character, up to and
including the split character at the next split point, are then wrapped to the
following line. This process repeats until the entire value is displayed.

When VALUEFITPOLICY=SPLITALWAYS, each value is split at every

instance of a split character in the value regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line.

The following figure shows how value Product*Group*1 is split when the
VALUESPLITCHARDROP=FALSE, VALUESPLITCHAR="*", and
VALUEHALIGN=CENTER options are specified with the SPLIT and
SPLITALWAYS fit policies.

In this example, when VALUEFITPOLICY=SPLIT, the label is split at the first

occurrence of the asterisk to make the value fit. No split is needed at the
second asterisk. The characters that follow the first asterisk wrap to the
next line. When VALUEFITPOLICY=SPLITALWAYS, the value is split at
every occurrence of the asterisk. Each asterisk remains as the last character
in the current line, and the characters that follow are wrapped to the next
line.

Default TRUE. A split character is dropped from the value display when a
split occurs at that character.

Requirement The VALUEFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The VALUESPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

VALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the value text within the blocks.
 BLOCKPLOT Statement 361

Default CENTER

Interaction For this option to have any effect, the DISPLAY= option must
include VALUE.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
A block plot contains information about X locations and text values to be
associated with corresponding intervals along the X axis. There is no Y-axis
information in this plot.

Examples

Example 1: BlockPlot Overlaid with SeriesPlot

When overlaid with plots that have a Y axis (a series plot for example), a block plot
expands vertically to fill the Y-axis range. In the BLOCKPLOT statement, the
BLOCK= argument can be used to reference text values from a column. The
resulting graph displays those values within the plot wall. This example shows how
a block plot can be specified with a series plot within an OVERLAY layout. Here is
the output for this example.
 362 Chapter 6 / Plot Statements

Example Output

Input Data
To prepare input data for the graph, “event” information must be added to the
existing data for stock prices. Here is the SAS code that generates the data for this
example. Notice that the first DATA step creates a Release column. That column is
later specified on the BLOCK= argument to display text values on the wall of the
block plot.
data MSevents;
input Date date9. Release $5.;
label Release="Windows Release";
datalines;
01jun1990 3.0
01sep1995 95
01jul1998 98
01mar2000 2000
01nov2001 XP
;

proc sort data=sashelp.stocks(keep=date stock close)

out=MSstock;
where stock="Microsoft";
by date;
run;

data events;
 BLOCKPLOT Statement 363

merge MSstock MSevents;

by date;
run;

In the merged input data set, the Release column value is missing for each
observation until the first event defined in the MSEvents data, 3.0, occurs in June of
1990. The Release column value is missing again for the subsequent observations
until the next event defined in the MSEvents data, 95, occurs in September of 1995.
Here is a sample of the merged data.
Obs Stock Date Close Release
...
45 Microsoft 02APR90 $58.00
46 Microsoft 01MAY90 $73.00
47 Microsoft 01JUN90 $76.00 3.0
48 Microsoft 02JUL90 $66.50
49 Microsoft 01AUG90 $61.50
50 Microsoft 04SEP90 $63.00
...

Example Program
Here is the code for this example.
proc template;
define statgraph blockplot1;
begingraph;
entrytitle "Microsoft Share Prices";
entrytitle "and Significant OS Releases";
layout overlay;
blockplot x=date block=release /
datatransparency=0.3 valuevalign=top
labelposition=top display=(fill values label)
extendblockonmissing=true ;
seriesplot x=date y=close;
endlayout;
endgraph;
end;
run;

proc sgrender data=events template=blockplot1;

format date year4.;
label date="Year";
run;

Details
In the GTL template code for this example, BLOCK=RELEASE is specified in the
BLOCKPLOT statement so that the RELEASE values are displayed on the wall of
the resulting block plot. In this template, the BLOCKPLOT statement sets
 364 Chapter 6 / Plot Statements

EXTENDBLOCKONMISSING= TRUE so that missing values in the data revert to the

previous nonmissing value in the block plot. Thus, in the block plot, values are
missing until 01JUN90, when the value changes from missing to 3.0. The block plot
retains that 3.0 value for subsequent observations until the next nonmissing value
replaces it (in this case, the value95 on 01SEP95). In the example output shown in
“Example Output” on page 362, the fill color for each of the nonmissing values is
determined by the style elements GraphData1–GraphData5 in this case. The fill
color for the missing values is determined by the next available GraphDataN style
element, which is GraphData6 in this case.

Example 2: Standalone BlockPlot in Lattice Layout

When used as a standalone plot in a LATTICE layout, the block plot’s height can be
controlled. Using this technique, it is possible to include two or more “event” strips
in a plot. This example demonstrates this technique by creating a series plot with
an event strip as shown in the following figure.

Example Output

Example Program
In this example, the lattice ROWWEIGHTS=(0.04 0.96) option apportions 4% of the
vertical space to the block plot. Here is the code for this example.
proc template;
 BOXPLOT Statement 365

define statgraph blockplot2;

begingraph;
entrytitle "Microsoft Share Prices";
entrytitle "and Significant OS Releases";
layout lattice / rowweights=(0.04 0.96);
blockplot x=date block=release / datatransparency=0.3
valuefitpolicy=shrink labelposition=left
display=(fill label outline values)
extendblockonmissing=true;
seriesplot x=date y=close;
endlayout;
endgraph;
end;
run;

proc sgrender data=events template=blockplot2;

format date year4.;
label date="Year";
run;

BOXPLOT Statement
Creates box plots that are computed from input data.

Tip: Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
BOXPLOT Y=numeric-column | expression < /options >;

BOXPLOT X=column | expression

Y=numeric-column | expression < /option(s) >;

Summary of Optional Arguments

Appearance options
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.
CAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the whisker caps.
CAPSHAPE=SERIF | LINE | BRACKET | NONE
366 Chapter 6 / Plot Statements

specifies the shape at the ends of the whiskers.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX
specifies that a connect line joins a statistic from box to box.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the lines connecting multiple boxes.
CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled boxes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill,
whiskers, mean, median, caps, connect lines, outliers, and data labels, if
displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which box-plot features to display.
EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the boxes.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of
the GraphData1–GranphDataN style elements.
INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is
specified.
MEANATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the marker representing the mean within the
box.
MEDIANATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line representing the median within the
box.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.
OUTLIERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the markers representing the outliers.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the box outline.
SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid
overlap.
WHISKERATTRS=style-element | style-element (line-options) | (line-options)
 BOXPLOT Statement 367

specifies the line properties of the whiskers and caps.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

OUTLIERTIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over an
outlier.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
box or whisker in the box plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column
specifies the labels of the outliers. Either a numeric or a character column
can be used.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the outlier labels.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
368 Chapter 6 / Plot Statements

specifies an amount to offset all boxes from the discrete X ticks.

GROUP=column | discrete-attr-var | expression
creates a box plot for each unique group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)
specifies the statistics to be displayed for each box.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.
PERCENTILE=1 | 2 | 3 | 4 | 5
specifies one of five definitions used to calculate percentiles.
WEIGHT=numeric-column | expression
specifies a column that contains a statistics calculation a priori weight for
each observation of the input data object.
WHISKERPERCENTILE=number
specifies the whisker length, in percentile units.

Required Arguments
Specifying only Y= creates a single box plot. Specifying both X= and Y= creates a box plot
for each unique value of X.
Y=numeric-column | expression
specifies the column for the Y values. This argument is required.

X=column | expression
specifies the column for the X values. This argument is required if you want to
create a box plot for each unique X value.

Note For interval X values, if a user-defined format is applied to the X column,

the format should map each X value to only one unique formatted value.
Otherwise, unexpected results might occur.

Optional Arguments
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.
 BOXPLOT Statement 369

Defaults For a discrete category axis, 0.4 for a nongrouped box plot or 0.6
for a grouped box plot.

For an interval category axis, 85% of the smallest interval between

any two boxes for the given plot.

Range 0–1, where 0 is the narrowest and 1 is the widest.

Interactions For a grouped box plot with a discrete category axis when
GROUPDISPLAY=CLUSTER, the box width is a percentage of the
cluster width. See CLUSTERWIDTH.

Prior to SAS 9.4M3, for an interval box plot, this option is ignored,
and the box width is controlled by the INTERVALBOXWIDTH=
option. Starting with SAS 9.4M3, for an interval box plot, this option
is ignored when the INTERVALBOXWIDTH= option is specified.
Otherwise, the BOXWIDTH= option is honored.

CAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the whisker caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1

Interaction This option is ignored when CAPSHAPE=NONE is in effect.

Tips Use the CAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the WHISKERATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the caps:

capscale=2

Reduce the size of the caps by 50%:

capscale=0.5

CAPSHAPE=SERIF | LINE | BRACKET | NONE

specifies the shape at the ends of the whiskers.

SERIF
specifies a short line perpendicular to the whisker.

LINE
specifies a line perpendicular to the whisker that extends the width of the
box.

BRACKET
specifies a line perpendicular to the whiskers that extends the width of the
box and that has short extensions at each end.

The extensions are drawn in the direction of the box.

370 Chapter 6 / Plot Statements

NONE
specifies that no shape appears at the ends of the whiskers.

The following figure shows each of the shapes.

Default SERIF

Interactions The cap color and the thickness are specified by the
WHISKERATTRS= option. The cap pattern is always solid.

The DISPLAY= option must include CAPS in order for cap lines to
be shown.

Starting with SAS 9.4M5, when CAPSCALE= is in effect, the display

width for the caps is the cap's default width multiplied by the value
specified in CAPSCALE=.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.

Default 0.7

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width
 BOXPLOT Statement 371

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Note When the X axis is an interval axis, the cluster width is a fraction
of the smallest data interval.

CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX

specifies that a connect line joins a statistic from box to box.

Default The GraphBox:Connect style reference.

Requirement The DISPLAY= option must contain the CONNECT display-options

value in order for the connect line to be displayed.

Interaction This option applies only when the X= argument is used to generate
multiple boxes.

Note Starting with SAS 9.4M3, the connect lines are drawn in axis order.
In prior releases, they are drawn in data order.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the lines connecting multiple boxes.

Default The GraphConnectLine style element.

Interactions If there is only one box, then this option is ignored.

If the DISPLAY= option does not include CONNECT, or if the

GROUP= option is used, then this option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement DISPLAY= must include CONNECT for this option to have any
effect.

Interaction This option is ignored when the GROUP= option is in effect.

DATALABEL=column
specifies the labels of the outliers. Either a numeric or a character column can
be used.

Default No data labels are displayed

372 Chapter 6 / Plot Statements

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See LABELFAR= option

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the outlier labels.

Default The GraphDataText style element.

Interactions This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

If one or more label options are specified and they do not include
all the font properties (color, family, size, weight, style), then the
non-specified properties are derived from the GraphDataText style
element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When
set to TRUE, the data label is split unconditionally at each occurrence of any of
the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

 BOXPLOT Statement 373

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.
374 Chapter 6 / Plot Statements

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

 BOXPLOT Statement 375

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled boxes. The following figure shows
boxes with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill, whiskers,
mean, median, caps, connect lines, outliers, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
interior fill area of the boxes. You can combine this option with
376 Chapter 6 / Plot Statements

FILLATTRS= to set one transparency for the box outlines and the
whiskers, mean, median, caps, and connect lines, but a different
transparency for the box fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.

Default 0 (no offset, all boxes are centered on the discrete ticks)

Range -0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right when
ORIENT=VERTICAL, and up when ORIENT=HORIZONTAL. (If the
layout's axis options set REVERSE=TRUE, then the offset direction
is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

See “About the DISCRETEOFFSET= Option” on page 400

DISPLAY=(display-options) | STANDARD | ALL

specifies which box-plot features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

CAPS
displays caps at the ends of the whiskers.

CONNECT
displays the line connecting multiple boxes.

FILL
displays the box fill color.

FILLPATTERN
displays the box fill pattern.

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the
boxes.
 BOXPLOT Statement 377

MEAN
displays the mean symbol within the box.

MEDIAN
displays the median line within the box.

NOTCHES
displays notched boxes.

See “Notches” on page 399.

OUTLIERS
displays markers for the outliers.

STANDARD
specifies CAPS, FILL, MEAN, MEDIAN, and OUTLIERS.

ALL
specifies all features: CAPS, CONNECT, FILL, FILLPATTERN, MEAN,
MEDIAN, NOTCHES, and OUTLIERS.

Default The GraphBox:DisplayOpts style reference. If this style element

does not exist, then the default is STANDARD.

Interaction If EXTREME= TRUE, then the OUTLIERS feature is ignored

Notes Starting with SAS 9.4M3, connect lines are drawn in axis order. They
are drawn in data order in prior releases.

Starting with SAS 9.4M3, when DISPLAY= includes MEAN, the

BOXPLOT statement contributes its mean markers to a discrete
legend when TYPE=MARKER is in effect in the DISCRETELEGEND
statement.

Tips To control the appearance of these features, use the

CONNECTATTRS= , FILLATTRS= , MEANATTRS= ,
MEDIANATTRS= , OUTLIERATTRS= , and WHISKERATTRS=
options. The WHISKERATTRS= option controls affects both CAPS
and WHISKERS. Starting with SAS 9.4M5, use the
FILLPATTERNATTRS= on page 380 option to control the
appearance of the box fill pattern.

Regardless of which display options are being displayed, this option

does not affect the axis range.

DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)

specifies the statistics to be displayed for each box.

NONE
does not display any statistics.

STANDARD
displays N, MEAN, and STD.

ALL
displays all available statistics.
378 Chapter 6 / Plot Statements

(statistics-list)
specifies a space-separated list of one or more statistics, enclosed in
parentheses.

The list can include one or more of the following:

DATAMAX
maximum data value that includes not only the maximum whisker values
but also the maximum outlier values.

This option is valid in SAS 9.4M1 and later releases. The DATAMAX value
is greater than or equal to the MAX value and is always represented in
the axis range.

DATAMIN
minimum data value that includes not only the minimum whisker values
but also the minimum outlier values.

This option is valid in SAS 9.4M1 and later releases. The DATAMIN value
is less than or equal to the MIN value and is always represented in the
axis range.

IQR
interquartile range (Q3–Q1).

MAX
maximum data value below the box upper fence.

MEAN
mean data value for the box.

MEDIAN
median data value for the box.

MIN
minimum data value above the box lower fence.

N
number of observations for the box.

Q1
lower quartile (25th percentile) for the box.

Q3
upper quartile (75th percentile) for the box.

RANGE
range of the data (MAX–MIN).

STD
standard deviation of the data for the box.

SUMWGT
sum of the weights for the box.

This option is valid in SAS 9.4M1 and later releases.

Default NONE
 BOXPLOT Statement 379

Restrictions Prior to SAS 9.4M5, this option is ignored if ORIENT=HORIZONTAL.

Starting with SAS 9.4M5, statistics can be displayed for both
vertical and horizontal box plots.

This option is ignored if the GROUP= option is also specified.

Note The notches in the box plot can extend beyond DATAMIN and
DATAMAX in some cases.

EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences. Fences are
locations above and below the box. The upper and lower fences are located at a
distance 1.5 times the Interquartile Range (IQR) (IQR = Q3 - Q1). The upper and
lower far fences are located at a distance 3 times the IQR .

FALSE
specifies that whiskers be drawn from the upper edge of the box to the
largest value within the upper fence, and from the lower edge of the box to
the smallest value within the lower fence.

This representation is sometime called a schematic box plot or a Tukey box

plot.

TRUE
specifies that whiskers be drawn to the largest and smallest data values,
whether these values are inside or outside the fences.

The outliers and far outliers are not displayed and are not labeled. This
representation is sometime called a skeletal box and whisker plot.

Default FALSE

Interaction This option overrides the DATALABEL= , DATALABELATTRS= ,

LABELFAR= , OUTLIERATTRS= , and SPREAD= options.

See “Statement Summary” on page 391

“boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the boxes.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of the GraphData1–

GraphDataN style elements.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tip The DATATRANSPARENCY option sets the transparency for the box
outlines, box fill, whiskers, mean, median, caps, connect lines, and
outliers. You can combine this option with DATATRANSPARENCY=
to set one transparency for the box outlines and the whiskers, mean,
median, caps, and connect lines, but a different transparency for the
box fills. Example:
380 Chapter 6 / Plot Statements

datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation. The marker for that observation is drawn n times,
which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a box plot for each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.
 BOXPLOT Statement 381

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct box is created for each group value by varying the visual attributes of
the box display features. The display features are controlled by the current ODS
style or by the DISPLAY= option. The default group appearance for each display
feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The Color attribute controls the fill color, and the ContrastColor
attribute controls the outline color. The box outline line style is
solid.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern
and box outline. The box outline line style is solid. If the active
ODS style does not specify FillPattern, fill patterns are not
displayed by default.

Median, caps, ContrastColor attribute of a GraphData1–GraphDataN style

and notches element or the GraphMissing style element with a solid line
style.

Mean and MarkerSymbol and ContrastColor attributes of a GraphData1–

outliers GraphDataN style element or the GraphMissing style element.

1 The default display features are determined by the DisplayOpts attribute of the GraphBox style
element.
2 Fill patterns are valid for box plots starting with SAS 9.4M5.

Interactions Connect lines are not drawn for grouped data.

The box plot display depends on the setting for the

GROUPDISPLAY= option.

By default, the group values are mapped in the order of the data.
The GROUPORDER= option can be used to control the sorting
order of the group values. The INDEX= option can be used to alter
the default sequence of colors and markers.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.
382 Chapter 6 / Plot Statements

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the box outlines, but the
PATTERN= setting in the OUTLIERATTRS= option could be used to
assign the same line pattern to all box outlines and connect lines.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “Example 2: Box Plot of City MPG and Vehicle Type Grouped by
Origin” on page 403

“DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs. The following example shows a box plot with
GROUPDISPLAY=CLUSTER:

OVERLAY
draws boxes for a given group value at the exact coordinate.

Depending on the data, boxes at a given coordinate might overlap.

CLUSTER
draws boxes for a given group value adjacent to each other.

Default OVERLAY

Interaction This option is ignored unless GROUP= is specified.

Tip Use the CLUSTERWIDTH= option to control the width of the

clusters when CLUSTER is in effect.
 BOXPLOT Statement 383

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
384 Chapter 6 / Plot Statements

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.

AUTO
the system determines the box width.

Prior to SAS 9.4M3, AUTO uses 85% of the smallest interval between any
two boxes for the given plot. Starting with SAS 9.4M3, if the BOXWIDTH=
option is specified, AUTO uses the BOXWIDTH= option value. Otherwise, it
uses 85% of the smallest interval between any two boxes for the given plot.

uses one of the following box-width values listed in the order of precedence:
 BOXPLOT Statement 385

dimension
sets the box width to the specified value.

Default AUTO

Requirement The category axis (X by default) must be an interval axis for the
INTERVALBOXWIDTH= option to have any effect.

Interaction Prior to SAS 9.4M3, this option controls the box width for an
interval box plot. Starting with SAS 9.4M3, this option overrides
the BOXWIDTH= option for an interval box plot.

Tip The box width that you specify with this option is honored even if
the boxes overlap. If the boxes overlap, reduce the interval box
width or use the BOXWIDTH= option instead.

See “dimension” on page xiii

LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.

FALSE
applies the labels specified by the DATALABEL= option to both the outliers
and the far outliers.

TRUE
applies the labels specified by the DATALABEL= option to the far outliers.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “DATALABEL=column ” on page 412

“Statement Summary” on page 391 for information about outliers.

“boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y= column label. If a label is not defined, then the Y= column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MEANATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the marker representing the mean within the box.

Defaults For non-grouped data, GraphBoxMean style element.

386 Chapter 6 / Plot Statements

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
mean.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MEDIANATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line representing the median within the box.

Defaults For non-grouped data, the GraphBoxMedian style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxMedian style element, and the ContrastColor attribute of
the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
median.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.

Default VERTICAL

OUTLIERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the markers representing the outliers.

Defaults For non-grouped data, GraphOutlier style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.
 BOXPLOT Statement 387

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

OUTLIERTIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over an
outlier. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the box plot
can be specified along with roles that do contribute.

(role-list)
an ordered, space-separated list of unique BOXPLOT roles. BOXPLOT roles
for OUTLIERTIP include X, Y, STAT, and DATALABEL.

Note In the data tip, the STAT role displays the text “outlier” or “far
outlier” as applicable.

Example The following example displays data tips only for the column that
is assigned to the X role:
OUTLIERTIP=(X)

Default The columns assigned to these roles are automatically included in

the data tip information: X and Y .

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that specifies the IMAGEMAP option, and write the
graphs to the ODS HTML destination.

Interaction The labels and formats for the OUTLIERTIP variables can be
controlled with the TIPLABEL= and TIPFORMAT= options.

See the TIP= option for specifying the information to display when the
cursor is positioned over a box or whisker in the box plot.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the box outline.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled boxes, the LineStyle and LineThickness
attributes of the GraphOutlines style element, and the ContrastColor
attribute of the GraphData1–GraphDataN style elements.

For grouped data and unfilled boxes, the LineThickness attribute of

the GraphOutlines style element, and the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.
388 Chapter 6 / Plot Statements

“Line Options” on page 1602 for available line-options.

PERCENTILE=1 | 2 | 3 | 4 | 5
specifies one of five definitions used to calculate percentiles.

Default 5 (empirical distribution function with averaging)

Note The percentile definition and default are the same as those that are
used by the PCTLDEF= option of the UNIVARIATE procedure and the
QNTLDEF= option of the SUMMARY procedure.

See “Calculating Percentiles” on page 392

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid overlap.
For vertical box plots this means offsetting the outliers horizontally. If this
option is false, then outliers with the same value are plotted in the same
position, which means only one is visible.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “boolean ” on page xii for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box or
whisker in the box plot. If this option is used, then it replaces all of the
information that is displayed by default.

(role-list)
an ordered, space-separated list of unique BOXPLOT roles.
 BOXPLOT Statement 389

BOXPLOT roles for TIP include DATAMAX, DATAMIN, MAX, MIN, MEAN,
MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT apply to SAS 9.4M1
and later releases.

Tip Statistics such as N, MIN, and MAX are special roles, They are not
column-based like the X role.

Example The following example displays data tips only for the columns that
are assigned to the roles X (CATEGORY) and the statistic MEAN:
TIP=(X MEAN)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X, N, STD, MIN, MAX, Q1, Q3, MEAN, and
MEDIAN.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

See the OUTLIERTIP= option for specifying the information to display

when the cursor is positioned over an outlier.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(X Y)
TIPFORMAT=(X=4. Y=4.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles.

390 Chapter 6 / Plot Statements

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(X)
TIPLABEL=(X="Type")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles.

WEIGHT=numeric-column | expression
specifies a column that contains a statistics calculation a priori weight for each
observation of the input data object.

Requirement The value must be nonnegative.

Note The values of the weight variable must be greater than zero. If the
weight value for an observation is zero, negative, or missing, that
observation is deleted from the analysis.

WHISKERATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the whiskers and caps.

Defaults For non-grouped data, the GraphBoxWhisker style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxWhisker style element, and the ContrastColor attribute of
the GraphData1–GraphDataN style elements.

Restriction The caps are always drawn with a solid line.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

WHISKERPERCENTILE=number
specifies the whisker length, in percentile units. When this option is specified,
number is used as the low percentile, and 100–number is used as the high
percentile.

Here are some examples of values and their effect:

0 specifies the high and low extremes

10 specifies the 10 th percentile low and the 90th percentile high
25 specifies the 25 th percentile low and the 75th percentile high
 BOXPLOT Statement 391

Default The whiskers are drawn from the box to the most extreme point that is
less than or equal to 1.5 times the IQR

Range 0–25

Notes When this option is specified, fences and far outliers are not drawn.

When this option is set to 25, no whiskers are drawn because the box
extends from the 25 th to the 75 th percentile.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Summary
The BOXPLOT statement displays a single box if given just a Y argument. It
displays multiple boxes if given both Y and X arguments and X has more than one
unique value.

By default for numeric or character columns, the category (X) axis is

TYPE=DISCRETE. You can override the default and set the TYPE= to LINEAR or
TIME in the parent layout, provided that the X column is numeric. The axis for the
analysis (Y) column is always LINEAR. When the X axis is LINEAR, you must use
the INTERVALBOXWIDTH= option to specify the box width.

When ORIENT= VERTICAL, the X (or X2) axis is used for the X column and the Y
(or Y2) axis is used for the Y column. When ORIENT=HORIZONTAL, the X (or X2)
axis is used for the Y column and the Y (or Y2) axis is used for the X column.
392 Chapter 6 / Plot Statements

Two basic box plot representations can be drawn with the BOXPLOT statement: a
schematic (Tukey) box plot and a skeletal box plot. See the EXTREME= option for
details.

Visual Description of Box Plot Percentile Boundaries

Box plots display the distribution of data by using a rectangular box and whiskers.
Whiskers are lines that indicate a data range outside of the box. The following
figure shows the parts of a box plot.

As shown in the figure, the bottom and top edges of the box are located at the 25th
and 75th percentiles of the sample. Within the box, you can display the median
(50th percentile) as a line and the mean as a marker (see DISPLAY= option).

You can also display markers and data labels for outliers. Outliers are observations
that are more extreme than the upper and lower fences (± 1.5 IQR). Outliers that
are beyond upper and lower far fences (± 3 IQR) are called FAR OUTLIERS and can
also be identified and labeled. From a graphical perspective, the locations of fences
along the axis are known, but there is no line or marker that displays a fence. (See
DISPLAY= , LABELFAR= , and DATALABEL= options).

Finally, you can control the range represented by the whiskers. By default, the
whiskers are drawn from the upper edge of the box to the MAX value, and from the
lower edge of the box to the MIN value. (See the EXTREME= option.)

Calculating Percentiles
You can specify one of five definitions for computing the percentiles with the
PERCENTILE= option. Let n be the number of nonmissing values for a variable, and
let X1, X2, ..., Xn represent the ordered values of the variable. X1 is the smallest
 BOXPLOT Statement 393

value, X2 is the next smallest, and Xn is the largest value. Let the tth percentile be y,
set:
t
p= 100

and let:
np = j + g

when PERCENTILE=1, 2, 3, or 5, or let:

(n + 1)p = j + g

when PERCENTILE=4, where j is the integer part of np, and g is the fractional part
of np. Then the PERCENTILE= option defines the tth percentile, y, as described in
the following table:

Percentile Description Equation Notes

1 Weighted average at y = (1 − g)x j + gx j + 1 x0 is taken to be x1

Xnp

2 Observation y = xj Used when g < 2 or

1
numbered closest to
1
np when g = 2 and j is
even

y = xj + 1 1
Used when g = 2 and j
1
is odd or when g > 2

3 Empirical distribution y = xj Used when g = 0

function
y = xj + 1 Used when g > 0

4 Weighted average y = (1 − g)x j + gx j + 1 xn + 1 is taken to be xn

aimed at x(n + 1) p

5 Empirical distribution 1
y = 2 (x j + x j + 1) Used when g = 0
function with
averaging
y = xj + 1 Used when g > 0

Changing Box Plot Display

SAS defines graphical style elements that control the display of box plots
generated with the BOXLOT or BOXPLOTPARM statement. Using these style
elements as a starting point, you can change the style attribute values to achieve a
very different appearance for your box plots. Using the DEFAULT style for an
394 Chapter 6 / Plot Statements

example, here is a portion of the style template for elements that are related to box
plots:
proc template;
define style Default;
...
style GraphBox /
capstyle = "serif"
connect = "mean"
displayopts = "fill caps median mean outliers";
style GraphBoxMean / ...;
style GraphBoxMedian / ...;
style GraphBoxOutlier / ...;
style GraphBoxWhisker / ...;
...
end;
run;

Table 6.1 Style Elements and Their Purpose

Style Element Purpose

GraphBox general box plot properties (see the next table)

GraphBoxMean marker properties of mean marker

(MARKERSYMBOL= , MARKERSIZE=, CONTRASTCOLOR=)

GraphBoxMedian line properties of the median line

(LINESTYLE=, LINETHICKNESS=, CONTRASTCOLOR=)

GraphBoxOutlier marker properties of outliers

(MARKERSYMBOL=, MARKERSIZE=, CONTRASTCOLOR=)

GraphBoxWhisker line properties of whiskers and caps

(LINESTYLE=, LINETHICKNESS=, CONTRASTCOLOR=)

Table 6.2 Attributes and Values for the GraphBox Style Element

Attribute Value(s) Description

CONNECT= "MEAN" | "MEDIAN" | "Q1" | "Q3" | statistic to connect with line

"MIN" | "MAX"

CAPSTYLE= "SERIF" | "LINE" | "BRACKET" shape at ends of whiskers

DISPLAYOPTS= "<CAPS> show caps at end of

whiskers

<FILL> show filled boxes

 BOXPLOT Statement 395

Attribute Value(s) Description

<MEAN> show a marker for the mean

<MEDIAN> show a line for the median

<OUTLIERS> show markers for the

outliers

<CONNECT> show line connecting same

statistic on multiple boxes

<NOTCHES>" show notched boxes

The DisplayOpts attribute of GraphBox lists the general features to be displayed.

The following diagram shows the standard display for box plots, as defined by the
DEFAULT style. The keywords that are related to the appearance features are
annotated:

The two display options that are not the default are CONNECT (show connect
lines) and NOTCHES.
396 Chapter 6 / Plot Statements

The STATISTICAL style is derived from the DEFAULT style and inherits the
GraphBox element from the parent DEFAULT style. The following code generates a
box plot for the STATISTICAL style:
/* Specify a path for the ODS output */
filename odsout "output-path";

proc template;
define statgraph boxplotdef;
begingraph;
entrytitle "Statistical Style";
layout overlay / xaxisopts=(label="Age" type=linear);
boxplot x=ageatstart y=cholesterol / intervalboxwidth=40;
endlayout;
endgraph;
end;

ods graphics / outputfmt=static;

ods _all_ close;
ods html path=odsout file="boxplot.html" style=statistical;

proc sgrender data=sashelp.heart template=boxplotdef;

where ageatstart between 50 and 55;
run;

ods html close;

ods html; /* Not required in SAS Studio */

Here is the output.

 BOXPLOT Statement 397

For this example, the following attributes on the default box plot are changed:
n By default, serif caps are displayed at the end of the fences. Those caps are
removed from the fence lines.
n By default, the boxes are filled. Empty, notched boxes are displayed.

n By default, the mean values are represented by hollow diamonds. Filled

diamonds with a slightly reduced size are displayed.
n By default, the marker symbols for the outliers are hollow black circles. The size
and shape of the marker symbols are changed, and the marker size is reduced.

To make these changes, you can derive a new style from the STATISTICAL style
and set the attributes that you want to change. Any attribute settings that you do
not change are inherited from the parent STATISTICAL style. The following style
template effects the desired changes:
proc template;
define style Boxplot;
parent = styles.statistical;
style GraphBox from GraphBox /
capstyle = "line"
displayopts = "caps median mean outliers notches";
style GraphBoxMean from GraphBoxMean /
markersymbol="diamondfilled"
contrastcolor=GraphColors("gcdata1")
markersize = 5px;
style GraphOutlier from GraphOutlier /
markersize = 5px
markersymbol = "x"
contrastcolor = GraphColors("gcdata2");
end;
run;

Note the following:

n The DEFINE STYLE statement assigns the name BOXPLOT to our new style,
and sets the STATISTICAL style as the parent style.
398 Chapter 6 / Plot Statements

n On the GraphBox style element, the CapStyle attribute is set to LINE, which
removes the serif caps from the end of the fences. The DisplayOpts attribute
drops the FILL value from the display list and adds the NOTCHES value; these
changes determine that the graph displays empty, notched boxes.
n On the GraphBoxMean style element, the marker symbol is changed to a filled
diamond and the marker size is reduced to 5 pixels (the default is 9 pixels). The
ContrastColor attribute is set to GCDATA1 (the default is GCDATA).
n On the GraphBoxOutlier style element, the marker symbol is changed to an X
and the marker size is reduced to 5 pixels (the default is 7 pixels). The
CONTRASTCOLOR= attribute is set to GCDATA2 (the default is GCOUTLIER).

The following code generates a box plot for the BOXPLOT style:
/* Specify a path for the ODS output */
filename odsout "output-path";

proc template;
define statgraph boxplotdef;
begingraph;
entrytitle "Boxplot Style";
layout overlay / xaxisopts=(label="Age" type=linear);
boxplot x=ageatstart y=cholesterol / intervalboxwidth=40;
endlayout;
endgraph;
end;

ods graphics / outputfmt=static;

ods _all_ close;
ods html path=odsout file="notchedboxplot.html" style=Boxplot;

proc sgrender data=sashelp.heart template=boxplotdef;

where ageatstart between 50 and 55;
run;

ods html close;

ods html; /* Not required in SAS Studio */

Here is the output.

 BOXPLOT Statement 399

When making such style changes, remember that you are affecting all box plot
displays for all procedures that produce box plots when this style is in effect. It is
possible to change the box plot appearance for specific procedures, but to do this, a
specific graph template must be modified, not a style template.

For a comprehensive description of the style elements affecting ODS graphics, see
“Graph Style Elements Used by ODS Graphics” in SAS Graph Template Language:
User’s Guide.

Notches
In a notched box plot, the endpoints of the notches are at the following computed
locations:
1.58 IQR
median ±
n

In the equation, IQR (IQR=Q3-Q1) is the interquartile range and n is the sample size.
The medians (central lines) of the two boxes are significantly different at
approximately the 0.05 level if the corresponding notches do not overlap.
400 Chapter 6 / Plot Statements

Maximum observation

th
75 percentile (upper
quartile)

1/2
1.58(IQR) / n
+ Mean
Median

1/2
1.58(IQR) / n
th
25 percentile (lower
quartile)

Minimum observation

Note: In some cases, the notches in the box plot can extend beyond DATAMIN and
DATAMAX.

About the DISCRETEOFFSET= Option

This feature is useful for graphing multiple response variables side by side on a
common axis. By default within an overlay-type layout, if multiple BOXPLOT
statements are used with different analysis variables, then the boxes for matching
X values are centered on the ticks. Depending on the data, the boxes might be
superimposed. The following code fragment shows the default box positioning:
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");
boxplot x=type y=mpg_city / name="City"
legendlabel="City";
boxplot x=type y=mpg_highway / name="Highway"
legendlabel="Highway";
discretelegend "City" "Highway";
endlayout;
 BOXPLOT Statement 401

To place the different response values side by side, you can assign a different offset
to each BOXPLOT statement. The BOXWIDTH= option can be used in conjunction
with the DISCRETEOFFSET= option to create narrower boxes when desired.
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");
boxplot x=type y=mpg_city / name="City"
discreteoffset=0.2 legendlabel="City";
boxplot x=type y=mpg_highway / name="Highway"
discreteoffset=-0.2 legendlabel="Highway";
discretelegend "City" "Highway";
endlayout;
 402 Chapter 6 / Plot Statements

Examples

Example 1: Box Plot of City MPG and Vehicle Type

The following graph was generated by the “Example Program” on page 402:

Example Program
proc template;
define statgraph boxplot;
begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay /
xaxisopts=(offsetmin=0.1 offsetmax=0.1);
boxplot y=mpg_city x=type /
datalabel=make spread=true;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=boxplot;

label type="Vehicle Type";
run;
 BOXPLOTPARM Statement 403

Example 2: Box Plot of City MPG and Vehicle Type Grouped by

Origin
The following graph was generated by the “Example Program” on page 403:

Example Program
proc template;
define statgraph boxplot;
begingraph;
entrytitle "City Mileage and Vehicle Type By Origin";
layout overlay /
xaxisopts=(offsetmin=0.1 offsetmax=0.1);
boxplot y=mpg_city x=type / name="box"
group=origin groupdisplay=cluster
spread=true datalabel=make;
discretelegend "box";
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars template=boxplot;
label type="Vehicle Type";
run;

BOXPLOTPARM Statement
Creates side-by-side box plots specified by parameters.
 404 Chapter 6 / Plot Statements

Requirements: The input data must be precomputed. See “Input Data Requirements for the
BOXPLOTPARM Statement” on page 435.
Nonmissing Y values for statistical observations of Q1 and Q3 are required in order
for a box to be drawn.
The statistical values, if present, must conform to the following rules in order for a
box to be displayed:
Q1 <= MEDIAN <= Q3
MIN <= MAX
STD >= 0
N > 0

Tip: Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
BOXPLOTPARM Y=numeric-column | expression
STAT=string-column < /option(s) >;

BOXPLOTPARM X=column | expression

Y=numeric-column | expression
STAT=string-column < /options >;

Summary of Optional Arguments

Appearance options
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.
CAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the whisker caps.
CAPSHAPE=SERIF | LINE | BRACKET | NONE
specifies the shape at the ends of the whiskers.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX
specifies that a connect line joins a statistic from box to box.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the lines connecting multiple boxes.
CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.
 BOXPLOTPARM Statement 405

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled boxes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill,
whiskers, mean, median, caps, connect lines, outliers, and data labels, if
displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which box-plot features to display.
EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the boxes.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of
the GraphData1–GranphDataN style elements.
INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is
specified.
MEANATTRS=style-element | style-element (marker-options) | (marker-options)
specifies the attributes of the marker representing the mean within the
box.
MEDIANATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line representing the median within the
box.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.
OUTLIERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the markers representing the outliers.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the box outline.
SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid
overlap.
WHISKERATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the whiskers and caps.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
406 Chapter 6 / Plot Statements

specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

OUTLIERTIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over an
outlier.
ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
box or whisker in the box plot.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column
specifies the labels of the values that are identified as outlier or far
outlier by the STAT= column.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the outlier labels.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.
GROUP=column | discrete-attr-var | expression
creates a box plot for each unique group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
 BOXPLOTPARM Statement 407

specifies the ordering of the groups within a category.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page that is displayed when a box or an outlier marker
is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)
specifies the statistics to be displayed for each box.

Required Arguments
Specifying only Y= creates a single box plot. Specifying both X= and Y= creates a box plot
for each unique value of X.
Y=numeric-column | expression
specifies the column for the Y values. The Y values must be the statistical
values needed for the box plot. At a minimum, there must be nonmissing values
for the 25th and 75th percentiles.

X=column | expression
specifies the column for the X values. The X values must qualify or classify the
values in the Y column. This optional argument is used to create a plot box for
each classifier.

Restriction When you specify a numeric column for the X= argument and the
category axis is an interval axis, you should not change the column
value format. Doing so might result in incorrect output.

STAT=string-column
specifies the name of a column that contains the statistic represented by the
value in the Y column. Valid STAT= column values include the following (see the
requirements listed at the end of this description):

BOXWIDTH
specifies the width of the boxes as a ratio of the maximum possible width.

The range of values is 0 (narrowest) to 1 (widest). The default is 0.4. If the Y

value corresponding to BOXWIDTH is in range, then it overrides the setting
that is specified in the BOXWIDTH= option.

DATAMAX
specifies the maximum data value that includes not only the maximum
whisker values but the maximum outlier values as well.
408 Chapter 6 / Plot Statements

The DATAMAX value is greater than or equal to the MAX value and is always
represented in the axis range.

Note This statistic is valid in SAS 9.4M1 and later releases.

DATAMIN
specifies the minimum data value that includes not only the minimum
whisker values but the minimum outlier values as well.

The DATAMIN value is less than or equal to the MIN value and is always
represented in the axis range.

Note This statistic is valid in SAS 9.4M1 and later releases.

FAROUTLIER
specifies the observations that are outside the lower and upper far fences.

The far fences are located at a distance 3 times the Interquartile Range (IQR
= Q3–Q1) above and below the box. The far outliers are labeled when the
DATALABEL= option is used. Specify that LABELFAR= TRUE to label only
the far outliers but not the outliers.

MAX
specifies the maximum data value less than or equal to the upper fence.

MEAN
specifies the data mean.

MEDIAN
specifies the data median.

MIN
specifies the minimum data value greater than or equal to the lower fence.

N
specifies the subgroup sample size.

The N value is not shown in the plot but is used to calculate notch locations
when the DISPLAY= option displays notches.

OUTLIER
specifies the observations that are outside the lower and upper fences.

The fences are located at a distance 1.5 times the Interquartile Range (IQR =
Q3–Q1) above and below the box. The outliers are labeled when the
DATALABEL= option is used.

Q1
specifies the 1st quartile (25th percentile).

The data must contain a nonmissing value for this quartile.

Q3
specifies the 3rd quartile (75th percentile).

The data must contain a nonmissing value for this quartile.

STD
specifies the data standard deviation.
 BOXPLOTPARM Statement 409

SUMWGT
specifies the sum of the weights for the box.

Note This statistic is valid in SAS 9.4M1 and later releases.

Requirements Nonmissing Y values for STAT observations of Q1 and Q3 are

required in order for a box to be drawn. Other STAT values can be
omitted or have missing Y values.

The STAT values, if present, must conform to the following rules

in order for a box to be displayed:
Q1 <= MEDIAN <= Q3
MIN <= MAX
STD >= 0
N > 0

Optional Arguments
BOXWIDTH=number
specifies the width of a box as a ratio of the maximum possible width.

Defaults For a discrete category axis, 0.4 for a nongrouped box plot or 0.6
for a grouped box plot.

For an interval category axis, 85% of the smallest interval between

any two boxes for the given plot.

Range 0–1, where 0 is the narrowest and 1 is the widest.

Interactions For a grouped box plot with a discrete category axis when
GROUPDISPLAY=CLUSTER, the box width is a percentage of the
cluster width. See CLUSTERWIDTH.

This option is overridden by the Y value when the STAT= column

value is BOXWIDTH and the corresponding Y value is within the
range of 0–1.

Prior to SAS 9.4M3, for an interval box plot, this option is ignored,
and the box width is controlled by the INTERVALBOXWIDTH=
option. Starting with SAS 9.4M3, for an interval box plot, this option
is ignored when the INTERVALBOXWIDTH= option is specified.
Otherwise, the BOXWIDTH= option is honored.

CAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the whisker caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1

Interaction This option is ignored when CAPSHAPE=NONE is in effect.

410 Chapter 6 / Plot Statements

Tips Use the CAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the WHISKERATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the caps:

capscale=2

Reduce the size of the caps by 50%:

capscale=0.5

CAPSHAPE=SERIF | LINE | BRACKET | NONE

specifies the shape at the ends of the whiskers.

SERIF
specifies a short line perpendicular to the whisker.

LINE
specifies a line perpendicular to the whisker that extends the width of the
box.

BRACKET
specifies a line perpendicular to the whiskers that extends the width of the
box and that has short extensions at each end.

The extensions are drawn in the direction of the box.

NONE
specifies that no shape appears at the ends of the whiskers.

The following figure shows each of the shapes.

Default SERIF

Interactions The cap color and the thickness are specified by the
WHISKERATTRS= option. The cap pattern is always solid.

The DISPLAY= option must include CAPS in order for cap lines to
be shown.

Starting with SAS 9.4M5, when CAPSCALE= is in effect, the display

width for the caps is the cap's default width multiplied by the value
specified in CAPSCALE=.

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.
 BOXPLOTPARM Statement 411

Default 0.7

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

CONNECT=MEAN | MEDIAN | Q1 | Q3 | MIN | MAX

specifies that a connect line joins a statistic from box to box.

Default The GraphBox:Connect style reference.

Requirement The DISPLAY= option must contain the CONNECT display-options

value in order for the connect line to be displayed.

Interaction This option applies only when the X= argument is used to generate
multiple boxes.

Note Starting with SAS 9.4M3, the connect lines are drawn in axis order.
In prior releases, they are drawn in data order.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the lines connecting multiple boxes.

Default The GraphConnectLine style element.

Interactions If there is only one box, then this option is ignored.

If the DISPLAY= option does not include CONNECT, or if the

GROUP= option is used, then this option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

412 Chapter 6 / Plot Statements

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement DISPLAY= must include CONNECT for this option to have any
effect.

Interaction This option is ignored when the GROUP= option is in effect.

DATALABEL=column
specifies the labels of the values that are identified as outlier or far outlier by
the STAT= column. Either a numeric or a character column can be used.

Default No data labels are displayed

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See the LABELFAR= option

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the outlier labels.

Default The GraphDataText style element.

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When
set to TRUE, the data label is split unconditionally at each occurrence of any of
the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
 BOXPLOTPARM Statement 413

as a separate split character unless the specified characters appear

consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE
414 Chapter 6 / Plot Statements

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.
 BOXPLOTPARM Statement 415

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled boxes. The following figure shows
boxes with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.
416 Chapter 6 / Plot Statements

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the box outlines, box fill, whiskers,
mean, median, caps, connect lines, outliers, and data labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
interior fill area of the boxes. You can combine this option with
FILLATTRS= to set one transparency for the box outlines and the
whiskers, mean, median, caps, and connect lines, but a different
transparency for the box fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all boxes from the discrete X ticks.

Default 0 (no offset, all boxes are centered on the discrete ticks)

Range -0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right when
ORIENT=VERTICAL, and up when ORIENT=HORIZONTAL. (If the
layout's axis options set REVERSE=TRUE, then the offset direction
is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

See “About the DISCRETEOFFSET= Option” on page 438

DISPLAY=(display-options) | STANDARD | ALL

specifies which box-plot features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

 BOXPLOTPARM Statement 417

CAPS
displays caps at the ends of the whiskers.

CONNECT
displays the line connecting multiple boxes.

FILL
displays the box fill color.

FILLPATTERN
displays the box fill pattern.

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the
boxes.

MEAN
displays the mean symbol within the box.

MEDIAN
displays the median line within the box.

NOTCHES
displays notched boxes.

See “Notches” on page 437.

OUTLIERS
displays markers for the outliers.

STANDARD
specifies CAPS FILL MEAN MEDIAN, and OUTLIERS.

ALL
specifies all features: CAPS, CONNECT, FILL, FILLPATTERN, MEAN,
MEDIAN, NOTCHES, and OUTLIERS.

Default The GraphBox:DisplayOpts style reference. If this style element

does not exist, then the default is STANDARD.

Restriction The display features requested can be displayed only if the input
data includes this information.

Interaction If EXTREME= TRUE, then the OUTLIERS feature is ignored

Notes Starting with SAS 9.4M3, the connect lines are drawn in axis order.
They are drawn in data order in prior releases.

Starting with SAS 9.4M3, when DISPLAY= includes MEAN, the

BOXPLOTPARM statement contributes its mean markers to a
discrete legend when TYPE=MARKER is in effect in the
DISCRETELEGEND statement.
418 Chapter 6 / Plot Statements

Tips To control the appearance of these features, use the

CONNECTATTRS= , FILLATTRS= , MEANATTRS= ,
MEDIANATTRS= , OUTLIERATTRS= , and WHISKERATTRS=
options. The WHISKERATTRS= option controls affects both CAPS
and WHISKERS. Starting with SAS 9.4M5, use the
FILLPATTERNATTRS= on page 420 option to control the appearance
of the box fill pattern.

Regardless of which display options are being displayed, this option

does not affect the axis range.

DISPLAYSTATS=NONE | STANDARD | ALL | (statistics-list)

specifies the statistics to be displayed for each box.

NONE
does not display any statistics.

STANDARD
displays N, MEAN, and STD.

ALL
displays all available statistics.

(statistics-list)
specifies a space-separated list of one or more statistics, enclosed in
parentheses.

The list can include one or more of the following:

DATAMAX
maximum data value that includes not only the maximum whisker values
but also the maximum outlier values.

This option is valid in SAS 9.4M1 and later releases. The DATAMAX value
is greater than or equal to the MAX value and is always represented in
the axis range.

DATAMIN
minimum data value that includes not only the minimum whisker values
but also the minimum outlier values.

This option is valid in SAS 9.4M1 and later releases. The DATAMIN value
is less than or equal to the MIN value and is always represented in the
axis range.

IQR
interquartile range (Q3–Q1).

MAX
maximum data value below the box upper fence.

MEAN
mean data value for the box.

MEDIAN
median data value for the box.
 BOXPLOTPARM Statement 419

MIN
minimum data value above the box lower fence.

N
number of observations for the box.

Q1
lower quartile (25th percentile) for the box.

Q3
upper quartile (75th percentile) for the box.

RANGE
range of the data (MAX–MIN).

STD
standard deviation of the data for the box.

SUMWGT
sum of the weights for the box.

This option is valid in SAS 9.4M1 and later releases.

Default NONE

Restrictions Prior to SAS 9.4M5, this option is ignored if ORIENT=HORIZONTAL.

Starting with SAS 9.4M5, statistics can be displayed for both
vertical and horizontal box plots.

This option is ignored if the GROUP= option is also specified.

Only those statistics that are included in the STAT= column can be
displayed. RANGE requires both MAX and MIN to be included. IQR
requires both Q1 and Q3 to be included.

Note The notches in the box plot can extend beyond DATAMIN and
DATAMAX in some cases.

EXTREME=TRUE | FALSE
specifies whether the whiskers can extend beyond the fences. Fences are
locations above and below the box. The upper and lower fences are located at a
distance 1.5 times the Interquartile Range (IQR) (IQR = Q3 - Q1). The upper and
lower far fences are located at a distance 3 times the IQR .

FALSE
specifies that whiskers be drawn from the upper edge of the box to the
largest value within the upper fence, and from the lower edge of the box to
the smallest value within the lower fence.

This representation is sometime called a schematic box plot or a Tukey box

plot.

TRUE
specifies that whiskers be drawn to the largest and smallest data values,
whether these values are inside or outside the fences.

The outliers and far outliers are not displayed and are not labeled. This
representation is sometime called a skeletal box and whisker plot.
420 Chapter 6 / Plot Statements

Default FALSE

Interaction This option overrides the DATALABEL= , DATALABELATTRS= ,

LABELFAR= , OUTLIERATTRS= , and SPREAD= options.

See “Statement Summary” on page 433

“boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the boxes.

Defaults For non-grouped data, the Color attribute of GraphDataDefault style

element.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for the

box outlines, box fill, whiskers, mean, median, caps, connect lines,
and outliers. You can combine this option with
DATATRANSPARENCY= to set one transparency for the box
outlines and the whiskers, mean, median, caps, and connect lines,
but a different transparency for the box fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
 BOXPLOTPARM Statement 421

information, see “Attribute Rotation Patterns” in SAS Graph

Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

GROUP=column | discrete-attr-var | expression

creates a box plot for each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

The box plot display depends on the setting for the GROUPDISPLAY= option.
This option can be used to group the box plots in the display.

A distinct box is created for each group value by varying the visual attributes of
the box display features. The display features are controlled by the current ODS
style or by the DISPLAY= option. The default group appearance for each display
feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The Color attribute controls the fill color, and the ContrastColor
attribute controls the outline color. The box outline line style is
solid.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern
and box outline. The box outline line style is solid. If the active
ODS style does not specify FillPattern, fill patterns are not
displayed by default.

Median, caps, ContrastColor attribute of a GraphData1–GraphDataN style

and notches element or the GraphMissing style element with a solid line
style.
422 Chapter 6 / Plot Statements

Display
Feature 1 Style Attributes That Control Default Group Appearance

Mean and MarkerSymbol and ContrastColor attributes of a GraphData1–

outliers GraphDataN style element or the GraphMissing style element.

1 The default display features are determined by the DisplayOpts attribute of the GraphBox style
element.
2 Fill patterns are valid for box plots starting with SAS 9.4M5.

Defaults If box fills are enabled by the ODS style or by the DISPLAY= option,
then each distinct group value is represented in the plot by a
different fill color. Starting with SAS 9.4M5, each distinct group
value can be represented by a different fill pattern, with or without
a different fill color. The fill color and fill pattern are defined by the
Color and FillPattern attributes of the GraphData1–GraphDataN
and GraphMissing style elements.

If only box outlines are enabled by the ODS style or by the

DISPLAY= option, then each distinct group value is represented in
the plot by a different box outline color and line style. The outline
colors and line styles are defined by the ContrastColor and
LineStyle attributes of the GraphData1–GraphDataN style
elements.

Each distinct group value is represented in the plot by a different

box outline color. The outline colors are defined by the
ContrastColor attribute of the GraphData1–GraphDataN and
GraphMissing style elements.

If the MEDIAN, CAPS, or NOTCHES are enabled by the DISPLAY=

option, then each of these features uses the same color as the box
outline. Line styles do not change by group value.

If the MEAN or OUTLIERS are enabled by the DISPLAY= option,

then each distinct group value is represented by a different marker.
The markers are defined by the MarkerSymbol and ContrastColor
attributes of the GraphData1–GraphDataN and GraphMissing style
elements. A marker is used for both MEAN and OUTLIERS, if
displayed.

Interactions Connect lines are not drawn for grouped data.

This option causes the DISPLAY=(CONNECT) and the CONNECT=

options to be ignored.

Tips By default, the group values are mapped in the order of the data.
The GROUPORDER= option can be used to control the sorting
order of the group values. The INDEX= option can be used to alter
the default sequence of colors and markers.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.
 BOXPLOTPARM Statement 423

The representations that are used to identify the groups can be

overridden individually. For example, each distinct group value is
represented by a different line pattern for the box outlines, but the
PATTERN= setting on the OUTLIERATTRS= option could be used
to assign the same line pattern to all box outlines and connect
lines.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See the GROUPDISPLAY= option to see the output for the grouped
boxes

“DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how to display the boxes that represent group values for the
coordinate pairs. The following example shows a box plot with
GROUPDISPLAY=CLUSTER:

OVERLAY
draws boxes for a given group value at the exact coordinate.

Depending on the data, boxes at a given coordinate might overlap.

CLUSTER
draws boxes for a given group value adjacent to each other.

This option is available only when the category (X) column is discrete.

Default OVERLAY

Interactions This option is ignored unless GROUP= is specified.

424 Chapter 6 / Plot Statements

The groups in the legend are shown in the order that is specified in
GROUPORDER by default.

Attributes such as color, symbol, and pattern are assigned to each

group in DATA order by default.

Tip Use the CLUSTERWIDTH= option to control the width of the

clusters when CLUSTER is in effect.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
 BOXPLOTPARM Statement 425

unpredictable results might occur. To generate consistent graphs in

that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping box attributes (fill and outline) to one of the
GraphData1–GranphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

INTERVALBOXWIDTH=AUTO | dimension
specifies the box width when an interval category (X) column is specified.
426 Chapter 6 / Plot Statements

AUTO
the system determines the box width.

uses one of the following box-width values listed in the order of precedence:
n the BOXWIDTH statistic, if it is included in the Stat column and if the Y
value is within the range of 0–1
n starting with SAS 9.4M3, the BOXWIDTH= option value, if it is specified

n otherwise, 85% of the smallest interval between any two boxes for the
given plot

dimension
sets the box width to the specified value.

Default AUTO

Requirement The category axis (X by default) must be an interval axis for the
INTERVALBOXWIDTH= option to have any effect.

Interactions Starting with SAS 9.4M3, this option is overridden by the Y value
when the STAT= column value is BOXWIDTH and the
corresponding Y value is within the range of 0–1.

Prior to SAS 9.4M3, this option controls the box width for an
interval box plot. Starting with SAS 9.4M3, this option overrides
the BOXWIDTH= option for an interval box plot.

Tip The box width that you specify with this option is honored even if
the boxes overlap. If the boxes overlap, reduce the interval box
width or use the BOXWIDTH= option instead.

See “dimension” on page xiii

LABELFAR=TRUE | FALSE
specifies whether all outliers or only far outliers are labeled.

FALSE
applies the labels specified by the DATALABEL= option to both the outliers
and the far outliers.

TRUE
applies the labels specified by the DATALABEL= option to the far outliers.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “DATALABEL=column ” on page 412

“Statement Summary” on page 391 for information about outliers.

“boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
 BOXPLOTPARM Statement 427

Default The Y= column label. If a label is not defined, then the Y= column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MEANATTRS=style-element | style-element (marker-options) | (marker-options)

specifies the attributes of the marker representing the mean within the box.

Defaults For non-grouped data, GraphBoxMean style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
mean.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MEDIANATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line representing the median within the box.

Defaults For non-grouped data, the GraphBoxMedian style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxMedian style element, and the ContrastColor attribute of
the GraphData1–GraphDataN style elements.

Interaction This option is ignored if the DISPLAY= option does not display the
median.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.
428 Chapter 6 / Plot Statements

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and of the boxes.

Default VERTICAL

OUTLIERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the markers representing the outliers.

Defaults For non-grouped data, GraphOutlier style element.

For grouped data, the MarkerSymbol, Markersize, and ContrastColor

attributes of the GraphData1–GraphDataN style elements.

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

OUTLIERTIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over an
outlier. If this option is used, then it replaces all of the information that is
displayed by default.

(role-list)
an ordered, space-separated list of unique BOXPLOTPARM roles and user-
defined roles. BOXPLOTPARM roles for OUTLIERTIP include X, Y, STAT, and
DATALABEL.

Note In the data tip, the STAT role displays the text “outlier” or “far
outlier” as applicable.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns that are
assigned to the X and Y roles, and also the data column Obs, which
is not assigned to any pre-defined BOXPLOTPARM role. The Obs
column must first be assigned a role:
ROLENAME=(TIP1=OBS)
OUTLIERTIP=(X Y TIP1)

Default The columns assigned to these roles are automatically included in

the data tip information: X and Y .

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that specifies the IMAGEMAP option, and write the
graphs to the ODS HTML destination.

Interaction The labels and formats for the OUTLIERTIP variables can be
controlled with the TIPLABEL= and TIPFORMAT= options.

See the ROLENAME= option for specifying user-defined roles.

 BOXPLOTPARM Statement 429

the TIP= option for specifying the information to display when the
cursor is positioned over a box or whisker in the box plot.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the box outline.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data and filled boxes, the LineStyle and LineThickness
attributes of the GraphOutlines style element, and the ContrastColor
attribute of the GraphData1–GraphDataN style elements.

For grouped data and unfilled boxes, the LineThickness attribute of

the GraphOutlines style element, and the ContrastColor and LineStyle
attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the OUTLIERTIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)
430 Chapter 6 / Plot Statements

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles DATAMAX, DATAMIN, MAX, MIN, MEAN,
MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT are valid in SAS
9.4M1 and in later releases.

SPREAD=TRUE | FALSE
specifies whether outliers with the same value are spread out to avoid overlap.
For vertical box plots this means offsetting the outliers horizontally. If this
option is false, then outliers with the same value are plotted in the same
position, which means only one is visible.

Default FALSE

Interaction This option is ignored if EXTREME= TRUE or the DISPLAY= option

does not display the outliers.

See “boolean ” on page xii for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a box or
whisker in the box plot. If this option is used, then it replaces all of the
information that is displayed by default. Roles for columns that do not
contribute to the box plot can be specified along with roles that do contribute.

(role-list)
an ordered, space-separated list of unique BOXPLOTPARM roles.

BOXPLOTPARM roles for TIP include DATAMAX, DATAMIN, MAX, MIN,

MEAN, MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Note The roles DATAMAX, DATAMIN, and SUMWGT are valid in SAS
9.4M1 and later releases.

Tip Statistics such as N, MIN, and MAX are special roles, They are not
column-based like the X role.

Example The following example displays data tips only for the columns that
are assigned to the roles X (CATEGORY) and the statistic MEAN:
TIP=(X MEAN)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: DATAMAX, DATAMIN, MAX, MIN, MEAN,
MEDIAN, N, Q1, Q2, STD, SUMWGT, and X.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.
 BOXPLOTPARM Statement 431

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

See the OUTLIERTIP= option for specifying the information to display

when the cursor is positioned over an outlier.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the OUTLIERTIP= or TIP= options are
used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page that is displayed when a box or an outlier marker is
selected.
432 Chapter 6 / Plot Statements

Note: This option is valid starting in SAS 9.4M1.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each box or outlier marker that is to have an active link.

Example http://www.sas.com

The character column can specify different target URLs for each box and outlier
marker.

Requirements The target URLs for the boxes must be specified in the Q1 statistic
observations, and the target URLs for the outlier markers must be
specified in the OUTLIER statistic observations. URLs that are
specified in observations other than Q1 and OUTLIER are ignored.

To generate selectable boxes and outlier markers, you must

include an ODS GRAPHICS ON statement that specifies the
IMAGEMAP option, and you must write the output to the ODS
HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for a box or outlier, meaning that no
action is taken when that box or outlier marker is selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET=
on page 1568.

Example The following vehicle mileage data sample shows box and outlier
URLs specified in column URL for Sedan.
STAT X VALUE DATALABEL URL
...
N Sedan 262.0
MEAN Sedan 08
MEDIAN Sedan 20.00
Q1 Sedan 18.00 ./mileageSedan.html
Q3 Sedan 24.00
STD Sedan 4.23
OUTLIER Sedan 36.00 Honda ./mileageHonda.html
OUTLIER Sedan 35.00 Toyota ./mileageToyota.html
OUTLIER Sedan 35.00 Toyota ./mileageToyota.html
OUTLIER Sedan 38.00 Volkswagen ./mileageVolkswagen.html
MIN Sedan 12.00 Volvo
MAX Sedan 33.00 Volvo
...
 BOXPLOTPARM Statement 433

WHISKERATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the whiskers and caps.

Defaults For non-grouped data, the GraphBoxWhisker style element.

For grouped data, the LineStyle and LineThickness attritubes of the

GraphBoxWhisker style element, and the ContrastColor attribute of
the GraphData1–GraphDataN style elements.

Restriction The caps are always drawn with a solid line.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Summary
The BOXPLOTPARM statement requires precomputed input data. One reason to
choose this statement over the BOXPLOT statement is that you can control the
computational technique used to compute various statistics for the box plot, such
as the mean, quartiles, location of fences, outlier definition, and so on. See
Appendix 5, “Generalized Macro for BOXPLOTPARM Data,” on page 1627 for
examples of such computations using PROC SUMMARY and multiple DATA steps.

The BOXPLOTPARM statement displays a single box if given just Y and a STAT
argument. It displays multiple boxes if given both Y and X and a STAT argument
and X has more than one unique value.
434 Chapter 6 / Plot Statements

By default for numeric or character columns, the category (X) axis is

TYPE=DISCRETE. You can override the default and specify TYPE=LINEAR in the
parent layout, provided that the X column is numeric. The axis for the analysis (Y)
column is always LINEAR. When the X axis is LINEAR, you can use the
INTERVALBOXWIDTH= option to specify the box width.

When ORIENT= VERTICAL, the X (or X2) axis is used for the X column and the Y
(or Y2) axis is used for the Y column. When ORIENT=HORIZONTAL, the X (or X2)
axis is used for the Y column and the Y (or Y2) axis is used for the X column.

Two basic box plot representations can be drawn with the BOXPLOTPARM
statement: a schematic (Tukey) box plot and a skeletal box plot. See the EXTREME=
option for details.

Visual Description of Box Plot Percentile Boundaries

Box plots display the distribution of data by using a rectangular box and whiskers.
Whiskers are lines that indicate a data range outside of the box. The following
figure shows the parts of a box plot.

As shown in the figure, the bottom and top edges of the box are located at the 1st
quartile (25th percentile) and 3rd quartile (75th percentile) of the sample. Within
the box, you can display the median (50th percentile) as a line and the mean as a
marker (see the DISPLAY= option).

You can also display markers and data labels for outliers. Outliers are observations
that are more extreme than the upper and lower fences (± 1.5 IQR). Outliers that
are beyond upper and lower far fences (±3 IQR) are called FAR OUTLIERS and can
also be identified and labeled. From a graphical perspective, the locations of fences
 BOXPLOTPARM Statement 435

along the axis are known, but there is no line or marker that displays a fence. (See
DISPLAY= , LABELFAR= , and DATALABEL= options).

Finally, you can control the range represented by the whiskers. By default, the
whiskers are drawn from the upper edge of the box to the MAX value, and from the
lower edge of the box to the MIN value. (See the EXTREME= option.)

Input Data Requirements for the BOXPLOTPARM Statement

At a minimum, valid data for the BOXPLOTPARM statement must provide a
numeric column (Y=) that contains calculated statistics for an analysis, and a string
column (STAT=) that identifies each statistic. The Y column must contain
nonmissing values for the Q1 (25th percentile) and Q3 (75th percentile) statistics. If
Y values are missing or not supplied for other statistic values, then those statistics
are not displayed in the plot, regardless of syntax requests to display them.

For example, a petroleum company uses a turbine to heat water into steam that is
pumped into the ground to make oil more viscous and easier to extract. This
process occurs 20 times daily, and the amount of power (in kilowatts) used to heat
the water to the desired temperature is recorded. The following data show the
statistics that are calculated for one day of this process:

PowerOutputs Statistic

3180.00 MIN

3340.00 Q1

3487.40 MEAN

3490.00 MEDIAN

3610.00 Q3

4050.00 MAX

20.00 N

To plot the data from the preceding table, the following BOXPLOTPARM statement
uses the Y= and STAT= arguments to generate a single box plot for the recorded
statistics:

BOXPLOTPARM Y=PowerOutputs STAT=Statistic;

436 Chapter 6 / Plot Statements

If the data contain statistics for multiple days of the process, then a third column in
the data must be present to identify the days that the statistics were recorded. For
example, the following data show the statistics that are calculated for two days of
this process:

Day PowerOutputs Statistic

04JUL 3180.00 MIN

04JUL 3340.00 Q1

04JUL 3487.40 MEAN

04JUL 3490.00 MEDIAN

04JUL 3610.00 Q3

04JUL 4050.00 MAX

04JUL 20.00 N

05JUL 3179.00 MIN

05JUL 3333.50 Q1

05JUL 3471.65 MEAN

05JUL 3419.50 MEDIAN

05JUL 3605.00 Q3

05JUL 3849.00 MAX

05JUL 20.00 N
 BOXPLOTPARM Statement 437

To plot the data from the preceding table, the BOXPLOTPARM statement needs
the Y=, STAT=, and X= arguments to generate a separate box plot for each day that
the statistics were recorded:

BOXPLOTPARM Y=PowerOutputs STAT=Statistic X=Day;

See Appendix 5, “Generalized Macro for BOXPLOTPARM Data,” on page 1627 for a
more complete example of providing input data for BOXPLOTPARM.

Changing Box Plot Display

SAS defines graphical style elements that control the display of box plots. Using
these style elements as a starting point, you can change the style attribute values
to achieve a very different appearance for your box plots. For more information, see
“Changing Box Plot Display” on page 393.

Notches
In a notched box plot, the endpoints of the notches are at the following computed
locations:
1.58 IQR
median ±
n

In the equation, IQR (IQR=Q3-Q1) is the interquartile range and n is the sample size.
The medians (central lines) of the two boxes are significantly different at
approximately the 0.05 level if the corresponding notches do not overlap.
438 Chapter 6 / Plot Statements

Maximum observation

th
75 percentile (upper
quartile)

1/2
1.58(IQR) / n
+ Mean
Median

1/2
1.58(IQR) / n
th
25 percentile (lower
quartile)

Minimum observation

Note: In some cases, the notches in the box plot can extend beyond DATAMIN and
DATAMAX.

About the DISCRETEOFFSET= Option

The DISCRETEOFFSET= option is useful for graphing multiple response variables
side by side on a common axis. By default within an overlay-type layout, if multiple
BOXPLOTPARM statements are used with different analysis variables, then the
boxes for matching X values are centered on the ticks. Depending on the data, the
boxes might be superimposed. The following code fragment shows the default box
positioning:
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");

boxplotparm x=type y=mpg_city stat=y_stat / name="City" ;

boxplotparm x=type y=mpg_highway stat=y_stat / name="Highway" ;

discretelegend "City" "Highway";

endlayout;
 BOXPLOTPARM Statement 439

To place the different response values side by side, you can assign a different offset
to each BOXPLOTPARM statement. The BOXWIDTH= option can be used in
conjunction with the DISCRETEOFFSET= option to create narrower boxes when
desired.
layout overlay / cycleattrs=true
yaxisopts=(label="Miles Per Gallon");

boxplotparm x=type y=mpg_city stat=y_stat / name="City"

discreteoffset=0.2 ;
boxplotparm x=type y=mpg_highway stat=y_stat / name="Highway"
discreteoffset=-0.2 ;

discretelegend "City" "Highway";

endlayout;
 440 Chapter 6 / Plot Statements

Example: BOXPLOTPARM Statement

The following graph was generated by the “Example Program” on page 440:

Use the %BOXCOMPUTE macro to generate the data for this example. See
Appendix 5, “Generalized Macro for BOXPLOTPARM Data,” on page 1627 to see the
code for creating all of the data.
%boxcompute(indsn=sashelp.cars,x=type,y=mpg_city,datalabel=make);

Example Program
Here is the code for this example.
/* Define the template for the plot. */
proc template;
define statgraph boxplotparm1;
begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true ;
endlayout;
endgraph;
end;
run;
 BUBBLEPLOT Statement 441

/* Generate the plot. */

proc sgrender data=boxdata template=boxplotparm1;
run;

BUBBLEPLOT Statement
Creates a bubble plot of the input data. The locations of the bubble centers correspond to the values
of X and Y columns in the data, and the bubble radii correspond to the values of a SIZE column.

Tip: Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
BUBBLEPLOT X=column | expression
Y=column | expression
SIZE=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
BUBBLERADIUSMAX=dimension
specifies the drawing size of the largest bubble.
BUBBLERADIUSMIN=dimension
specifies the drawing size of the smallest bubble.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variable to use to
determine the bubble colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bubbles.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bubble fills and bubble
outlines
DISPLAY=(display-options) | STANDARD | ALL
specifies which bubble features to display.
DRAWORDER=SIZE | DATA
442 Chapter 6 / Plot Statements

specifies whether the bubbles are drawn according to bubble size or

according to data order.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled bubble areas.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
INDEX=positive-integer-column | expression
specifies indices for mapping bubble attributes (fill and outline) to one of
the GraphData1–GraphDataN style elements.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bubble outlines.
RELATIVESCALE=TRUE | FALSE
specifies whether the SIZE= column values are interpreted as relative
values.
RELATIVESCALETYPE=LINEAR | PROPORTIONAL
specifies the type of scaling that is to be applied to the SIZE= column
values.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
SIZETHRESHOLDMAX=numeric-value
specifies a SIZE= column value threshold at which bubble size is clamped
to the BUBBLERADIUSMAX= option value.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over
the bubbles.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
 BUBBLEPLOT Statement 443

specifies a column for bubble labels.

DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the bubble labels.
DATALABELPOSITION=TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER | RIGHT
| BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the bubble labels relative to the bubble.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate bubble color for each unique grouping that is
specified.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a bubble is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values of the bubble centers.

Y=column | expression
specifies the column for the Y values of the bubble centers.

SIZE=numeric-column | expression
specifies the bubble SIZE values.
444 Chapter 6 / Plot Statements

Optional Arguments
BUBBLERADIUSMAX=dimension
specifies the drawing size of the largest bubble.

Default Three times as large as the size set by GraphDataDefault:markerSize

Restriction The dimension value must be greater than the

BUBBLERADIUSMIN=dimension value.

Interaction This option is ignored when RELATIVESCALE= FALSE.

Tip A maximum size that is specified as a percent is interpreted as a

percent of the graph's height. The height can be adjusted with the
DESIGNHEIGHT= option in the BEGINGRAPH statement or the
HEIGHT= option in the ODS GRAPHICS statement.

See “dimension” on page xiii

BUBBLERADIUSMIN=dimension
specifies the drawing size of the smallest bubble.

Default GraphDataDefault:markerSize

Restriction The BUBBLERADIUSMIN= value must be less than the

BUBBLERADIUSMAX= value.

Interaction This option is ignored when RELATIVESCALE= FALSE.

Tip A maximum size that is specified as a percent is interpreted as a

percent of the graph's height. The height can be adjusted with the
DESIGNHEIGHT= option in the BEGINGRAPH statement or the
HEIGHT= option in the ODS GRAPHICS statement.

See “dimension” on page xiii

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.
 BUBBLEPLOT Statement 445

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element for filled bubbles and

ThreeColorAltRamp for unfilled bubbles

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variable to use to
determine the bubble colors.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Interactions Prior to SAS 9.4M5, when the GROUP= option is specified with the
COLORRESPONSE= option, the GROUP= option is ignored. Starting
with SAS 9.4M5, the COLORRESPONSE= option controls the color
of the group fill patterns when the fill patterns are displayed.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides

suboption COLOR= in the OUTLINEATTRS= option and varies the
outline color according to the color gradient or the attribute map.

Notes When both the fill and outline are displayed, the bubble fill color
varies according to the color gradient or attribute map but the
446 Chapter 6 / Plot Statements

bubble outline color remains fixed on the color specified in option

OUTLINEATTRS=.

The mapped color used for the bubbles is also used for the data
labels.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATALABEL=column | expression
specifies a column for bubble labels. The label positions are adjusted to prevent
them from overlapping.

Default No labels are displayed.

Note The BUBBLEPLOT statement does not use a data-label collision

avoidance algorithm. Data labels in bubble plots might overlap.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the bubble labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the text color is derived from the GraphData1–
GraphDataN style elements. The data label color changes to match
the group color derived from the ContrastColor attribute of the style
element that is in effect.

Interaction The default attributes are overridden if the COLORRESPONSE=

option is used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER | RIGHT |

BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the bubble labels relative to the bubble.

Default TOPRIGHT

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When
set to TRUE, the data label is split unconditionally at each occurrence of any of
the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

See “boolean ” on page xii for other Boolean values that you can use.
 BUBBLEPLOT Statement 447

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interaction The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A
448 Chapter 6 / Plot Statements

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.
 BUBBLEPLOT Statement 449

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bubbles. The following figure
shows bubbles with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, DISPLAY= FILL must be in
effect. Otherwise, this option is ignored.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

450 Chapter 6 / Plot Statements

This option is ignored if the RELATIVESCALE= option is set to

FALSE.

When a data skin is applied, all bubble outlines are set by the skin,
and the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bubble fills and bubble outlines

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

Note This option does not affect the data labels.

Tip The FILLATTRS= option can be used to set transparency for just the
filled bubble areas. You can combine this option with FILLATTRS= to
set one transparency for the bubble outlines but a different
transparency for the bubble fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which bubble features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE
displays the bubble outline

FILL
displays the bubble fill color

FILLPATTERN
displays the bubble fill pattern

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the
bubbles.

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: FILL, FILLPATTERNATTRS, and OUTLINE
 BUBBLEPLOT Statement 451

Default STANDARD

Tip Use the DATASKIN= , OUTLINEATTRS= , and FILLATTRS= options to

control the appearance of the bubble outline and fill color. Starting with
SAS 9.4M5, use the FILLPATTERNATTRS= on page 452 option to
control the appearance of the bubble fill pattern.

DRAWORDER=SIZE | DATA
specifies whether the bubbles are drawn according to bubble size or according
to data order.

Note: This feature applies to SAS 9.4M3 and to later releases.

SIZE
draws the bubbles according to bubble size, from the largest to the smallest

DATA
draws the bubbles according to data order

The following figure shows the effect of SIZE and DATA on four bubbles. The
bubble labels indicate the data order, and the bubble sizes increase linearly
starting with 1.

Default SIZE

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled bubble areas.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the bubble fill colors vary according to the gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

bubble fills and the bubble outlines. You can combine this option
with DATATRANSPARENCY= to set one transparency for the
outlines but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
452 Chapter 6 / Plot Statements

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interactions For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Starting with SAS 9.4M5, when COLORRESPONSE= is in effect and

the DISPLAY= option enables FILLPATTERN display, the
FILLPATTERNATTRS= suboption COLOR= is ignored, and the fill-
pattern colors vary according to the gradient.

Tip In order to change the fill pattern for each group value in ODS
styles such as HTMLBlue that use color-priority attribute rotation,
specify ATTRPRIORITY=NONE in an ODS GRAPHICS statement or
in the BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

GROUP=column | discrete-attr-var | expression

creates a separate bubble color for each unique grouping that is specified.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct bubble is created for each group value by varying the visual attributes
of the bubble display features. The display features are controlled by the
current ODS style or by the DISPLAY= option. The default group appearance for
each display feature is shown in the following table.
 BUBBLEPLOT Statement 453

Display
Feature1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.

Outline n if the outline is enabled with fill

color, fill pattern, or both,
ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is the only display
feature enabled, ContrastColor and
LineStyle attributes of a
GraphData1–GraphDataN style
element or the GraphMissing style
element

1 Fill color and the outline are displayed by default.

2 Fill patterns are valid for bubble plots starting with SAS 9.4M5.

Interactions If a discrete attribute map variable is specified, then the color

mapping for the bubbles is defined by the associated
DISCRETEATTRMAP statement. See “DISCRETEATTRMAP
Statement” on page 1505.

The mapped color that is used for outlines is also used as the color
of the data labels.

This option is ignored if the COLORRESPONSE= option is also

used.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Notes All bubbles have only one fill and one outline color as specified by
the FILLATTRS= and OUTLINEATTRS options.

If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tip By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
454 Chapter 6 / Plot Statements

GROUPMAX= option in the ODS GRAPHICS statement to change

the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping bubble attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
 BUBBLEPLOT Statement 455

Default The SIZE= column label. If a label is not defined, then the SIZE=
column name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bubble outlines.

Defaults For non-grouped data, the ContrastColor and LineStyle attributes

of the GraphOutlines style element.

For grouped data and filled bubbles, the ContrastColor attribute of

the GraphData1–GraphDataN style elements and a solid line style.
If the COLORRESPONSE= option is specified, then the bubble
outline attributes are derived from the GraphDataDefault style
element.

For grouped data and unfilled bubbles, the ContrastColor and

LineStyle attributes of the GraphData1–GraphDataN style
elements.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bubble outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

456 Chapter 6 / Plot Statements

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

RELATIVESCALE=TRUE | FALSE
specifies whether the SIZE= column values are interpreted as relative values.
Relative means that the size values do not translate directly into bubble radii.
Rather, the bubble sizes are scaled to represent the value range of the SIZE=
column.

For example, when RELATIVESCALE=TRUE, if only two bubbles are drawn with
sizes of 2 and 4, then they appear the same as only two bubbles with sizes of
4000 and 8000. By contrast, when RELATIVESCALE=FALSE, the size values
are interpreted in the same units as the axes.

If you set this option to FALSE, then it is recommended that you also place the
BUBBLEPLOT statement in a LAYOUT OVERLAYEQUATED container. This
ensures that the X axis and Y-axis units are the same. For more information, see
“Details” on page 460.

Default TRUE

Interactions When this option is set to TRUE, the BUBBLERADIUSMAX= and

BUBBLERADIUSMIN= options can be used to fix the drawing size of
the smallest bubble and largest bubble. If
RELATIVESCALE=FALSE, then the BUBBLERADIUSMAX= and
BUBBLERADIUSMIN= options are ignored.

If this option is set to FALSE, then the DATASKIN= option is

ignored.

If one or both axes are discrete, then RELATIVESCALE=FALSE is

ignored.

Tip If you specify RELATIVESCALE=FALSE, then it is recommended

that you also place the BUBBLEPLOT statement in a LAYOUT
OVERLAYEQUATED container to ensure that the X axis and Y-axis
units are the same. If you place the BUBBLEPLOT statement in a
 BUBBLEPLOT Statement 457

LAYOUT OVERLAY container instead, then the bubbles might be

drawn as ellipses because the X axis and Y-axis units are different.

See “boolean ” on page xii for other Boolean values that you can use.

RELATIVESCALETYPE=LINEAR | PROPORTIONAL
specifies the type of scaling that is to be applied to the SIZE= column values.

LINEAR
increases the size of the bubbles in linear proportion to the range of the
SIZE= column values.

For example, if only two bubbles are drawn with sizes of 2 and 4, then they
appear the same as only two bubbles with sizes of 4000 and 8000.

PROPORTIONAL
increases the size of each bubble in direct proportion to its corresponding
SIZE= column value.

For example, if only two bubbles are drawn with sizes of 50 and 100, then
the bubble for SIZE=50 is drawn to half the size of the bubble for SIZE=100.

Default LINEAR

Interactions This option is ignored when RELATIVESCALE=FALSE.

When theSIZETHRESHOLDMAX= option is specified, for any SIZE=

column value that is greater than the SIZETHRESHOLDMAX=
value, the proportional scale is adjusted so that the size of the
bubble for that value is clamped to theBUBBLERADIUSMAX=
value.

If all the values for the SIZE= column are negative, then
RELATIVESCALETYPE=PROPORTIONAL is ignored, and the
default value is used.

When RELATIVESCALETYPE=PROPORTIONAL is specified, the

BUBBLERADIUSMIN= option specifies the minimum bubble size. In
that case, when a SIZE= column value results in a bubble of a size
that is less than the BUBBLERADIUSMIN= value, the bubble size
for that value is changed to the BUBBLERADIUSMIN= value.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.
458 Chapter 6 / Plot Statements

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y , SIZE, GROUP, DATALABEL, and
COLORRESPONSE.

SIZETHRESHOLDMAX=numeric-value
specifies a SIZE= column value threshold at which bubble size is clamped to the
BUBBLERADIUSMAX= option value. The size of the bubbles for all SIZE=
column values that equal or exceed the specified threshold value is set to the
BUBBLERADIUSMAX= value.

Default The maximum SIZE= column value is mapped to the

BUBBLERADIUSMAX= option value.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
bubbles. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the bubble
plot can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique BUBBLEPLOT and user-defined
roles.

BUBBLEPLOT roles include X, Y, SIZE, GROUP, DATALABEL, and

COLORRESPONSE.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned
to the roles X, Y, and SIZE, as well as the column Pop_2009. The
POP_2009 column is not assigned to any pre-defined
BUBBLEPLOT role, so it must first be assigned a role:

ROLENAME=(TIP1=POP_2009)
TIP=(TIP1 X Y SIZE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X , Y , SIZE , GROUP , and
COLORRESPONSE .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.
 BUBBLEPLOT Statement 459

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a bubble is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each bubble that is to have an active link.

Example http://www.sas.com
460 Chapter 6 / Plot Statements

Requirement To generate a plot with selectable bubbles, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the bubbles for those X and Y pairs
are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The BUBBLEPLOT statement displays one bubble for each row in the data,
provided that row contains nonmissing values for X, Y, and SIZE. By default, the
bubbles are displayed as filled, outlined circles. Regardless of the data order, the
bubbles are always drawn from the largest size to the smallest size.
 BUBBLEPLOT Statement 461

By default, the minimum and maximum values of the SIZE= column establish a
range over which the bubble radii increase in linear proportion. The actual drawing
size of the smallest and largest bubble is set automatically. You can adjust the
smallest and largest bubble sizes with the BUBBLERADIUSMIN= and
BUBBLERADIUSMAX= options. In these cases where the bubble sizes are
proportional to each other, the default setting RELATIVESCALE=TRUE is
appropriate.

If the SIZE= values are in the same units as the X and Y values, and both X and Y
are numeric, then you can generate a plot where the bubble-radius units match the
axis-scale units. To do so, specify the BUBBLEPLOT statement within a LAYOUT
OVERLAYEQUATED block, and in the BUBBLEPLOT statement, set
RELATIVESCALE=FALSE.

By default, for character columns, the X and Y axes are always discrete. For
numeric columns, the X and Y axes are linear. You can change axis type for numeric
axes with the layout options XAXISOPTS= and YAXISOPTS=.

Note: Within a LAYOUT OVERLAY, the unit-interval of the X and Y axes are not
necessarily the same and the bubbles might be distorted into ellipses when
RELATIVESCALE=FALSE. The OVERLAYEQUATED container ensures that the
bubbles are displayed as circles, assuming that both the X= and Y= arguments
specify numeric columns.

data influence;
input x y radius category;
datalines;
2 4 1 1
5 5 2 1
6 3 2 2
12 7 3 2
;
proc template;
define statgraph equatedbubbles;
begingraph;
entrytitle 'Radius of Influence';
entrytitle 'Bubbles Show Distance Covered by Observation';
layout overlayequated /
xaxisopts = (griddisplay=on)
yaxisopts = (griddisplay=on);
bubbleplot x=x y=y size=radius /
group=category datatransparency=0.5
relativescale=false ;
endlayout;
endgraph;
end;

proc sgrender data=influence template=equatedbubbles;

run;
462 Chapter 6 / Plot Statements

Example: BUBBLEPLOT Statement

The following graph was generated by the “Example Program” on page 463 :
 CONTOURPLOTPARM Statement 463

Example Program
data bubbleintro;
input Engineer $ Salary number;
format Salary dollar7.0 number comma6.0;
datalines;
Electric 59000 89382
Civil 54000 73273
Software 56000 34833
Chemical 62000 25541
Mechanical 60000 19601
;
proc template;
define statgraph engineer;
begingraph;
entrytitle 'Median Salary for Entry Level Engineers';
entrytitle 'Bubbles Show Number of Engineers in Survey';
layout overlay;
bubbleplot x=engineer y=salary
size=number / datalabel=number;
endlayout;
endgraph;
end;

proc sgrender data=bubbleintro template=engineer;

run;

CONTOURPLOTPARM Statement
Creates a contour plot representing a response variable evaluated over a grid of X and Y values.

Restriction: Contour plots do not support data tips.

Tips: By default, the CONTOURPLOTPARM statement assumes that the X-Y grid is
complete and does not contain any missing or irregular values. If the X-Y grid is not
complete, specify GRIDDED=FALSE in the CONTOURPLOTPARM statement so
that the values needed to complete the grid are calculated. Otherwise, unexpected
results might occur.
You can use a legend to display the contour level values. For CONTOURTYPE=LINE
and CONTOURTYPE=LABELEDLINE, use a DISCRETELEGEND statement to add a
legend. For all other contour types, use a CONTINUOUSLEGEND statement to add
a legend.
For filled contour types, there might be small, visible gaps between the axes and
the contour boundaries. To eliminate the gaps, specify the following axis options in
the layout statement for the plots parent layout:
XAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))
YAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))
464 Chapter 6 / Plot Statements

Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
CONTOURPLOTPARM X=numeric-column | expression
Y=numeric-column | expression
Z=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used to determine the colors of filled
or gradient contours.
CONTOURTYPE=LINE | FILL | GRADIENT | LINEFILL | LINEGRADIENT |
LABELEDLINE | LABELEDLINEFILL | LABELEDLINEGRADIENT
specifies how the contour is displayed.
GRIDDED=TRUE | FALSE
specifies whether the X and Y values are equally spaced in a rectangular
grid.
LEVELS=(contour-value-list)
specifies a list of contour level values.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the contour lines.
NHINT=integer
specifies the suggested number of contour levels for the Z column.
NLEVELS=integer
specifies the actual number of contour levels for the Z column.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
 CONTOURPLOTPARM Statement 465

specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
LINELABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the contour line labels.
LINELABELBASELINE=HORIZONTAL | TANGENT
specifies the text alignment of the contour line labels.
LINELABELFORMAT=format
specifies the format to use for the contour line labels.
LINELABELPOSITION=MIDDLE | BEGIN | END
specifies the position for the contour line labels.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the X coordinates for the grid.

Y=numeric-column | expression
specifies the Y coordinates for the grid.

Z=numeric-column | expression
specifies the contour response values.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used to determine the colors of filled or
gradient contours.

style-element
specifies the name of a style element. The style element can contain these
style attributes:

STARTCOLOR specifies a color for the smallest data value of the

Z= column.
NEUTRALCOLOR specifies a color for the midpoint of the range of the
Z= column.
ENDCOLOR specifies a color for the highest data value of the Z=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp.
466 Chapter 6 / Plot Statements

You can use style attribute references such as GraphData2:Color, color

names, or RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color.
The list can contain a mix of style attribute references, color names, and
color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element.

Interactions The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

The NeutralColor attribute is not used for two-color ramps.

Note For a contour plot of type FILL, LINEFILL, or LABELEDLINEFILL,

the color that is chosen for the first segment is the color for the N+1
inflection point and not the starting color in the color model.

CONTOURTYPE=LINE | FILL | GRADIENT | LINEFILL | LINEGRADIENT |

LABELEDLINE | LABELEDLINEFILL | LABELEDLINEGRADIENT
specifies how the contour is displayed.

LINE
displays contour levels as unlabeled lines.

FILL
displays the area between the contour levels as filled.

Each contour interval is filled with one color.

GRADIENT
displays a smooth gradient of color to represent contour levels.

LINEFILL
combines the LINE and FILL types.

Each contour interval is filled with one color. Displays contour levels as
unlabeled lines.

LINEGRADIENT
combines the LINE and GRADIENT types.

Displays contour levels as unlabeled lines.

LABELEDLINE
adds labels to the LINE type, displaying contour levels as labeled lines.

LABELEDLINEFILL
adds labels to the LINEFILL type.

Each contour interval is filled with one color. Displays contour levels as lines
with labels showing contour level values.

LABELEDLINEGRADIENT
adds labels to the LINEGRADIENT type.

Displays contour levels as lines with labels showing contour level values.
 CONTOURPLOTPARM Statement 467

The following figure shows the effect of each of the values.

Default The GraphContour:DisplayOpts style reference.

Interactions The fill colors of the types that enable FILL or GRADIENT are
controlled by the COLORMODEL= option.

The line properties of the types that enable LINE or LABELEDLINE

are controlled by the LINEATTRS= option.

The label properties of the types that enable LABELEDLINE are

controlled by the LINELABELATTRS= and LINELABELBASELINE=
options.

If a DISCRETELEGEND statement is associated with the contour,

then the legend is NOT displayed if CONTOURTYPE= is set to FILL
or GRADIENT.

If a CONTINUOUSLEGEND statement is associated with the

contour, then the legend is NOT displayed if CONTOURTYPE is set
to LINE or LABELEDLINE.

Note For a contour plot of type FILL, LINEFILL, or LABELEDLINEFILL,

the color that is chosen for the first segment is the color for the N+1
inflection point and not the starting color in the color model.

GRIDDED=TRUE | FALSE
specifies whether the X and Y values are equally spaced in a rectangular grid. If
set to FALSE, then additional calculations are performed in order to complete
the grid.

For information about the algorithm used to calculate the grid, see “Mesoscale
Objective Map Analysis Using Weighted Time-Series Observations.”1

Default TRUE

Tip By default, the CONTOURPLOTPARM statement assumes that the X-Y

grid is complete and does not contain any missing or irregular values. If

1. Barnes, Stanley L March 1973. “Mesoscale Objective Map Analysis Using Weighted Time-Series Observations.” Technical
Memorandum (NOAA TM ERLNSSL-62), United States National Oceanic and Atmospheric Administration, Environmental
Research Labs, Norman, OK..
468 Chapter 6 / Plot Statements

the X-Y grid is not complete, then specify GRIDDED=FALSE so that the
plot calculates the values needed to complete the grid. Otherwise,
unexpected results might occur.

See “boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Z-column label. If a label is not defined, then the Z-column
name is used.

Restrictions This option applies only to an associated DISCRETELEGEND

statement.

This option applies to contour types LINE and LABELEDLINE only.

LEVELS=(contour-value-list)
specifies a list of contour level values.

(contour-value-list)
a space-separated list of numeric values, enclosed in parentheses.

Default The number of levels and the level values are determined internally
using the NHINT= or NLEVELS= option value.

Interaction This option overrides the NHINT= and NLEVELS= options.

Notes Values that are outside of the data range are ignored.

When the LEVELS= option is not specified, the data range might be
extended in order to accommodate the internally computed levels.

Example The following example specifies six levels for a contour plot:
levels=(0.0001 0.0004 0.0007 0.0010 0.0013 0.0016)

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the contour lines.

Default The GraphAxisLines style element.

Interaction This option is honored only if the CONTOURTYPE= displays lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

LINELABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the contour line labels.

Default The GraphValueText style element.

Interaction This option is honored only if the CONTOURTYPE= displays labels.

 CONTOURPLOTPARM Statement 469

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LINELABELBASELINE=HORIZONTAL | TANGENT
specifies the text alignment of the contour line labels. Each contour line has a
precomputed label point.

HORIZONTAL
specifies that each label is parallel to the X axis.

The label intersects its contour line and is centered at the label point.

TANGENT
specifies that each label is drawn tangent to the contour line at the label
point.

This reduces intersection with the contour line and causes labels to be tilted
at various angles in relation to the X axis.

Default HORIZONTAL

Interaction This option is honored only if the CONTOURTYPE= displays labels.

LINELABELFORMAT=format
specifies the format to use for the contour line labels.

Default The format associated with the Z column or BEST6. if no format is

assigned.

Interaction This option is honored only if the CONTOURTYPE= displays labels.

LINELABELPOSITION=MIDDLE | BEGIN | END

specifies the position for the contour line labels.

Default MIDDLE

Interaction This option is honored only if the CONTOURTYPE= option specifies

labels.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
indicate the colors associated with the Z values.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.
470 Chapter 6 / Plot Statements

NHINT=integer
specifies the suggested number of contour levels for the Z column.

Default 7

Interaction This option is ignored if the LEVELS= or NLEVELS= option is

specified.

Note The actual number of levels is adjusted to provide an appropriate

number of levels for the data.

NLEVELS=integer
specifies the actual number of contour levels for the Z column.

Default The number of levels is determined internally, using the NHINT=

value.

Interactions This option overrides the NHINT= option.

This option is ignored if CONTOURTYPE= GRADIENT.

This option is ignored if the LEVELS= option is specified.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.
 CONTOURPLOTPARM Statement 471

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
A CONTOURPLOTPARM statement uses the CONTOURTYPE= option to specify
the type of contour plot to generate. Contour types that display fills or gradients
but no contour lines can use only a CONTINUOUSLEGEND statement to represent
the contour level values in a legend. Contour types that display lines can use either
a CONTINUOUSLEGEND or DISCRETELEGEND statement to identify contour level
values.

By default, the CONTOURPLOTPARM statement assumes that the X-Y grid is

complete. If the grid is not complete, then set the GRIDDED= option to FALSE so
that the plot calculates the values needed to complete into a grid with a bounding
rectangle.

For filled or gradient contour types, small gaps might be visible between the axes
and the bounding box of the contour data. To eliminate these gaps, set the axis
options of the LAYOUT OVERLAY statement as follows:

XAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))

YAXISOPTS=(OFFSETMIN=0 OFFSETMAX=0
LINEAROPTS=(THRESHOLDMIN=0 THRESHOLDMAX=0))

Contour plots do not support the data tips that are enabled by the IMAGEMAP=
option in the ODS GRAPHICS statement.

Example: CONTOURPLOTPARM Statement

The following graph was generated by the “Example Program” on page 472:
 472 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph contourplotparm;
begingraph;
entrytitle "Contour Plot of Height and Weight";
layout overlay /
xaxisopts=(offsetmin=0 offsetmax=0
linearopts=(thresholdmin=0 thresholdmax=0))
yaxisopts=(offsetmin=0 offsetmax=0
linearopts=(viewmax=250
thresholdmin=0 thresholdmax=0));
contourplotparm x=height y=weight z=density /
contourtype=fill nhint=12
name="Contour" colormodel=twocolorramp;
continuouslegend "Contour" / title="Density";
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.gridded template=contourplotparm;
run;
 DENDROGRAM Statement 473

DENDROGRAM Statement
Creates a tree diagram that is typically used to display the results of a hierarchical clustering
analysis.

Restriction: The DENDROGRAM statement does not support CAS in-memory tables. In SAS
9.4M6 and earlier releases, unexpected results might occur when data stored in a
CAS in-memory table is used. Starting with SAS Viya 3.5 and SAS Studio 5.2, the
dendrogram is dropped when data stored in a CAS in-memory table is used.

Syntax
DENDROGRAM NODEID=column | expression
PARENTID=column | expression
CLUSTERHEIGHT=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
CLUSTERS=numeric-column | expression
specifies a numeric column containing the resultant number of clusters at
each node.
CUT=TRUE | FALSE
specifies whether the tree is to be cut.
CUTOPTS=(pruning-options)
specifies pruning options for cutting the dendrogram.
DATATRANSPARENCY=number
specifies the degree of the transparency of the dendrogram lines.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the dendrogram lines.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the dendrogram leaf axis.
TREETYPE=RECTANGULAR | TRIANGULAR
specifies the type of tree structure to draw.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
474 Chapter 6 / Plot Statements

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
dendrogram line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
NODEID=column | expression
specifies a column for the ID values of the nodes. Each node ID value must be
unique. If duplicate NODEID values are found, then the dendrogram is not
rendered . The column can be numeric or character, but it must be of the same
type and have the same formatted length as the ParentID column.

The maximum number of nodes that are supported by the dendrogram is

determined by the DISCRETEMAX= option in the ODS GRAPHICS statement.
The default value is DISCRETEMAX=1000. If the graph data contains more than
1000 discrete values, then the dendrogram is not drawn and a warning is written
to the SAS log. In that case, you can use the DISCRETEMAX= option to increase
the maximum number of discrete values that are allowed.

Note If the data is a SAS data set, the nodes are displayed in data order. If the
data is a CAS in-memory table, the nodes are displayed in ascending
order of the NODEID values by default.

PARENTID=column | expression
specifies a column for the parent ID values of the nodes. The column can be
numeric or character, but it must be of the same type and have the same
formatted length as the NodeID column.

CLUSTERHEIGHT=numeric-column | expression
specifies the column for the height values for each node.
 DENDROGRAM Statement 475

Optional Arguments
CLUSTERS=numeric-column | expression
specifies a numeric column containing the resultant number of clusters at each
node.

Interaction For this option to take effect, the pruning options in the CUTOPTS=
option must set TYPE=NCLUSTERS and specify a number for the
NCLUSTERS= setting.

CUT=TRUE | FALSE
specifies whether the tree is to be cut.

Default FALSE

Tip To set the properties of the CUT, use the CUTOPTS= option.

See “boolean ” on page xii for other Boolean values that you can use.

CUTOPTS=(pruning-options)
specifies pruning options for cutting the dendrogram. The following pruning-
options must be specified as a space-separated list of option = value pairs
enclosed in parentheses.

CUTHEIGHT=number
specifies the height at which the tree is to be pruned.

Default The tree is not pruned.

Requirement You must include pruning option TYPE=CUTHEIGHT with this

option.

Interaction This option is ignored when CUT=FALSE or when pruning

option TYPE=CUTHEIGHT is not explicitly specified.

Example CUTOPTS=(TYPE=CUTHEIGHT CUTHEIGHT=0.75)

NCLUSTERS=number
specifies the number of clusters to use for pruning the tree.

Default The tree is not pruned.

Interaction For this setting to take effect, pruning-option TYPE=NCLUSTERS

must also be set. In addition, the CLUSTERS= option must be
used, and the CUT= option must be set to TRUE.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the cut lines.

Default The GraphCutLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

476 Chapter 6 / Plot Statements

TYPE=CUTHEIGHT | NCLUSTERS
specifies which rule to use to prune the tree.

Default CUTHEIGHT

DATATRANSPARENCY=number
specifies the degree of the transparency of the dendrogram lines.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The CLUSTERHEIGHT= column label. If a label is not defined, then

the CLUSTERHEIGHT= column name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the dendrogram lines.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the dendrogram leaf axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.
 DENDROGRAM Statement 477

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
dendrogram line. If this option is used, then it replaces all of the information
that is displayed by default.

(role-list)
an ordered, space-separated list of unique DENDROGRAM roles, which
include NODEID , PARENTID , and CLUSTERHEIGHT .

Example The following example displays data tips for the columns assigned
to the roles NODEID and PARENTID.
TIP=(NODEID PARENTID)

NONE
suppresses data tips from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: NODEID , PARENTID , and
CLUSTERHEIGHT .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.
478 Chapter 6 / Plot Statements

Example TIPFORMAT=(CLUSTERHEIGHT=4.1)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(CLUSTERHEIGHT="Height")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TREETYPE=RECTANGULAR | TRIANGULAR
specifies the type of tree structure to draw.

Default RECTANGULAR

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.
 DENDROGRAM Statement 479

Details
A dendrogram is a tree diagram that is typically used to show the cluster
arrangements in hierarchical data. The DENDROGRAM statement supports clusters
with only a single root. If multiple roots are found in the data, then a warning is
written to the SAS log and the dendrogram is not drawn. The maximum number of
nodes that is supported by the DENDROGRAM statement is determined by the
DISCRETEMAX= option in the ODS GRAPHICS statement, which is 1000 by
default.

In the Graph Template Language, a DENDROGRAM plot typically appears by itself

in a LAYOUT OVERLAY container. You can overlay REFERENCELINE or
BANDPLOT statements on a DENDROGRAM, but overlaying other plot types might
produce unexpected results.

Using the DENDROGRAM statement in layouts where the axis ranges are merged
across cells might produce unexpected results.

Example: DENDROGRAM Statement

The following graph was generated by the “Example Program” on page 479:

Example Program
data clustree;
input id $ parent $7-12 height nClus;
 480 Chapter 6 / Plot Statements

label id="Cluster ID" parent="Parent ID";

datalines;
clus1 3 1
clus2 clus1 0.2 7
clus3 clus1 1.75 2
clus4 clus3 0.7 4
clus5 clus3 0.8 3
clus6 clus4 0.4 5
clus7 clus6 0.1 9
clus8 clus5 0.25 6
clus9 clus8 0.15 8
1 clus9 0 10
2 clus6 0 10
3 clus2 0 10
4 clus7 0 10
5 clus7 0 10
6 clus2 0 10
7 clus4 0 10
8 clus5 0 10
9 clus8 0 10
10 clus9 0 10
run;

proc template;
define statgraph dendrogram;
begingraph;
layout overlay;
dendrogram nodeID=id parentID=parent clusterheight=height;
endlayout;
endgraph;
end;
run;

proc sgrender data=clustree template=dendrogram;

run;

DENSITYPLOT Statement
Creates a univariate probability density curve computed from input data.

Requirement: If you include a histogram and a density plot in your graph and both plots specify
the same data column, you must either specify SCALE=DENSITY in the
HISTOGRAM statement or place the HISTOGRAM statement before the
DENSITYPLOT statement in your template code. Otherwise, the density plot
values might not scale properly to the histogram bins on the dependent axis. This
requirement applies even when the histogram and density plots are in different
cells in a multicell graph.
Tips: If the data density is not known, then use the KERNEL distribution option in the
DENSITYPLOT statement.
Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
 DENSITYPLOT Statement 481

SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.

Syntax
DENSITYPLOT numeric-column | expression </<distribution-option> <options>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the density curve and curve
label.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the density curve.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the density curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the density curve label.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the density curve label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the density curve label relative to the curve line.
482 Chapter 6 / Plot Statements

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label
block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate density curve for each unique group value in the
specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.
WEIGHT=numeric-column | expression
specifies a column that contains a density-curve calculation a priori
weight for each observation of the input data object.

Required Arguments
numeric-column
specifies a numeric column of data values that are used to calculate the
parameters for the probability distribution.

expression
specifies an expression that calculates values when those values are not stored
in the data.

Optional Arguments
CURVELABEL="string"
specifies a label for the density curve.

Default No curve label is displayed.

 DENSITYPLOT Statement 483

Restriction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the density curve label.

Default The GraphValueText style element.

Interaction For this option to take effect, the CURVELABEL=curvelabel option

must also be used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the density curve label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the DENSITYPLOT is used in multi-

cell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes can be external to the grid.

Interaction This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve label
appears. For more information, see “Location and Position of Curve
Labels” on page 213.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the density curve label relative to the curve line.

AUTO
positions the density label automatically near the end of the density curve
along unused axes whenever possible (typically Y2 or X2) to avoid collision
with tick values.

Restriction This option is used only when

CURVELABELLOCATION=OUTSIDE.

MAX
forces the density label to appear near maximum density X-values (typically,
to the right).
484 Chapter 6 / Plot Statements

MIN
forces the density label to appear near minimum density X-values (typically,
to the left).

START
forces the density label to appear near the beginning of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the density label to appear near the end of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the density
label appears. For more information, see “Location and Position of
Curve Labels” on page 213.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the curve label might fall outside of the graph area. In
that case, the curve label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When
a curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

 DENSITYPLOT Statement 485

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the curve label. In that case, all of the specified split characters
together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve

label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters,
then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.

TRUE
drops the split characters from the curve label text.

FALSE
includes the split characters in the curve label text.
486 Chapter 6 / Plot Statements

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a curve label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

 DENSITYPLOT Statement 487

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the curve for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the density curve and curve label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation. The marker for that observation is drawn n times,
which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a separate density curve for each unique group value in the specified
column.
488 Chapter 6 / Plot Statements

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value is represented in the plot by a different line
color and pattern. The line color is determined by the ContrastColor
attribute of the GraphData1–GraphDataN and GraphMissing style
elements. The line pattern is determined by the LineStyle attribute of
the GraphData1–GraphDataN and GraphMissing style elements.

Notes The group values are mapped in the order in which they appear in the
data.

If you specify a column in a SAS data set, the visual attributes for each
group value are assigned in data order. If you specify a column in a CAS
in-memory table, the visual attributes for each group value are
assigned in ascending order of the group column character values or of
unformatted numeric values.

Tips You can individually override the representations that are used to
identify the groups. For example, in some ODS styles, each distinct
group value is represented by a different line color and pattern. In that
case, you can use the PATTERN= setting on the LINEATTRS= option to
assign the same line pattern to all of the curves.

By default, ODS Graphics limits the number of groups to 1000. When

this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change the
maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

“Example 3: Grouped Density Plot” on page 496

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.
 DENSITYPLOT Statement 489

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the MISSING= system option changes the default missing
character, or a user-defined format is applied to the group value. In
those cases, the attributes of the missing group value are
determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the density curve.

Defaults For non-grouped data, the GraphFit style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphFit:LineThickness style reference.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.
490 Chapter 6 / Plot Statements

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Note The columns assigned to the X and Y roles are automatically

included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X and Y roles are automatically included
in the data tip information.

WEIGHT=numeric-column | expression
specifies a column that contains a density-curve calculation a priori weight for
each observation of the input data object.

Requirement The value must be nonnegative.

 DENSITYPLOT Statement 491

Note The values of the weight variable must be greater than zero. If the
weight value for an observation is zero, negative, or missing, that
observation is deleted from the analysis.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Distribution Options
KERNEL (<C=positive-number> <WEIGHTFUNCTION=NORMAL | QUADRATIC |
TRIANGULAR> <MAXPOINTS=positive-integer>)
specifies a nonparametric kernel density estimate. By default, the kernel density
estimates are computed by minimizing the approximate mean integrated square
error (AMISE). It is the same method that the UNIVARIATE procedure uses to
compute kernel density estimates. The general form of the kernel density
estimator is as follows.
n
x − xi
fλ x =
100h %
nλ ∑ K0 λ
i=1

In the equation, K0 ⋅ is the weight function, λ is the bandwidth, n is the sample

size, and xi is the ith observation. You can use the C= suboption to specify the
bandwidth and the WEIGHTFUNCTION= suboption to specify the weight
function K0 ⋅ .

For more information about the kernel density estimate, see Kernel Density
Estimates in the UNIVARIATE procedure description in Base SAS Procedures
Guide: Statistical Procedures.

TIP Use the KERNEL distribution option when the data density is not
known.
492 Chapter 6 / Plot Statements

C=positive-number
specifies a positive number (0 < number <= 100) that represents the
standardized bandwidth.

The value of λ, referred to as the bandwidth parameter, determines the

degree of smoothness in the estimated density function. You specify λ
indirectly by specifying a standardized bandwidth c with the C=kernel-
option. If Q is the interquartile range and n is the sample size, then c is
related to λ by the following equation:
1
λ = cQn − 5

Default Calculated from the data as the bandwidth that minimizes the
AMISE.

Range 0 < number <= 100

WEIGHTFUNCTION=NORMAL | QUADRATIC | TRIANGULAR

specifies one of the weight functions NORMAL, QUADRATIC, or
TRIANGULAR.

The following table shows the formulas for the weight functions.

Weight Function Formula

NORMAL 1 1
K0 t = exp − 2 t2 for ‐∞ < t < ∞
2π

QUADRATIC 3
K0 t = 4 1 − t2 for t ≤ 1

TRIANGULAR K0 t = 1 − t for t ≤ 1

Default NORMAL

Note In prior SAS releases, the weight function was specified with the
WEIGHT= option. In SAS 9.4 and later releases, the WEIGHT=
option is not valid as a distribution option. You must use the
WEIGHTFUNCTION= option instead.

MAXPOINTS=positive-integer
specifies the maximum number of points generated for the curve.

Default 512

Requirements Prior to SAS 9.4M5, positive-integer must be greater than or

equal to 2.

Starting with SAS 9.4M5, positive-integer must be greater than

or equal to 16.
 DENSITYPLOT Statement 493

NORMAL (<MU=number> <SIGMA=number> <MAXPOINTS=number>)

specifies a normal density estimate, with mean and standard deviation. The
fitted density function equation is as follows:
100h % 1 x−μ 2
px = exp − 2 σ
for ‐∞ < x < ∞
σ 2π

In the equation, μ is the mean and σ is the standard deviation (σ > 0). You can
specify μ with the MU= suboption and σ with the SIGMA= suboption. By default,
ODS estimates μ with the sample mean and σ with the sample standard
deviation.
For more information, see the discussion of Kernel Density Estimates for the
UNIVARIATE procedure in the documentation for Base SAS.

MU=number
specifies the mean.

Default The value is calculated from the data.

SIGMA=number
specifies the standard deviation.

Default The value is calculated from the data.

MAXPOINTS=positive-integer
specifies the maximum number of points generated for the curve.

Default 200

Requirements Prior to SAS 9.4M5, positive-integer must be greater than or

equal to 2.

Starting with SAS 9.4M5, positive-integer must be greater than

or equal to 16.

Details
A typical DENSITYPLOT statement specifies either the NORMAL or the KERNEL
distribution option. If no distribution option is specified, then the NORMAL() option
is used. The following syntax explicitly shows the default case:
DENSITYPLOT numeric-column / NORMAL()

To specify a kernel distribution, use the following plot syntax:

DENSITYPLOT numeric-column / KERNEL()

If more than one distribution option is specified, then the last distribution option
specified is used.
 494 Chapter 6 / Plot Statements

Examples

Example 1: Overlaid Density Plots

The following graph was generated by the “Example Program” on page 494:

Example Program
When used as a standalone plot or overlaid with other density plots, the dependent
axis shows the computed density values.
proc template;
define statgraph densityplot1;
begingraph;
entrytitle "Fitted Density Curves";
entrytitle "of Patient Weight";
entryfootnote halign=left "Framingham Heart Study";
layout overlay;
densityplot weight / normal()
lineattrs=graphfit name="n" legendlabel="Normal";
densityplot weight / kernel()
lineattrs=graphfit2 name="k" legendlabel="Kernel";
discretelegend "n" "k";
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.heart template=densityplot1;
label weight="Patient Weight";
 DENSITYPLOT Statement 495

run;

Example 2: Density Plot and Histogram

When one or more density plots are overlaid on a histogram, the dependent axis
shows the statistic indicated by the histogram’s SCALE= option. The area under
each density curve is equal to the area of the histogram. The following graph was
generated by the “Example Program” on page 495:

Example Program
proc template;
define statgraph densityplot2;
begingraph;
entrytitle "Patient Weight Distribution";
entrytitle "with Fitted Normal Curve";
entryfootnote halign=left "Framingham Heart Study";
layout overlay;
histogram weight / primary=true scale=count;
densityplot weight / normal() lineattrs=graphfit;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.heart template=densityplot2;
label weight="Patient Weight";
run;
 496 Chapter 6 / Plot Statements

Example 3: Grouped Density Plot

The following graph was generated by the “Example Program” on page 496:

Example Program
proc template;
define statgraph densityplot;
begingraph;
entrytitle "Highway Mileage Distribution by Origin";
layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot))
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
densityplot mpg_highway / name="densityplot" group=origin;
discretelegend "densityplot" / title="Origin:";
endlayout;
endgraph;
end;
run;

proc sgrender template=densityplot data=sashelp.cars;

run;

DROPLINE Statement
Creates a horizontal or vertical drop line from a point to an axis.

Requirement: A DROPLINE statement must be used within a 2-D layout (for example, an
OVERLAY, OVERLAYEQUATED, DATALATTICE, or DATAPANEL layout).
 DROPLINE Statement 497

Syntax
DROPLINE X=x-axis-value | column | expression
Y=y-axis-value | column | expression </options>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the line are considered when determining
the data ranges for the axes.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the drop line.
DATATRANSPARENCY=number
specifies the degree of the transparency of the drop line.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the drop line.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset all drop lines from discrete X values, or Y
values, or both.
DROPTO=X | Y | BOTH
specifies the axis to which the line is dropped.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LABEL="string" | string-column
specifies a label for the point(s) indicated by the X= and Y= arguments.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the drop line label(s).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
498 Chapter 6 / Plot Statements

Required Arguments
X=x-axis-value | column | expression
specifies the X coordinate of the drop line(s).

Requirement Values must agree in type with the X-axis data type. For example,
you should use numeric SAS date or time values (or SAS date/time
constants) for a time axis.

Note When a character value is specified, leading blanks are honored

and trailing blanks are ignored when the specified value is
compared with the data values.

Tip By default, if a specified value is outside of the X-axis data range,

then the data range is extended to include the value. This behavior
can be changed with the CLIP= option.

Y=y-axis-value | column | expression

specifies the Y coordinate of the drop line(s).

Requirement Values must agree in type with the Y-axis data type.

Note When a character value is specified, leading blanks are honored

and trailing blanks are ignored when the specified value is
compared with the data values.

Tip By default, if a specified value is outside of the Y-axis data range,

then the data range is extended to include the value. This behavior
can be changed with the CLIP= option.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the line are considered when determining the
data ranges for the axes.

FALSE
specifies that the reference line values are to be considered when the axis
range is determined. The reference lines are drawn as follows based on the
axis type:
n For a discrete axis, the reference line values that are not already on the
axis are added to the end of the axis data list. When applicable, the axis
values are then sorted:
o If the axis values are numeric values, then they are sorted ordinally.
o If the axis values are character values and a sorting option is applied to
the axis, then they are sorted as specified by the sorting option.
Reference lines are then drawn at the specified locations.
n For a linear, log, or time axis, a new axis data list is created by performing
a mathematical union of the data values and the reference line values.
The reference lines are then drawn at the locations specified.
 DROPLINE Statement 499

TRUE
specifies that the reference line values are not to be considered when the
axis range is determined. The reference lines are drawn as follows based on
the axis type:
n For a discrete axis, if the reference line value exactly matches a value on
the axis, then a reference line is drawn at that location. Otherwise, the
reference line is not drawn.

Note: If the axis values are formatted, then the reference line value must
exactly match the formatted axis value in order for the line to be drawn.

n For a linear, log, or time axis, if the reference line value is within the axis
data range, then the reference line is drawn at the specified location.
Otherwise, the reference line is not drawn.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the drop line. The following figure shows
drop lines with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data skin
is not applied to that plot. In that case, use the DATASKINMAX=
option in your ODS GRAPHICS statement to increase the maximum
limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.
500 Chapter 6 / Plot Statements

DATATRANSPARENCY=number
specifies the degree of the transparency of the drop line.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the point labels.

DISCRETEOFFSET=number
specifies an amount to offset all drop lines from discrete X values, or Y values,
or both.

Default 0 (no offset, all drop lines are centered on discrete X values, or
discrete Y values, or both)

Range -0.5 to +0.5 where 0.5 represents half the distance between discrete
ticks. If the X axis is discrete, then a positive offset is to the right. If
the Y axis is discrete, then a positive offset is up. If REVERSE=TRUE
on the X or Y axis, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DROPTO=X | Y | BOTH
specifies the axis to which the line is dropped.

X
draws one or more droplines to an X axis.

Tip The XAXIS= option determines whether the X axis or X2 axis is the
endpoint for the line.

Y
draws one or more droplines to a Y axis.

Tip The YAXIS= option determines whether the Y axis or Y2 axis is the
endpoint for the line.

BOTH
draws one or more droplines to both axes.

Note This option is valid starting in SAS 9.4M1.

Default X

LABEL="string" | string-column
specifies a label for the point(s) indicated by the X= and Y= arguments.
 DROPLINE Statement 501

Default No label is specified

Interaction If drawing multiple drop lines using X=column or Y=column, then you
can assign corresponding labels by using a column to define the
labels.

Note Starting with SAS 9.4M1, space is reserved at the maximum end of
the X axis to accommodate the length of the labels regardless of
where the labels appear in the plot.

Tips You can use the OFFSETMAX= axis option to adjust the amount of
space that is reserved on the X axis for the labels.

The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the drop line label(s).

Default The GraphValueText style element.

Interaction For this option to have any effect, the LABEL= option must also be
specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the drop line.

Default The GraphReference style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.
502 Chapter 6 / Plot Statements

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
A drop line is always drawn perpendicular from the specified point to the X
(bottom), X2 (top), Y (left), or Y2 (right) axis. Axis offsets do not apply to drop
lines, so they always meet the axis line.

The DROPTO= option controls whether a horizontal or vertical drop line is created.
DROPTO=X specifies the horizontal axis for a vertical drop line, and DROPTO=Y
specifies the vertical axis for a horizontal drop line.

A DROPLINE statement can be used only within a 2-D overlay-type layout

(OVERLAY, OVERLAYEQUATED, PROTOTYPE). Another plot statement that is
derived from data values that provide boundaries for the axis area must be included
in the layout. For example, it can be used with a scatter plot or a histogram. You can
generate multiple drop lines by specifying a column for X and Y. The column type
(numeric or string) must agree with the type of data presented on the axis.

To generate both a vertical and a horizontal drop line from a single point, use
multiple DROPLINE statements.
 DROPLINE Statement 503

Example: DROPLINE Statement

The following graph was generated by the “Example Program” on page 503. The
graph shows two DROPLINE statements originating from the same point (X=3,
Y=5). One statement uses DROPTO=X and the other uses DROPTO=Y.

Example Program
proc template;
define statgraph dropline;
begingraph;
entrytitle "Drop lines at Inflection Point";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
dropline x=3 y=5 / dropto=x
lineattrs=(color=blue pattern=dot) label="(3,5)";
dropline x=3 y=5 / dropto=y
lineattrs=(color=blue pattern=dot);
endlayout;
endgraph;
end;
run;

data test;
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
 504 Chapter 6 / Plot Statements

run;

proc sgrender data=test template=dropline;

run;

ELLIPSE Statement
Creates a confidence ellipse computed from input data.

Requirements: An ELLIPSE statement must be used in a two-dimensional overlay-type layout (for

example, an OVERLAY, OVERLAYEQUATED, or PROTOTYPE layout).
The ELLIPSE statement must be overlaid with another plot that is derived from
data values that provide boundaries for the axis area. It is typically overlaid with a
scatter plot.

Syntax
ELLIPSE X=numeric-column | expression
Y=numeric-column | expression < /options >;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when
determining the data ranges for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and
outline.
DISPLAY=(display-options) | STANDARD | ALL
specifies which ellipse features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the ellipse.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the ellipse outline.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
 ELLIPSE Statement 505

specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate ellipse for each unique group value in the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
ALPHA=positive-number
sets a significance value for the confidence level to compute for the
ellipse.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.
TYPE=MEAN | PREDICTED
specifies the type of ellipse.

Required Arguments
X=numeric-column | expression
specifies the numeric column for the X values.

Y=numeric-column | expression
specifies the numeric column for the Y values.

Optional Arguments
ALPHA=positive-number
sets a significance value for the confidence level to compute for the ellipse.

Default 0.05

Range 0 < number < 1

Note ALPHA=0.05 represents a 95% confidence level.

See TYPE=
506 Chapter 6 / Plot Statements

CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the
data ranges for the axes.

FALSE
the data for the ellipse contribute to the data range for each axis.

Each axis might be extended to force the display of the entire ellipse.

TRUE
the data for the ellipse are ignored when establishing axis scales.

Each axis scale is determined by the other plots in the parent layout. This
might result in the ellipse not being entirely displayed (clipped) if its data
range is not within the data ranges of the other plots.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
ellipse fill. You can combine this option with FILLATTRS= to set one
transparency for the ellipse outline but a different transparency for the
ellipse fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which ellipse features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE
displays the ellipse outline

FILL
displays the ellipse fill color

FILLPATTERN
displays the ellipse fill pattern

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
 ELLIPSE Statement 507

FILLPATTERNATTRS= option to specify a fill pattern for the

ellipse.

STANDARD
specifies OUTLINE only

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Default GraphEllipse:DisplayOpts style reference.

Tip Use FILLATTRS= and OUTLINEATTRS= to control the appearance of

the ellipse outline and fill color. Starting with SAS 9.4M1, use the
FILLPATTERNATTRS= on page 507 option to control the appearance
of the ellipse fill pattern.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the ellipse.

Default The GraphDataDefault style element.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tip The DATATRANSPARENCY= option sets the transparency for the

ellipse fill and ellipse outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outline but
a different transparency for the fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
508 Chapter 6 / Plot Statements

information, see “Attribute Rotation Patterns” in SAS Graph

Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation. The marker for that observation is drawn n times,
which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a separate ellipse for each unique group value in the specified column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct ellipse is created for each group value by varying the visual attributes
of the ellipse display features. The display features are controlled by the current
ODS style or by the DISPLAY= option. The default group appearance for each
display feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.
 ELLIPSE Statement 509

Display
Feature 1 Style Attributes That Control Default Group Appearance

If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.

Outline ContrastColor and LineStyle attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.

1 The default display features are determined by the DisplayOpts attribute of the GraphEllipse
style element.
2 Fill patterns are valid for ellipse plots starting with SAS 9.4M5.

Notes The group values are mapped in the order in which they appear in the
data.

If you specify a column in a SAS data set, the visual attributes for each
group value are assigned in data order. If you specify a column in a CAS
in-memory table, the visual attributes for each group value are assigned
in ascending order of the group column character values or of
unformatted numeric values.

Tips You can individually override the representations that are used to
identify the groups. For example, in some ODS styles, each distinct
group value is represented by a different line color and pattern. In that
case, you can use the PATTERN= setting on the OUTLINEATTRS= option
to assign the same line pattern to all of the ellipses.

By default, ODS Graphics limits the number of groups to 1000. When

this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change the
maximum number of groups that can be processed.

See “Example 2: 95% Confidence Ellipses by Plant Species” on page 514

“DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the MISSING= system option changes the default missing
character, or a user-defined format is applied to the group value. In
those cases, the attributes of the missing group value are
510 Chapter 6 / Plot Statements

determined by a GraphData1–GraphDataN style element instead of

by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the ellipse outline.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the LineThickness
attribute of the GraphDataDefault style element.

Interaction For this option to have any effect, the outlines must be enabled by
the ODS style or the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

TYPE=MEAN | PREDICTED
specifies the type of ellipse.

MEAN
specifies a confidence ellipse of the mean

PREDICTED
specifies a prediction ellipse for a new observation

Default MEAN

See ALPHA= option for specifying a confidence level.

 ELLIPSE Statement 511

For statistical details about how the ellipse is calculated, see

“Confidence and Prediction Ellipses ” on page 511.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Description
The ELLIPSE statement can be used only within 2-D overlay-type layouts. It
computes an ellipse for a set of points specified by the X and Y columns and a
confidence level specified by the ALPHA= option. Use the TYPE= option to control
whether a predicted or confidence ellipse is generated.

Confidence and Prediction Ellipses

Two types of ellipses can be computed for the input data (where observations
correspond to points in a scatter plot). One is a confidence ellipse for the
population mean (TYPE=MEAN), and the other is a prediction ellipse for a new
observation (TYPE=PREDICT). Both assume a bivariate normal distribution.

Let Z and S be the sample mean and sample covariance matrix of a random sample
of size n from a bivariate normal distribution with mean μ and covariance matrix Σ.
The variable Z − μ is distributed as a bivariate normal variate with mean zero and
1
covariance ( n )Σ, and it is independent of S. Using Hotelling’s T 2 statistic, which is
defined as follows:

T 2 = n(Z − μ)′S−1(Z − μ)

A 100(1 − α) % confidence ellipse for μ is computed from the following equation:

n −1 2
n−1
(Z − μ)′S (Z − μ) = F
n − 2 2, n − 2
(1 − α)
512 Chapter 6 / Plot Statements

where F2, n − 2(1 − α) is the (1 − α) critical value of an F distribution with degrees of

freedom 2 and n − 2.

A prediction ellipse is a region for predicting a new observation in the population. It

also approximates a region containing a specified percentage of the population.

Denote a new observation as the bivariate random variable Znew. The following
variable:
Znew − Z = (Znew − μ) − (Z − μ)

is distributed as a bivariate normal variate with mean zero (the zero vector) and
1
covariance (1 + n )Σ, and it is independent of S. A 100(1 − α) % prediction ellipse is
then given by the following equation:
n −1 2(n + 1)
n−1
(Z − μ)′S (Z − μ) = F
n − 2 2, n − 2
(1 − α)

The family of ellipses generated by different critical values of the F distribution has
a common center (the sample mean) and common major and minor axis directions.

The shape of an ellipse depends on the aspect ratio of the plot. The ellipse
indicates the correlation between the two variables if the variables are
standardized (by dividing the variables by their respective standard deviations). In
this situation, the ratio between the major and minor axis lengths is:
1+ r
1− r

In particular, if r = 0, then the ratio is 1, which corresponds to a circular confidence

contour and indicates that the variables are uncorrelated. A larger value of the ratio
indicates a larger positive or negative correlation between the variables.

Examples

Example 1: Prediction Ellipses

The following graph was generated by the “Example Program” on page 513:
 ELLIPSE Statement 513

Example Program
This example overlays two ELLIPSE statements on a SCATTERPLOT of the same
data. Both ELLIPSE statements use TYPE= PREDICTED. One ELLIPSE statement
uses ALPHA= 0.2, and the other uses ALPHA=0.05.
proc template;
define statgraph ellipse;
begingraph;
entrytitle "Prediction Ellipses";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth /
datatransparency=0.5;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.2
name="p80" legendlabel="80%"
outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.05
name="p95" legendlabel="95%"
outlineattrs=graphconfidence2;
discretelegend "p80" "p95" /
location=inside autoalign=(topleft);
endlayout;
entryfootnote halign=left "Fisher's Iris Data";
endgraph;
end;
run;
proc sgrender data=sashelp.iris template=ellipse;
run;
 514 Chapter 6 / Plot Statements

Example 2: 95% Confidence Ellipses by Plant Species

The following graph was generated by the “Example Program” on page 514:

Example Program
proc template;
define statgraph ellipsegroup;
begingraph;
entrytitle "95% Prediction By Species";
layout overlay;
scatterplot x=petallength y=petalwidth / name="sp"
group=species;
ellipse x=petallength y=petalwidth / group=species
type=predicted alpha=0.05 name="p95";
mergedlegend "sp" "p95" / location=inside
autoalign=(topleft) across=1;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=ellipsegroup;

run;

ELLIPSEPARM Statement
Creates an ellipse specified by slope, axis, and origin parameters.
 ELLIPSEPARM Statement 515

Requirements: An ELLIPSEPARM statement must be used in a two-dimensional overlay-type

layout (for example, an OVERLAY, OVERLAYEQUATED, or PROTOTYPE layout).
The ELLIPSEPARM statement must be overlaid with another plot that is derived
from data values that provide boundaries for the axis area. It is typically overlaid
with a scatter plot.
Tip: You can generate a single ellipse by specifying a constant for each required
argument. You can generate multiple ellipses by specifying a numeric column for
any or all required arguments. If any of the SEMIMAJOR= , SEMIMINOR= ,
XORIGIN= or YORIGIN= constants or columns contains a missing value, then no
ellipse is drawn. To request a vertical major axis, specify SLOPE=. (missing value) as
a constant or column value.

Syntax
ELLIPSEPARM SEMIMAJOR=number | numeric-column | expression
SEMIMINOR=number | numeric-column | expression
SLOPE=number | numeric-column | expression
XORIGIN=number | numeric-column | expression
YORIGIN=number | numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when
determining the data ranges for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and
outline.
DISPLAY=(display-options) | STANDARD | ALL
specifies which ellipse features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the ellipse.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
INDEX=positive-integer-column | expression
specifies indices for mapping ellipse attributes (fill and outline) to one of
the GraphData1–GraphDataN style elements.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the ellipse outline.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
516 Chapter 6 / Plot Statements

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate ellipse for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
SEMIMAJOR=number | numeric-column | expression
specifies half the length of the major axis for the ellipse. Values are in the same
units as the data. The SEMIMAJOR value can be greater than, smaller than, or
equal to the SEMIMINOR value.

Restriction If a numeric column is specified and the column contains negative

values, then no ellipses are drawn. In SAS 9.4M2 and in earlier
releases, the same is true if the column contains missing values.
Starting with SAS 9.4M3, if the column contains a missing value, the
ellipse for that observation is not drawn. However, the remaining
ellipses that have nonmissing values for this argument and for
SEMIMINOR= are drawn.

SEMIMINOR=number | numeric-column | expression

specifies half the length of the minor axis for the ellipse. Values are in the same
units as the data. The SEMIMINOR value can be greater than, smaller than, or
equal to the SEMIMAJOR value.

Restriction If a numeric column is specified and the column contains negative

values, then no ellipses are drawn. In SAS 9.4M3 and in earlier
releases, the same is true if the column contains missing values.
Starting with SAS 9.4M3, if the column contains a missing value, the
ellipse for that observation is not drawn. However, the remaining
ellipses that have nonmissing values for this argument and for
SEMIMAJOR= are drawn.

SLOPE=number | numeric-column | expression

specifies the slope of the major axis for the ellipse. Slope can be positive or
negative.
 ELLIPSEPARM Statement 517

Note The slope value is in the data space and might or might not be maintained
in the screen space. Thus, setting SLOPE=1 does not always generate a 45-
degree line on the screen.

Tip Setting SLOPE=0 creates a major axis parallel to the X axis. Setting
SLOPE=. (missing value) creates a major axis parallel to the Y axis.

XORIGIN=number | numeric-column | expression

specifies the X coordinate of the center of the ellipse. Values are in the units of
the data.

Tip By default, if the value specified for the XORIGIN= option is outside of the
X-axis data range, then the data range is extended to include the specified
point. This behavior can be changed with the CLIP= option.

YORIGIN=number | numeric-column | expression

specifies the Y coordinate of the center of the ellipse. Values are in the units of
the data.

Tip By default, if the value specified for the YORIGIN= option is outside of the
Y-axis data range, then the data range is extended to include the specified
point. This behavior can be changed with the CLIP= option.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the ellipse are considered when determining the
data ranges for the axes.

FALSE
The data for the ellipse contribute to the data range for each axis.

Each axis might be extended to force the display of the entire ellipse.

TRUE
The data for the ellipse are ignored when establishing axis scales.

Each axis scale is determined by the other plots in the parent layout. This
might result in the ellipse not being entirely displayed (clipped) if its data
range is not within the data ranges of the other plots.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

DATATRANSPARENCY=number
specifies the degree of the transparency of the ellipse fill color and outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
ellipse fill. You can combine this option with FILLATTRS= to set one
518 Chapter 6 / Plot Statements

transparency for the ellipse outline but a different transparency for the
ellipse fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which ellipse features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE
displays the ellipse outline

FILL
displays a filled ellipse

FILLPATTERN
displays the ellipse fill pattern

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the
ellipse.

STANDARD
displays OUTLINE only

ALL
displays all features: FILL, FILLPATTERN, and OUTLINE

Default GraphEllipse:DisplayOpts style reference.

Tip Use FILLATTRS= and OUTLINEATTRS= to control the appearance of

the ellipse outline and fill color. Starting with SAS 9.4M5, use the
FILLPATTERNATTRS= on page 519 option to control the appearance
of the fill pattern.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the ellipse.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the GraphData1–GraphDataN style elements.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tip The DATATRANSPARENCY option sets the transparency for the

ellipse fill and ellipse outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outline but
a different transparency for the fill. Example:
 ELLIPSEPARM Statement 519

datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

GROUP=column | discrete-attr-var | expression

creates a separate ellipse for each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct ellipse is created for each group value by varying the visual attributes
of the ellipse display features. The display features are controlled by the current
ODS style or by the DISPLAY= option. The default group appearance for each
display feature is shown in the following table.
520 Chapter 6 / Plot Statements

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.
If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.

Outline ContrastColor and LineStyle attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.

1 The default display features are determined by the DisplayOpts attribute of the GraphEllipse
style element.
2 Fill patterns are valid for ellipse plots starting with SAS 9.4M5.

The representations that are used to identify the groups can be overridden
individually. For example, each distinct group value might be represented by a
different line pattern for the ellipses, but the PATTERN= suboption of the
OUTLINEATTRS= option could be used to assign the same line pattern to all
ellipse outlines, letting outline color distinguish group values.

Defaults If ellipse fills are enabled by the ODS style or by the DISPLAY=
option, then each distinct group value is represented in the plot by a
different fill color. Starting with SAS 9.4M5, each distinct group
value can be represented by a different fill pattern, with or without a
different fill color. The fill color and fill pattern are defined by the
Color and FillPattern attributes of the GraphData1–GraphDataN and
GraphMissing style elements.

If only ellipse outlines are enabled by the ODS style or by the

DISPLAY= option, then each distinct group value is represented in
the plot by a different ellipse outline color and line style. The outline
colors and line styles are defined by the ContrastColor and
LineStyle attributes of the GraphData1–GraphDataN style elements.

Restriction When the GROUP= option is specified, the group value must be a
character or numeric column. For each group value, there must be a
numeric column that does not contain missing values for
SEMIMAJOR=, SEMIMINOR=, XORIGIN=, and YORIGIN=.The
SLOPE=column can contain missing values. Under these
circumstances, an ellipse is drawn for each group value.

Interaction The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
 ELLIPSEPARM Statement 521

are assigned in ascending order of the group column character

values or of unformatted numeric values.

Tip By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping ellipse attributes (fill and outline) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.
522 Chapter 6 / Plot Statements

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the ellipse outline.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.
 ELLIPSEPARM Statement 523

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The ELLIPSEPARM statement plots an ellipse with specified semimajor and
semiminor axis lengths, a specified slope for the major axis, and the ellipse center.
The ellipse center is the point of intersection of the semimajor and semiminor axes.
It can be used only within a two-dimensional overlay-type layout (for example, an
OVERLAY, OVERLAYEQUATED, or PROTOTYPE layout). The ELLIPSEPARM
statement does not perform computations on input data to derive the location and
shape of the ellipse. Another plot that is derived from data values that provide
boundaries for the axis area must be included in the layout. It is typically overlaid
with a scatter plot. See “ELLIPSE Statement” on page 504 for information about
computed ellipses.

Example: ELLIPSEPARM Statement

The following graph was generated by the “Example Program” on page 524:
 524 Chapter 6 / Plot Statements

Overview
This is a simplified version of the CorrLoadPlot template for PROC PLS in the
SAS/STAT product. It consists of overlaid scatter plots of the scores of the first
two factors, the loadings of the model effects, and the loadings of the dependent
variables. The loadings are scaled so that the amount of variation in the variables
that is explained by the model is proportional to the distance from the origin;
circles indicating various levels of explained variation are also overlaid.

The circles are drawn with ELLIPSEPARM statements by setting the SEMIMAJOR
and SEMIMINOR lengths to be the same. Notice that all circles are concentric
because they share the same origin (0,0). The radius of each circle is passed
dynamically.

Also note that an OVERLAYEQUATED layout was used to force the length of unit
intervals on both axes to be the same.

The input data shown is representative of that computed by the PLS procedure for
the Correlation Loadings Plot shown in the PLS procedure example Spectrometric
Calibration in SAS/STAT User’s Guide.

Example Program
Here is the SAS program for this example.
proc template;
define statgraph ellipseparm;
 ELLIPSEPARM Statement 525

dynamic RADIUS1 RADIUS2 RADIUS3 RADIUS4 ;

begingraph;
entrytitle "Correlation Loading Plot";
layout overlayequated / equatetype=square
commonaxisopts=
(tickvaluelist=(-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1)
viewmin=-1 viewmax=1)
xaxisopts=
(label="Factor 1" offsetmin=0.05 offsetmax=0.05)
yaxisopts=
(label="Factor 2" offsetmin=0.05 offsetmax=0.05);
ellipseparm semimajor=RADIUS1 semiminor=RADIUS1 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS2 semiminor=RADIUS2 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS3 semiminor=RADIUS3 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;
ellipseparm semimajor=RADIUS4 semiminor=RADIUS4 slope=0
xorigin=0 yorigin=0 / clip=true display=(outline)
outlineattrs=(pattern=dash) datatransparency=0.75;

scatterplot x=xcirclelabel y=ycirclelabel / primary=true

markercharacter=circlelabel datatransparency=0.75 ;
scatterplot x=corr1 y=corr2 / name="ScatterVars"
group=corrgroup markercharacter=corrlabel;
discretelegend "ScatterVars";
endlayout;
endGraph;
end;
run;

data corrplot;
infile cards missover dlm="#";
length CorrGroup $ 18 CorrLabel CorrLabel CircleLabel $ 8;
input Corr1 Corr2 CorrGroup CorrLabel xCircleLabel
yCircleLabel CircleLabel;
datalines;
-0.179 # -0.268 # Predictor Loading # S1 # 0 # 0.5 # 25%
0.105 # 0.332 # Predictor Loading # S2 # 0 # -0.5 # 25%
-0.654 # 0.094 # Predictor Loading # S3 # 0 # 0.707 # 50%
-0.653 # 0.685 # Predictor Loading # S4 # 0 # -0.707 # 50%
0.096 # 0.059 # Predictor Loading # S5 # 0 # 0.866 # 75%
0.132 # 0.036 # Predictor Loading # L1 # 0 # -0.866 # 75%
0.087 # 0.156 # Predictor Loading # L2 # 0 # 1 # 100%
0.940 # 0.160 # Predictor Loading # L3 # 0 # -1 # 100%
0.607 # -0.350 # Predictor Loading # L4
0.096 # 0.059 # Predictor Loading # L5
-0.111 # -0.534 # Predictor Loading # P1
0.003 # 0.256 # Predictor Loading # P2
0.293 # 0.551 # Predictor Loading # P3
-0.480 # 0.643 # Predictor Loading # P4
-0.096 # -0.059 # Predictor Loading # P5
0.946 # 0.279 # Response Loading # log_RAI
 526 Chapter 6 / Plot Statements

-0.196 # 0.403 # Observation # 1

0.020 # -0.001 # Observation # 2
-0.195 # 0.324 # Observation # 3
0.021 # -0.079 # Observation # 4
-0.009 # -0.274 # Observation # 5
0.567 # 0.294 # Observation # 6
-0.096 # -0.059 # Observation # 7
0.258 # 0.210 # Observation # 8
-0.104 # -0.309 # Observation # 9
-0.187 # -0.458 # Observation # 10
0.051 # -0.078 # Observation # 11
0.017 # -0.260 # Observation # 12
-0.621 # 0.372 # Observation # 13
0.392 # 0.138 # Observation # 14
0.080 # -0.221 # Observation # 15
;
run;

proc sgrender data=corrplot template=ellipseparm;

dynamic radius1=0.50 radius2=0.71 radius3=0.87 radius4=1;
run;

FRINGEPLOT Statement
Creates a fringe plot on the X axis of an X-Y plot.

Syntax
FRINGEPLOT numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the fringe.
FRINGEHEIGHT=dimension
specifies the height of the fringe lines.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the fringe lines for the data points.

Axes options
XAXIS=X | X2
 FRINGEPLOT Statement 527

specifies whether data are mapped to the primary X (bottom) axis or to

the secondary X2 (top) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
fringe line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of lines for each unique group value of the specified
column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Argument
numeric-column | expression
specifies a column that provides the X coordinates of the data values.

Optional Arguments
DATATRANSPARENCY=number
specifies the degree of the transparency of the fringe.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FRINGEHEIGHT=dimension
specifies the height of the fringe lines.

Default 10 px
528 Chapter 6 / Plot Statements

See “dimension” on page xiii

GROUP=column | discrete-attr-var | expression

creates a distinct set of lines for each unique group value of the specified
column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value is represented in the plot by a different

color. The colors are determined by the ContrastColor attribute of
the GraphData1–GraphDataN and GraphMissing style elements.
The line style and line thickness are determined by the
GraphDataDefault style element.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of colors and
line patterns.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Notes You can override the representations that are used to identify the
groups. For example, you can use the
LINEATTRS=(PATTERN=pattern) option to assign the same line
pattern to all of the plot’s line patterns, letting line color indicate
group values. Likewise, you can use LINEATTRS=(COLOR=color) to
assign the same color to all lines, letting line pattern indicate group
values.

If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tip By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE
 FRINGEPLOT Statement 529

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The X-column label. If a label is not defined, then the X-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the fringe lines for the data points.
530 Chapter 6 / Plot Statements

Defaults For ungrouped data, the GraphDataDefault style element.

For grouped data, the LineStyle and LineThickness attributes of the

GraphDataDefault style element, and the ContrastColor attribute of
the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined role X.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a fringe
line. If this option is used, then it replaces all of the information that is displayed
by default. Roles for columns that do not contribute to the fringe plot can be
specified along with roles that do.

role-list
an ordered, space-separated list of unique FRINGEPLOT and user-defined
roles.

The FRINGEPLOT role is X.

Tip User-defined roles are defined with the ROLENAME= option.

 FRINGEPLOT Statement 531

Example The following example displays data tips for the columns assigned
to the roles X and Y, as well as the column Pct, which is not
assigned to any pre-defined FRINGEPLOT role. The Pct column
should appear first in the data tip:

ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.

Default The column assigned to the X role is automatically included in the

data tip information.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")
532 Chapter 6 / Plot Statements

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Note The display of the fringe lines is always anchored on the X axis (at
the bottom of the plot area), even when the FRINGEPLOT’s X-
column data range is mapped to the X2 axis.

Details
In a FRINGEPLOT, each fringe line represents the location of the corresponding raw
data value on the X axis. All fringe lines are of equal length.

Example: FRINGEPLOT Statement

The following graph was generated by the “Example Program” on page 533:
 HEATMAP Statement 533

Example Program
proc template;
define statgraph fringeplot;
dynamic VAR VARLABEL;
begingraph;
entrytitle "Histogram and Fringeplot";
layout overlay / xaxisopts=(label=VARLABEL)
yaxisopts=(offsetmin=0.03);
fringeplot VAR / datatransparency=0.75
fringeheight=3pct;
histogram VAR;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=fringeplot;

dynamic var="weight" varlabel="Weight (LBS)";
run;

HEATMAP Statement
Creates a plot of color-coded rectangles for the response variable of a pair of X and Y variables after
it bins the data in two dimensions.
 534 Chapter 6 / Plot Statements

Note: This feature applies to SAS 9.4M3 and to later releases.

Syntax
HEATMAP X=numeric-column | expression
Y=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies a numeric column that is used to color the heat map regions.
COLORSTAT= FREQ | PCT | PROPORTION | SUM | MEAN
specifies the statistic to be calculated for the COLORRESPONSE=
column.
DATATRANSPARENCY=number
specifies the degree of the transparency of the outline and fill for each
region.
DISPLAY=(display-options) | STANDARD | ALL
specifies which region features to display.
FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the interior fill area of the regions.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the rectangle outlines.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient that is specified by the ODS
style that is in effect or by the COLORMODEL= option.
WEIGHT=numeric-column
specifies a variable in the input data set that contains values to be used
as a priori weights for the colored-region calculations.
XGAP=number
specifies the amount of horizontal space on either side of each color-
coded region in the heat map.
YGAP=number
specifies the amount of horizontal space on either side of each color-
coded region in the heat map.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot are used for determining
default axis features.
XAXIS=X | X2
 HEATMAP Statement 535

specifies whether data is mapped to the primary X (bottom) axis or to

the secondary X2 (top) axis.
XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks.
XENDLABELS=TRUE | FALSE
specifies whether the X-axis ticks and tick values are placed at the bin
end-points or at the bin mid-points.
YAXIS=Y | Y2
specifies whether data is mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks.
YENDLABELS=TRUE | FALSE
specifies whether the Y-axis ticks and tick values are placed at the bin
end-points or at the bin mid-points.

Binning options
NXBINS=positive-integer
specifies the number of bins to use for the X role.
NYBINS=positive-integer
specifies the number of bins to use for the Y role.
XBINSIZE=positive-number
specifies the size of bins along the X role, in data units.
XBINSTART=number
specifies the data value for the first bin of the X role.
XBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an
X bin.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the XBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.
YBINSIZE=positive-number
specifies the size of bins along the Y role, in data units.
YBINSTART=number
specifies the data value for the first bin of the Y role.
YBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an
Y bin.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the YBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.

Data option
DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric
column.
DISCRETEY=TRUE | FALSE
536 Chapter 6 / Plot Statements

specifies whether the Y axis is discrete when Y= specifies a numeric

column.

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
rectangle.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Plot reference options

NAME="string"
assigns a name to a HEATMAP statement for reference in other template
statements and log messages.

Statistics options
FREQ=numeric-column | expression
specifies a column for the frequency count for each observation in the
input data.

Required Arguments
X=column | expression
specifies the column for the X values.

Requirement If X= specifies a numeric column, then the DISCRETEX= option

must be set correctly for the X-axis type. When it is not set
correctly, the heat map is not drawn. If the X-axis type is discrete,
then you must specify DISCRETEX=TRUE in the HEATMAP
statement. Otherwise, DISCRETEX=FALSE must be in effect.

See DISCRETEX= on page 538

Y=column | expression
specifies the column for the Y values.

Requirement If Y= specifies a numeric column, then the DISCRETEY= option

must be set correctly for the Y-axis type. When it is not set
correctly, the heat map is not drawn. If the Y-axis type is discrete,
then you must specify DISCRETEY=TRUE in the HEATMAP
statement. Otherwise, DISCRETEY=FALSE must be in effect.

See DISCRETEY= on page 538

Optional Arguments
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
 HEATMAP Statement 537

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric column that is used to color the heat map regions.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.
538 Chapter 6 / Plot Statements

Use the COLORSTAT= option to specify the statistic to compute for the
COLORRESPONSE= column.

COLORSTAT= FREQ | PCT | PROPORTION | SUM | MEAN

specifies the statistic to be calculated for the COLORRESPONSE= column.

Default FREQ

Interaction When FREQ, PCT, or PROPORTION is specified, the colors are

controlled by the FREQ= and WEIGHT= option values. When SUM or
MEAN is specified, the colors are controlled by the FREQ=,
WEIGHT=, and COLORRESPONSE= option values.

See COLORRESPONSE=

FREQ=

WEIGHT=

DATATRANSPARENCY=number
specifies the degree of the transparency of the outline and fill for each region.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip Use the FILLATTRS= option to set transparency for just the rectangle
fills. You can combine this option with FILLATTRS= to set one
transparency for the rectangle outlines and a different transparency for
the rectangle fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.

Default FALSE

Requirement If X= specifies a numeric column and the X-axis type is discrete,

then you must specify DISCRETEX=TRUE. Otherwise, the heat
map might not be drawn.

Interactions If X= specifies a character column, then this option is ignored, and

the X axis is considered to be discrete.

When this option is set to TRUE, the XENDLABELS= option is

ignored.

See X= on page 536

DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.

Default FALSE
 HEATMAP Statement 539

Requirement If Y= specifies a numeric column and the Y-axis type is discrete,

then you must specify DISCRETEY=TRUE. Otherwise, the heat
map might not be drawn.

Interactions If Y= specifies a character column, then this option is ignored, and

the Y axis is considered to be discrete.

When this option is set to TRUE, the YENDLABELS= option is

ignored.

See Y= on page 536

DISPLAY=(display-options) | STANDARD | ALL

specifies which region features to display.

display-options
a space-separated list of options, enclosed in parentheses.

Currently, only OUTLINE is supported, which displays outlined, filled

rectangles.

STANDARD
displays the region fill color

ALL
displays the region fill color and outline

Default STANDARD

FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the interior fill area of the regions.

Default The DATATRANSPARENCY= option value

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction Only the TRANSPARENCY= sub option is honored. If a style

element or any other fill sub option is specified, it is ignored by the
HEATMAP statement.

Interaction This option overrides the DATATRANSPARENCY= option only for

the colored region. It does not affect the region outlines.

Tip Use this option when you want the region fill to have a different
degree of transparency than the outline.

FREQ=numeric-column | expression
specifies a column for the frequency count for each observation in the input
data.

Restrictions If the value of FREQ for a given observation is missing or is less

than 1, that observation is not used in the analysis.

If the value is not an integer, only the integer portion is used.

540 Chapter 6 / Plot Statements

NAME="string"
assigns a name to a HEATMAP statement for reference in other template
statements and log messages.

Restriction The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

Tip This option is used mostly with legend statements in order to

coordinate the use of colors and line patterns between the graph
and the legend.

NXBINS=positive-integer
specifies the number of bins to use for the X role. The system determines the
XBINSIZE= and XBINSTART= values if they are not specified. The bins always
span the range of the data.

Default Determined by the system.

Restriction The number of bins is the product of the X and Y bins. Starting
with SAS 9.4M4, the number of bins is limited to 100,000, by
default.

Requirement You must specify a value of 2 or greater. Otherwise, this option is

ignored.

Tip Starting with SAS 9.4M4, the limit on the number of bins is
specified by the ODS GRAPHICS statement NXYBINSMAX=
option, which defaults to 100,000. See NXYBINSMAX= on page
1573.

NYBINS=positive-integer
specifies the number of bins to use for the Y role. The system determines the
YBINSIZE= and YBINSTART= values if they are not specified. The bins always
span the range of the data.

Default Determined by the system.

Restriction The number of bins is the product of the X and Y bins. Starting
with SAS 9.4M4, the number of bins is limited to 100,000, by
default.

Requirement You must specify a value of 2 or greater. Otherwise, this option is

ignored.

Tip Starting with SAS 9.4M4, the limit on the number of bins is
specified by the ODS GRAPHICS statement NXYBINSMAX=
option, which defaults to 100,000. See NXYBINSMAX= on page
1573.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the rectangle outlines.

Default The ContrastColor, LineStyle, and LineThickness attributes of the

GraphOutlines style element.
 HEATMAP Statement 541

Interaction For this option to have any effect, outlines must be enabled by the
DISPLAY= option.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element value

“Line Options” on page 1602 for available line-options values

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot are used for determining default
axis features.

Default FALSE. If no plot in an OVERLAY layout is designated as primary,

the data columns associated with the first plot that could be
primary is considered primary on a per-axis basis.

Restrictions Only one plot in an overlay can be primary on a per-axis basis. If

multiple plots specify PRIMARY=TRUE for the same axis, the last
one specified is considered the primary plot.

This option is ignored if the plot is placed in a GRIDDED or

LATTICE layout block.

Tip This option is needed only when there are two or more plots within
an overlay that contribute to a common axis. If PRIMARY=TRUE for
one of them, then that plot’s data columns are used to determine
the axis features, regardless of where this plot statement occurs
within the OVERLAY layout block.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient that is specified by the ODS style that
is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL= on page 536

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
rectangle. If this option is used, the specified information replaces all the
information that is displayed by default. You can specify roles for columns that
do not contribute to the bar chart along with roles that do.

(role-list)
an ordered, space-separated list of unique HEATMAP roles and user-defined
roles.

HEATMAP roles include X, Y, and the implicit count.

Example To display data tips for the columns assigned to the roles X and Y
as well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
542 Chapter 6 / Plot Statements

TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) in the graph output.

Default The columns assigned to these roles are automatically included in

the data tip information: X and Y

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement that specifies the IMAGEMAP option, and write the
graphs to the ODS HTML destination.

Tip You can control the labels and formats for the TIP variables with
the TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.

(role-format-list)
a space-separated list of role-name = format pairs.

Example To specify a format for the user-defined TIP1 role:

ROLENAME=(TIP1=PCT)
TIP=(TIP1 X Y)
TIPFORMAT=(TIP1=PERCENT7.2)

Default The column format of the variable assigned to the role or BEST6. if
no format is assigned to a numeric column.

Requirement This option provides a way to control the formats of columns that
appear in data tips. Only the roles that appear in the TIP= option
are used.

TIPLABEL=(role-label-list)
specifies display labels for tip columns.

(role-label-list)
a space-separated list of role-name ="string" pairs.

Example To specify a label for the user-defined TIP1 role:

ROLENAME=(TIP1=PCT)
TIP=(TIP1 X Y)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the variable assigned to the
role.

Requirement This option provides a way to control the labels of columns that
appear in data tips. Only the roles that appear in the TIP= option
are used.

WEIGHT=numeric-column
specifies a variable in the input data set that contains values to be used as a
priori weights for the colored-region calculations.
 HEATMAP Statement 543

Requirement The values of the weight variable must be nonnegative.

Note If an observation's weight is zero, negative, or missing, the

observation is deleted from the analysis.

XAXIS=X | X2
specifies whether data is mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks. When this
option is FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the XENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option
is FALSE, the XENDLABELS= option is ignored.

When this option is TRUE, the system attempts to compute nice

values for the bin end points. In that case, the sum of the bin data
ranges might exceed the actual data range.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page xii for other Boolean values that you can use.

XBINSIZE=positive-number
specifies the size of bins along the X role, in data units. The system determines
the NXBINS= and XBINSTART= values if they are not specified. The bins always
span the X data range.

Default Determined by the system.

Restriction Starting with SAS 9.4M1, the number of bins is limited to

approximately 100,000. Starting with SAS 9.4M4, the limit is
specified by the ODS GRAPHICS statement option NXYBINSMAX=,
which defaults to 100,000. If the total number of bins that are
computed from the data, the XBINSTART= value, and the
XBINSIZE= value exceeds the limit, SAS computes a new bin-size
value that yields approximately the maximum number of bins. A
warning of the change is written to the SAS log in that case.

Interaction The XGAP= option is applied after this option.

544 Chapter 6 / Plot Statements

Tip Starting with SAS 9.4M4, you can use the ODS GRAPHICS
statement NXYBINSMAX= option to change the limit on the number
of bins. See NXYBINSMAX= on page 1573.

XBINSTART=number
specifies the data value for the first bin of the X role. The system determines
the NXBINS= and XBINSIZE= values if they are not specified. The bins always
span the X data range.

Default Determined by the system.

Interaction The XVALUES= option specifies how this value is interpreted.

XBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an X bin.
If this option is set to UPPER, then the value is counted as one of the values in
the upper bin. Otherwise, it is counted in the lower bin.

Default UPPER

XENDLABELS=TRUE | FALSE
specifies whether the X-axis ticks and tick values are placed at the bin end-
points or at the bin mid-points.

Default FALSE. The axis ticks and tick values are placed at the bin mid-
points.

Restriction This option is ignored when the X axis is discrete.

Interaction This option is ignored when XBINAXIS=FALSE is in effect.

Tip The axis ticks and tick value placements are independent of the
XVALUES= option.

XGAP=number
specifies the amount of horizontal space on either side of each color-coded
region in the heat map.

Default 0

Interaction If the XBINSIZE= option is specified, the XGAP= value is subtracted

from the specified X bin size.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the XBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.

Default MIDPOINTS

See XBINSTART= on page 544

XENDLABELS= on page 544

 HEATMAP Statement 545

YAXIS=Y | Y2
specifies whether data is mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for x-axis tick marks. When this
option is FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the YENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option
is FALSE, the YENDLABELS= option is ignored.

When this option is TRUE, the system attempts to compute nice

values for the bin end points. In that case, the sum of the bin data
ranges might exceed the actual data range.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page xii for other Boolean values that you can use.

YBINSIZE=positive-number
specifies the size of bins along the Y role, in data units. The system determines
the NYBINS= and YBINSTART= values if they are not specified. The bins always
span the Y data range.

Default Determined by the system.

Restriction Starting with SAS 9.4M1, the number of bins is limited to

approximately 100,000. Starting with SAS 9.4M4, the limit is
specified by the ODS GRAPHICS statement option NXYBINSMAX=,
which defaults to 100,000. If the total number of bins that are
computed from the data, the YBINSTART= value, and the
YBINSIZE= value exceeds the limit, SAS computes a new bin-size
value that yields approximately the maximum number of bins. A
warning of the change is written to the SAS log in that case.

Interaction The YGAP= option is applied after this option.

Tip Starting with SAS 9.4M4, you can use the ODS GRAPHICS
statement NXYBINSMAX= option to change the limit on the number
of bins. See NXYBINSMAX= on page 1573.
546 Chapter 6 / Plot Statements

YBINSTART=number
specifies the data value for the first bin of the Y role. The system determines
the NYBINS= and YBINSIZE= values if they are not specified. The bins always
span the Y data range.

Default Determined by the system.

Interaction The YVALUES= option specifies how this value is interpreted.

YBOUNDARY=UPPER | LOWER
specifies how an input value is counted when it lies on the endpoint of an Y bin.
If this option is set to UPPER, then the value is counted as one of the values in
the upper bin. Otherwise, it is counted in the lower bin.

Default UPPER

YENDLABELS=TRUE | FALSE
specifies whether the Y-axis ticks and tick values are placed at the bin end-
points or at the bin mid-points.

Default FALSE. The axis ticks and tick values are placed at the bin mid-
points.

Restriction This option is ignored when the Y axis is discrete.

Interaction This option is ignored when YBINAXIS=FALSE is in effect.

Tip The axis ticks and tick value placements are independent of the
YVALUES= option.

YGAP=number
specifies the amount of horizontal space on either side of each color-coded
region in the heat map.

Default 0

Interaction If the YBINSIZE= option is specified, the YGAP= value is subtracted

from the specified Y bin size.

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the YBINSTART= value represents the midpoint, lower
endpoint, or upper endpoint of the bin.

Default MIDPOINTS

See YBINSTART= on page 546

YENDLABELS= on page 546

Example: HEATMAP Statement

The following graph was generated by the “Example Program” on page 547.
 HEATMAPPARM Statement 547

Example Program
proc template;
define statgraph heatmap;
begingraph;
entrytitle "Vehicle Mileage By Curb Weight";
layout overlay /
xaxisopts=(label="Curb Weight (LBS)");
heatmap x=weight y=mpg_city / name="heatmap"
nybins=11 ybinstart=10 ybinsize=5
nxbins=11 xbinstart=2000 xbinsize=500;
continuouslegend "heatmap" / title="Count"
location=outside;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=heatmap;

run;

HEATMAPPARM Statement
Creates a two-dimensional plot that represents the values of three variables. Generating an X, Y grid
of rectangles from the values of two independent variables, it colors the rectangles to represent the
values of a third variable, which can be a response variable or a group variable.
 548 Chapter 6 / Plot Statements

Restriction: Starting with SAS 9.4M4, the number of bins is limited to 100,000. If the number of
bins exceeds the maximum, the heat map is not drawn. A warning message is
written to the SAS log in that case.
Requirements: The COLORGROUP= or the COLORRESPONSE= role must be specified.
The data for a parameterized heat map in a single-cell graph or in a cell in a
classification panel must have at least two bins for both the X and Y axes.
Otherwise, the heat map is not drawn.
Note: The data for a parameterized heat map should contain only one observation for
each X and Y value pair.
Tips: Starting with SAS 9.4M4, the limit on the number of bins is specified by the ODS
GRAPHICS statement NXYBINSMAX= option. The default value is 100,000. See
NXYBINSMAX= on page 1573.
Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
HEATMAPPARM X=column | expression
Y=column | expression
COLORGROUP=column | discrete-attr-var | expression </options>;

HEATMAPPARM X=column | expression

Y=column | expression
COLORRESPONSE=numeric-column | range-attr-var | expression </options>;

Summary of Optional Arguments

Appearance options
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
DATATRANSPARENCY=number
specifies the degree of the transparency of the filled rectangles.
DISPLAY=(display-options) | STANDARD | ALL
specifies which region features to display.
FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the area fill in the rectangles.
INCLUDEMISSINGCOLOR=TRUE | FALSE
specifies whether missing values of the color-group variable or of the
color-response variable are included in the plot.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the outlines of the filled rectangles.
 HEATMAPPARM Statement 549

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
XGAP=number
specifies the amount of horizontal space on either side of each filled
rectangle.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower
endpoints, or upper endpoints of the bins.
YGAP=number
specifies the amount of vertical space on either side of each filled
rectangle.
YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input Y values represent midpoints, lower
endpoints, or upper endpoints of the bins.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for X-axis tick marks.
XBOUNDARIES=(numeric-list)
specifies the boundaries of the X-value bins.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for Y-axis tick marks.
YBOUNDARIES=(numeric-list)
specifies the boundaries of the Y-value bins.

Data option
DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric
column.
DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric
column.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
550 Chapter 6 / Plot Statements

specifies the information to display when the cursor is positioned over a

filled rectangle.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
XENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the
endpoints of the bins or midpoints of the bins.
YENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the
endpoints of the bins or midpoints of the bins.

ODS options
URL=string-column
specifies an HTML page to display when a rectangle is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.

Requirement If X= specifies a numeric column, then the DISCRETEX= option

must be set correctly for the X-axis type. When it is not set
correctly, the heat map is not drawn. If the X-axis type is discrete,
then you must specify DISCRETEX=TRUE in the HEATMAP
statement. Otherwise, DISCRETEX=FALSE must be in effect.

See DISCRETEX= on page 552

Y=column | expression
specifies the column for the Y values.

Requirement If Y= specifies a numeric column, then the DISCRETEY= option

must be set correctly for the Y-axis type. When it is not set
correctly, the heat map is not drawn. If the Y-axis type is discrete,
then you must specify DISCRETEY=TRUE in the HEATMAP
statement. Otherwise, DISCRETEY=FALSE must be in effect.

See DISCRETEY= on page 553

COLORGROUP=column | discrete-attr-var | expression

specifies a column or a discrete attribute variable that is used to discretely color
the regions in the heat map.
 HEATMAPPARM Statement 551

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Requirement This argument is required when the response variable is of type

discrete.

Note If you specify a column in a SAS data set, colors are mapped to
group values in data order. If you specify a column in a CAS in-
memory table, colors are mapped to group values in ascending
order of the group column character values or unformatted
numeric values.

See “DISCRETEATTRVAR Statement” on page 1516

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric column or a range attribute variable that is used to color the
regions of the heat map.

range-attr-var
specifies a range attribute variable that is defined in a RANGEATTRVAR
statement.

Restriction A range attribute variable specification must be a direct

reference to the attribute variable. It cannot be set by a dynamic
variable.

See “RANGEATTRVAR Statement” on page 1531

When numeric-column or expression is specified, the colors for each region are
computed by mapping the values to a color ramp that is defined by the
COLORMODEL= option. When attr-variable is specified, the colors defined in
the associated RANGEATTRVAR or DISCRETEATTRVAR statement are used to
color the regions.

Requirement This argument is required when the response variable is of type

interval.

See “RANGEATTRVAR Statement” on page 1531

Optional Arguments
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

552 Chapter 6 / Plot Statements

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the filled rectangles.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

DISCRETEX=TRUE | FALSE
specifies whether the X axis is discrete when X= specifies a numeric column.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default FALSE

Requirements If X= specifies a numeric column and the X-axis type is discrete,

then you must specify DISCRETEX=TRUE. Otherwise, the heat
map might not be drawn.

In SAS programs that were written before SAS 9.4M3, if the

HEATMAPPARM statement is used to plot data on a discrete
category axis, you must add the DISCRETEX=TRUE option to the
 HEATMAPPARM Statement 553

HEATMAPPARM statement. Otherwise, the heat map is not

drawn.

Interactions If X= specifies a character column, then this option is ignored, and

the X axis is considered to be discrete.

When this option is set to TRUE, the XENDLABELS= option is

ignored.

See X= on page 550

DISCRETEY=TRUE | FALSE
specifies whether the Y axis is discrete when Y= specifies a numeric column.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default FALSE

Requirements If Y= specifies a numeric column and the Y-axis type is discrete,

then you must specify DISCRETEY=TRUE. Otherwise, the heat
map might not be drawn.

In SAS programs that were written before SAS 9.4M3, if the

HEATMAPPARM statement is used to plot data on a discrete
response axis, you must add the DISCRETEY=TRUE option to the
HEATMAPPARM statement. Otherwise, the heat map is not
drawn.

Interactions If Y= specifies a character column, then this option is ignored, and

the Y axis is considered to be discrete.

When this option is set to TRUE, the YENDLABELS= option is

ignored.

See Y= on page 550

DISPLAY=(display-options) | STANDARD | ALL

specifies which region features to display.

display-options
a space-separated list of options, enclosed in parentheses.

Currently, only OUTLINE is supported, which displays outlined, filled

rectangles.

STANDARD
displays the region fill color

ALL
displays the region fill color and outline

Default STANDARD

FILLATTRS=(TRANSPARENCY=number)
specifies the transparency of the area fill in the rectangles.
554 Chapter 6 / Plot Statements

Default The DATATRANSPARENCY= option value

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction Only the TRANSPARENCY= suboption is honored. If a style element

or any other fill suboption is specified, then it is ignored by the
HEATMAPPARM statement.

Note The fill colors are determined by the COLORRESPONSE= or

COLORGROUP= or column.

INCLUDEMISSINGCOLOR=TRUE | FALSE
specifies whether missing values of the color-group variable or of the color-
response variable are included in the plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default TRUE

Tip The attributes of the missing color-group or color-response value are

determined by the GraphMissing style element unless a discrete
attribute map is in effect, the MISSING= system option changes the
default missing character, or a user-defined format is applied to the
group value. In those cases, the attributes of the missing group value
are determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See COLORGROUP= on page 550

COLORRESPONSE= on page 551

“boolean ” on page xii for other Boolean values that you can use.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the outlines of the filled rectangles.

Default The ContrastColor, LineStyle, and LineThickness attributes of the

GraphOutlines style element.

Interaction For this option to have any effect, outlines must be enabled by the
DISPLAY= option.
 HEATMAPPARM Statement 555

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, COLORGROUP=, and
COLORRESPONSE=.
556 Chapter 6 / Plot Statements

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a filled
rectangle. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the heat map
can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique HEATMAPPARM and user-
defined roles. HEATMAPPARM roles include X and Y.

Example The following example displays data tips for the columns X and
PCT. The Pct column is not assigned to any pre-defined
HEATMAPPARM role, so it must first be assigned a role:
ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X and Y, and COLORGROUP= or
COLORRESPONSE=.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP variables can be controlled with
the TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)
 HEATMAPPARM Statement 557

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a rectangle is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each rectangle that is to have an active link.

Example http://www.sas.com

Requirement To generate a plot with selectable rectangles, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs. In that case, no
action is taken when the corresponding rectangle is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the rectangle for those X and Y
pairs is selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.
558 Chapter 6 / Plot Statements

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

XBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for X-axis tick marks. When this
option is FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the XENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option
is FALSE, the XENDLABELS= option is ignored.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page xii for other Boolean values that you can use.

XBOUNDARIES=(numeric-list)
specifies the boundaries of the X-value bins. The boundaries are specified as a
space-separated list of values enclosed in parentheses. The keywords MIN and
MAX can be used as one of the values in the list of boundaries. Keywords MIN
and MAX indicate the minimum and maximum data values for the X variable.

Interaction This option is ignored if the X values are not numeric.

Tip This option can be used to specify unequal bins.

Example xboundaries=(MIN 20 200 250 MAX)

XENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the endpoints of
the bins or midpoints of the bins.

Default FALSE. The axis ticks and values labels are drawn at the bin
midpoints.

Restriction This option is ignored when the X axis is discrete.

Interactions If this option is set to FALSE, then the axis ticks and value labels
are drawn at the bin midpoints, regardless of whether the
XVALUES= option identifies the X data as endpoint values or
midpoint values.

This option is ignored when XBINAXIS=FALSE is in effect.

 HEATMAPPARM Statement 559

See “boolean ” on page xii for other Boolean values that you can use.

XGAP=number
specifies the amount of horizontal space on either side of each filled rectangle.

Default 0

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Default MIDPOINTS

See XENDLABELS=

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YBINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for Y-axis tick marks. When this
option is FALSE, a standard axis is used, ignoring bin boundaries and midpoints.

Default TRUE

Interactions When this option is TRUE, the YENDLABELS= option determines

how the axis ticks and value labels are displayed. When this option
is FALSE, the YENDLABELS= option is ignored.

When this option is TRUE, the axis ticks are in predetermined

locations and are not changed when the parent layout specifies axis
suboptions such as TICKVALUELIST=, TICKVALUESEQUENCE=,
and INCLUDERANGES=.

See “boolean ” on page xii for other Boolean values that you can use.

YBOUNDARIES=(numeric-list)
specifies the boundaries of the Y-value bins. The boundaries are specified as a
space-separated list of values enclosed in parentheses. The keywords MIN and
MAX can be used as one of the values in the list of boundaries. Keywords MIN
and MAX indicate the minimum and maximum data values for the Y variable.
Example:
yboundaries=(MIN 20 200 250 MAX)

Interaction This option is ignored if the Y values are not numeric.

Tip This option can be used to specify unequal bins.

560 Chapter 6 / Plot Statements

YENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the endpoints of
the bins or midpoints of the bins.

Default FALSE. The axis ticks and values labels are drawn at the bin
midpoints.

Restriction This option is ignored when the Y axis is discrete.

Interactions If this option is set to FALSE, then the axis ticks and value labels
are drawn at the bin midpoints, regardless of whether the
YVALUES= option identifies the Y data as endpoint values or
midpoint values.

This option is ignored when YBINAXIS=FALSE is in effect.

See “boolean ” on page xii for other Boolean values that you can use.

YGAP=number
specifies the amount of vertical space on either side of each filled rectangle.

Default 0

YVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input Y values represent midpoints, lower endpoints, or
upper endpoints of the bins.

Default MIDPOINTS

See YENDLABELS=

Details
A heat map is useful for visualizing the magnitude of a response variable in relation
to two independent variables. For example, in molecular biology, heat maps can be
used to track the expression of genes across multiple sample studies. In the
HEATMAPPARM statement, you specify the independent variables in the X and Y
arguments. For a response variable that has discrete values, use the COLORGROUP
argument, or for a response variable that has interval values, use the
COLORRESPONSE argument. For interval response variables, you can use the
COLORMODEL= option to define the color ramp that is used to fill the rectangles.

Example: HEATMAPPARM Statement

The following graph was generated by the “Example Program” on page 561:
 HIGHLOWPLOT Statement 561

Example Program
proc template;
define statgraph heatmapparm;
begingraph;
layout overlay;
heatmapparm x=height y=weight colorresponse=count /
name="heatmapparm" xbinaxis=false ybinaxis=false;
continuouslegend "heatmapparm" / location=outside
valign=bottom;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.gridded template=heatmapparm;

run;

HIGHLOWPLOT Statement
Creates a display of floating vertical or horizontal lines or bars that connect the minimum and
maximum response values for each value of a categorical variable.

Requirements: Either the X= or Y= argument must be specified, but you cannot specify both on the
same HIGHLOWPLOT statement.
The HIGH= and LOW= arguments are required.
 562 Chapter 6 / Plot Statements

Note: Specifying the X= option creates a vertical high-low chart, which would typically be
used in the financial industry to plot stock values over time. Specifying the Y=
option creates a horizontal high-low chart, which would typically be used in the
Health and Life Sciences industry to display over time the duration of adverse
events or of adverse reactions to medication.
Tips: For charts that have a large number of bars that are very close together, slight
variations in spacing that normally occur due to integer rounding can become more
obvious. Subpixel rendering provides more precise bar spacing in that case. In SAS
9.4M2 and in earlier releases, specify SUBPIXEL=ON in the BEGINGRAPH
statement to enable subpixel rendering. See SUBPIXEL= on page 39. Starting with
SAS 9.4M3, subpixel rendering is enabled by default.
To disable subpixel rendering in SAS 9.4M3 and in later releases, specify
SUBPIXEL=OFF in the BEGINGRAPH statement or in an ODS GRAPHICS
statement. For information about the BEGINGRAPH statement SUBPIXEL= option,
see SUBPIXEL= on page 39. For information about the ODS GRAPHICS statement
SUBPIXEL= option, see “ODS GRAPHICS Statement” on page 1561.

Syntax
HIGHLOWPLOT X=column | expression
LOW=numeric-column | expression
HIGH=numeric-column | expression < /options >;

HIGHLOWPLOT Y=column | expression

LOW=numeric-column | expression
HIGH=numeric-column | expression < /options >;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
CLIPCAP=TRUE | FALSE
specifies whether a special clip cap is displayed to indicate where
clipping occurred.
CLIPCAPSHAPE=DEFAULT | BARBEDARROW | CLIPPEDARROW |
CLOSEDARROW | FILLEDARROW | OPENARROW | SERIF
specifies the shape of the arrowhead on the clipped end of a line or bar
when CLIPCAP=TRUE.
CLOSE=numeric-column | expression
specifies a column or expressions whose values are used to display a
closing-value indicator.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction
of the midpoint spacing. On an interval axis, specifies the width of the
 HIGHLOWPLOT Statement 563

group clusters as a fraction of the minimum interval between adjacent

data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the bar or
line colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars or lines of a high-low
chart.
DATATRANSPARENCY=number
specifies the degree of the transparency of the colored regions, and the
high-end and low-end labels.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
ENDCAPDISPLAYPOLICY=AUTO | ALWAYS
specifies the policy for displaying end caps when end caps are present.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
HIGHCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |
OPENARROW | CLOSEDARROW
specifies the type of cap used at the high end of the bar or line.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) or
bar attributes (fill and outline) to one of the GraphData1–GraphDataN
style elements.
INTERVALBARWIDTH=dimension
specifies the width of the floating bars.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the floating plot lines.
LOWCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |
OPENARROW | CLOSEDARROW
specifies the type of cap used at the low end of the bar or line.
OPEN=numeric-column | expression
specifies a column or expressions whose values are used to display an
opening-value indicator.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the outlines of the filled regions.
TYPE=LINE | BAR
specifies whether data values should be represented by bars or lines.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
564 Chapter 6 / Plot Statements

specifies whether data are mapped to the primary X (bottom) axis or to

the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
bar or line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
HIGHLABEL=column | expression
specifies the label to display at the high end of the bar or line.
LABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the labels for the bars or lines.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
LOWLABEL=column | expression
specifies the label to appear at the low end of each floating bar or line.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all bars or lines from the category
midpoints when graphing multiple response variables side by side on a
common axis.
GROUP=column | discrete-attr-var | expression
creates a distinct set of floating bars or lines for each unique group value
in the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped bars or lines are overlaid or clustered around
the category midpoints.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
 HIGHLOWPLOT Statement 565

Required Arguments
Either the X= or Y= argument must be specified, but you cannot specify both on the same
HIGHLOWPLOT statement. The HIGH= and LOW= arguments are required.
X=column | expression
specifies a column or expression representing the values in a vertical high-low
chart. The values can be character or numeric.

Y=column | expression
specifies a column or expression representing the values in a horizontal high-
low chart. The values can be character or numeric.

LOW=numeric-column | expression
specifies a column or expression representing the values of the lower end of the
floating bar or line.

HIGH=numeric-column | expression
specifies a column or expression representing the values of the higher end of
the floating bar or line.

Optional Arguments
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Restriction This option has an effect only when TYPE= BAR.

Interaction Prior to SAS 9.4M3, this option is ignored for an interval high-low
plot, and the bar width is controlled by the INTERVALBARWIDTH=
option. Starting with SAS 9.4M3, this option is honored for an
interval high-low plot but it can be overridden by the
INTERVALBARWIDTH= option.

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number

of bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

See DISCRETEOFFSET= option for examples of using this option.

CLIPCAP=TRUE | FALSE
specifies whether a special clip cap is displayed to indicate where clipping
occurred. When the VIEWMIN= and VIEWMAX= axis options are specified for
an axis and a data value exceeds the specified axis range, the bar for that value
is clipped. In that case, when CLIPCAP=TRUE, a special clip cap is displayed at
the clipped end of the bar. If the bar already has a high or low cap, then it is
replaced by the clip cap.
566 Chapter 6 / Plot Statements

For vertical bars, the clip cap is added to the end of the bar that is clipped by the
Y-axis range. The default clip cap is a vertical clipped arrowhead that points
toward the clip edge ( or ). For horizontal bars, the cap is added to the end of
the bar that is clipped by the X-axis range. The default clip cap is a horizontal
clipped arrowhead that points toward the clip edge ( or ). If an entire bar is
clipped, then a clip cap is displayed at the high or low side where the bar was
clipped.

The following figure shows a side-by-side comparison of a vertical line high-low

chart with no clipping and with clipping when CLIPCAP=TRUE.

The first graph shows the default case in which no clipping occurs. The second
graph shows the case in which VIEWMIN= and VIEWMAX= specify a range that
causes clipping on the Y axis. In the second graph, the clip caps indicate clipping
for each category line as follows:

A the high end of the line is clipped.

B the low end of the line is clipped.
C both ends of the line are clipped.
D the entire line is clipped at the low end.
E the entire line is clipped at the high end.

Default FALSE

Interaction Clip caps appear only when CLIPCAP=TRUE and the data values
exceed the axis range that is specified by the VIEWMIN= and
VIEWMAX= options. When the VIEWMIN= and VIEWMAX= options
are not specified, the axis range is adjusted to accommodate the
data values and clipping does not occur.

Note If the HIGHLABEL= and LOWLABEL= options are in effect and the
bar is clipped, then the label value at the clipped end is drawn at the
tip of the clip cap. If the entire bar is clipped, then the labels are not
shown.
 HIGHLOWPLOT Statement 567

Tip Use the CLIPCAPSHAPE= option to specify a different clip-cap

arrowhead.

See “boolean ” on page xii for other Boolean values that you can use.

CLIPCAPSHAPE=DEFAULT | BARBEDARROW | CLIPPEDARROW |

CLOSEDARROW | FILLEDARROW | OPENARROW | SERIF
specifies the shape of the arrowhead on the clipped end of a line or bar when
CLIPCAP=TRUE. The following figure shows each of the clip-cap arrowhead
shapes for vertical lines and bars that are clipped at both ends.

In SAS 9.4M1 and earlier releases, this option specifies the arrowhead shape
only for the clipped end of clipped lines (TYPE=LINE). For bars (TYPE=BAR),
this option is ignored, and CLIPPEDARROW is always used as the arrowhead
shape for the clipped end of clipped bars.

Starting with SAS 9.4M2, this option specifies the arrowhead shape for the
clipped end of clipped lines or clipped bars. For bars, you can specify the
CLIPPEDARROW shape for filled and unfilled bars. When any value other than
CLIPPEDARROW or DEFAULT is specified for bars, the FILLEDARROW shape is
used for filled bars, and the CLOSEDARROW shape is used for unfilled bars.

Default DEFAULT, which is the same as CLIPPEDARROW

Interaction This option is ignored when CLIPCAP=FALSE.

CLOSE=numeric-column | expression
specifies a column or expressions whose values are used to display a closing-
value indicator. This option is typically used when TYPE=LINE, but it can be
used when TYPE=BAR. For vertical high-low charts, the value is represented by
a short horizontal line extending from the side that displays the higher X values.
For horizontal high-low charts, the value is represented by a short vertical line
extending from the side that displays the higher Y values.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters
as a fraction of the minimum interval between adjacent data values.
568 Chapter 6 / Plot Statements

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interaction For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Defaults For lines or for outline-only bars, the ThreeColorAltRamp style

element.

For bars with fill, the ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the bar or line
colors.

Note: This feature applies to SAS 9.4M3 and to later releases.

 HIGHLOWPLOT Statement 569

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outlines are displayed, this option overrides

suboption COLOR= in the OUTLINEATTRS= option and varies the
outline color according to the color gradient or the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the outline color gradient.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars or lines of a high-low chart.
The following figure shows a high-low chart that contains a filled bar and a line
with each of the skins applied.
570 Chapter 6 / Plot Statements

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

When TYPE=BAR, the DISPLAY= option in effect must include FILL

in order for the DATASKIN= option to have any effect.

For filled bars, the skin appearance is based on the FILLATTRS=

option color specification.

DATATRANSPARENCY=number
specifies the degree of the transparency of the colored regions, and the high-
end and low-end labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS option can be used to set transparency for just the
colored regions. You can combine this option with FILLATTRS= to set
one transparency for the outlines but a different transparency for the
region fills . Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISCRETEOFFSET=number
specifies an amount to offset all bars or lines from the category midpoints when
graphing multiple response variables side by side on a common axis.
 HIGHLOWPLOT Statement 571

Default 0 (no offset, all bars or lines are centered on the category
midpoints)

Range –0.5 to +0.5, where 0.5 represents half the distance between
category ticks. Normally, a positive offset is to the right for a vertical
plot and up for a horizontal plot. (If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.)

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

See “About the DISCRETEOFFSET= Option” on page 585

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the bar fill color

OUTLINE
displays the bar outline

FILLPATTERN
displays the bar fill pattern

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the bars.

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Defaults In SAS 9.4M4 and in earlier releases, STANDARD

Starting in SAS 9.4M5, the GraphHighLow:DisplayOpts style

reference, if defined in the active ODS style. Otherwise,
STANDARD.
572 Chapter 6 / Plot Statements

Requirement You must also specify TYPE=BAR. Otherwise, this option is

ignored.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the bar outline and fill color. Starting with SAS
9.4M5, use the FILLPATTERNATTRS= on page 573 option to
control the appearance of the bar fill pattern.

ENDCAPDISPLAYPOLICY=AUTO | ALWAYS
specifies the policy for displaying end caps when end caps are present.

AUTO
draws the end caps only for the elements with a low-to-high range that is
large enough to accommodate the end caps.

Draws just the bar or line for the rest.

ALWAYS
always draws the end caps.

Note When the low-to-high range is not large enough for the end caps, the
end caps might overlap. The data ranges that are smaller than the
end-cap size might not be resolvable beyond the rendered size of the
end caps.

Default AUTO

Interactions This option is honored only when the option TYPE=LINE is in effect.

This option is ignored if the LOWCAP= or HIGHCAP= option is not

set or is effectively set to NONE.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to take effect, TYPE=BAR must also be specified,
and the DISPLAY= option must allow the fill to be displayed.

When COLORRESPONSE= is in effect, the FILLATTRS= suboption

COLOR= is ignored, and the bar fill colors vary according to the
gradient.

Tip The DATATRANSPARENCY= option sets the transparency for the

colored regions and the outlines around them. You can combine this
option with DATATRANSPARENCY= to set one transparency for
the outlines but a different transparency for the colored regions.
Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
 HIGHLOWPLOT Statement 573

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interactions For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Starting with SAS 9.4M5, when COLORRESPONSE= is in effect and

the DISPLAY= option enables FILLPATTERN display, the
FILLPATTERNATTRS= suboption COLOR= is ignored, and the fill-
pattern colors vary according to the gradient.

Tip In order to change the fill pattern for each group value in ODS
styles such as HTMLBlue that use color-priority attribute rotation,
specify ATTRPRIORITY=NONE in an ODS GRAPHICS statement or
in the BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-

options.

GROUP=column | discrete-attr-var | expression

creates a distinct set of floating bars or lines for each unique group value in the
specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

When TYPE= on page 583LINE (default), combinations of line color and line
style are used to create a distinct line for each group value. The ContrastColor
574 Chapter 6 / Plot Statements

and LineStyle attributes of a GraphData1–GraphDataN style element or the

GraphMissing style element control the color and line style of each group line.

When TYPE=BAR, a distinct bar is created for each group value by varying the
visual attributes of the bar display features. The display features are controlled
by the current ODS style or by the DISPLAY= option. The default group
appearance for each display feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.

Outline n if the outline is enabled with fill

color, fill pattern, or both,
ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is the only display
feature enabled, ContrastColor and
LineStyle attributes of a
GraphData1–GraphDataN style
element or the GraphMissing style
element

1 Fill color and the bar outline are displayed by default.

2 Fill patterns are valid for high-low plots starting with SAS 9.4M5.

Interactions If the X or Y role is discrete, then the bars or lines representing the
group can be drawn in a cluster or overlaid, based on the setting for
the GROUPDISPLAY= option. However, if the X or Y role is interval,
then the lines representing the group are overlaid and the
GROUPDISPLAY= option is ignored.

By default, the group values are mapped in the order of the data.
Use the GROUPORDER= option to control the sorting order of the
grouped bar segments. Use the INDEX= option to alter the default
sequence of colors and line patterns.

The INCLUDEMISSINGGROUP= option determines whether

missing group values are considered a distinct group value.

When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.
 HIGHLOWPLOT Statement 575

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden. For example, if each distinct group value is represented
by a different line pattern, you can use the
LINEATTRS=(PATTERN=pattern) specification to assign the same
line pattern to all the plot lines. In that case, the line color denotes
group values. Likewise, you can use LINEATTRS=(COLOR=color) to
assign the same color to all lines, letting line pattern denote group
values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped bars or lines are overlaid or clustered around the
category midpoints.

OVERLAY
centers the bars or lines for matching category values on the midpoints.

The bars or lines in each set of group values are superimposed on each other.

CLUSTER
clusters the bars or lines for matching category values around the midpoints.

Each cluster of group values is centered at the midpoint for the category.

The following example shows the effect of clustering the lines in a stock report
when the category values are grouped into a single response variable. Note that
if your category values are not grouped in the same column but are stored in
separate columns, then you can get this same effect by using the
DISCRETEOFFSET= option.
layout overlay /
yaxisopts=(label="Stock Value");
highlowplot x=month high=high low=low /
close=close open=open
legendlabel="Stock" name="cluster"
group=stock groupdisplay=cluster
lineattrs=(pattern=solid);
discretelegend "cluster" / title="Stock"
location=inside halign=right valign=top;
endlayout;
576 Chapter 6 / Plot Statements

Default OVERLAY

Interaction For this option to take effect, the GROUP= option must also be
specified.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.
 HIGHLOWPLOT Statement 577

This option is ignored when GROUPDISPLAY=OVERLAY.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

HIGHCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |

OPENARROW | CLOSEDARROW
specifies the type of cap used at the high end of the bar or line. All of the
keywords can be specified for any high-low chart. The effect of each keyword
depends on the setting for the TYPE= and DISPLAY= options as follows:
n When TYPE=BAR and DISPLAY= includes FILL, FILLEDARROW is used for
all settings other than NONE.
n When TYPE=BAR and DISPLAY= does not include FILL, CLOSEDARROW is
used for all settings other than NONE.
n When TYPE=LINE and CLOSEDARROW is specified, FILLEDARROW is used
instead.

The following figure shows the effect of each cap value on horizontal lines,
filled bars, and unfilled bars.
578 Chapter 6 / Plot Statements

Figure 6.1 Horizontal High and Low Cap Shapes for Lines, Filled Bars, and
Unfilled Bars

Default NONE

Interaction When TYPE=BAR, the caps are drawn to fit within the bar width.
The width of the bar itself is reduced.

Note If the length of the high-low element is smaller than the cap, then
the cap is not drawn.

HIGHLABEL=column | expression
specifies the label to display at the high end of the bar or line.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) or bar
attributes (fill and outline) to one of the GraphData1–GraphDataN style
elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.
 HIGHLOWPLOT Statement 579

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped
according to the setting of the GROUPORDER= option.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

INTERVALBARWIDTH=dimension
specifies the width of the floating bars.

Defaults Prior to SAS 9.4M3, 85% of the smallest interval between any two
bars for the given plot

Starting with SAS 9.4M3, the BARWIDTH= option setting

Restriction For this option to take effect, TYPE=BAR must be set, and the
independent variable must be of type interval.

Interaction Prior to SAS 9.4M3, this option controls the bar width for a high-low
plot. Starting with SAS 9.4M3, this option overrides the
BARWIDTH= option.

Tip The bar width that you specify with this option is honored even if
the bars overlap. If the bars overlap, reduce the interval bar width or
use the BARWIDTH= option instead.

See “dimension” on page xiii

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the labels for the bars or lines.

Defaults In SAS 9.4M2 and earlier releases, the GraphDataText style element.

Starting with SAS 9.4M2, the GraphDataText style element is the

default for non-grouped data. For grouped data, the data label color is
determined by the ContrastColor attribute of the GraphData1–
GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax for
using a style-element.
580 Chapter 6 / Plot Statements

“Text Options” on page 1606 for available text-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the floating plot lines.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

Interaction For this option to have any effect, TYPE= LINE must also be
specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

LOWCAP=column | NONE | SERIF | BARBEDARROW | FILLEDARROW |

OPENARROW | CLOSEDARROW
specifies the type of cap used at the low end of the bar or line. All of the
keywords can be specified for any high-low chart. The effect of each keyword
depends on the setting for the TYPE= and DISPLAY= options as follows:
n When TYPE=BAR and DISPLAY= includes FILL, FILLEDARROW is used for
all settings other than NONE.
n When TYPE=BAR and DISPLAY= does not include FILL, CLOSEDARROW is
used for all settings other than NONE.
n When TYPE=LINE and CLOSEDARROW is specified, FILLEDARROW is used
instead.

Figure 6.4 on page 578 shows the effect of each cap value on horizontal lines,
filled bars, and unfilled bars.

Default NONE

Interaction When TYPE=BAR, the caps are drawn to fit within the bar width.
The width of the bar itself is reduced.

Note If the length of the high-low element is smaller than the cap, then
the cap is not drawn.

LOWLABEL=column | expression
specifies the label to appear at the low end of each floating bar or line.
 HIGHLOWPLOT Statement 581

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

OPEN=numeric-column | expression
specifies a column or expressions whose values are used to display an opening-
value indicator. This option is typically used when TYPE=LINE, but it can be
used when TYPE=BAR. For vertical high-low charts, the value is represented by
a short horizontal line extending from the side that displays the lower X values.
For horizontal high-low charts, the value is represented by a short vertical line
extending from the side that displays the lower Y values.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the outlines of the filled regions.

Defaults For non-grouped data, the ContrastColor, LineThickness, and

LineStyle attributes of the GraphOutlines style element.

For grouped data when FILL is displayed, the LineThickness and

LineStyle attributes of the GraphOutlines style element, and the
ContrastColor attribute of the GraphData1–GraphDataN style
elements.

For grouped data when FILL is not displayed, the LineThickness

attribute of the GraphOutlines style element, and the
ContrastColor and LineStyle attributes of the GraphData1–
GraphDataN style elements.

Interactions For this option to have any effect, TYPE=BAR must be specified,
and outlines must be enabled by the ODS style or the DISPLAY=
option.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the bar outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE
582 Chapter 6 / Plot Statements

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, LOW, HIGH, OPEN, and CLOSE.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar or
line. If this option is used, then the information specified replaces all of the
information that is displayed by default. You can specify roles for columns that
do not contribute to the bar chart along with roles that do.

(role-list)
an ordered, space-separated list of unique HIGHLOWPLOT and user-defined
roles.

HIGHLOWPLOT roles include X, Y, LOW, HIGH, OPEN, CLOSE, GROUP, and

COLORRESPONSE.

Note The COLORRESPONSE role is valid starting with SAS 9.4M3.

Example The following example displays data tips for the columns X and
Pct. The Pct column is not assigned to any pre-defined
HIGHLOWPLOT role, so it must first be assigned a role.
ROLENAME=(TIP1=PCT)
TIP=(X TIP1)

NONE
suppresses data tips from the plot.
 HIGHLOWPLOT Statement 583

Default The columns assigned to these roles are automatically included in

the data tip information: X or Y and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TYPE=LINE | BAR
specifies whether data values should be represented by bars or lines.
584 Chapter 6 / Plot Statements

BAR
uses fill and outline attributes.

LINE
uses line attributes.

Default LINE

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interactions When the Y= argument is specified, the HIGH= and LOW=

arguments are mapped to the axis that is specified on this option (X
or X2).

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interactions When the X= argument is specified, the HIGH= and LOW=

arguments are mapped to the axis that is specified on this option (Y
or Y2).

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

Statement Description
A high-low chart specifies that floating vertical or horizontal lines or bars connect
the minimum and maximum response values for each value of a categorical
variable. The data should have at least two response values for every category
value. Otherwise, the single value is displayed without the connecting line or bar. In
the statement syntax, the X and Y arguments can specify a column with character
or numeric values. The LOW and HIGH arguments must specify a numeric column.

The HIGHLOWPLOT statement can be used to create a plot showing the high and
low response values for observations along a time axis. The independent variable
can be interval or discrete. The HIGHLOWPLOT has two common uses:
n One typical use of a high-low chart is in the financial industry to plot stock
values by day. The OPEN= and CLOSE= options are typically used in the plot,
 HIGHLOWPLOT Statement 585

and the TYPE= option is set to LINE (the default), as illustrated in “Example 1:
Vertical High-low Chart” on page 586 .
n Another typical use of a high-low chart is in the Health and Life Sciences
industry to display over time the duration of adverse events or of adverse
reactions to medication. The HIGHCAP= and LOWCAP= options are typically
used in the plot, and the TYPE= option is set to BAR, as illustrated in “Example
2: Horizontal High-low Chart” on page 587 .

For both vertical and horizontal high-low charts, the data values can be grouped
using the GROUP= option. For grouped data, the data values are not summarized.
Each observation is plotted independently, and the line or bar segment for each
category value can use different display characteristics.

About the DISCRETEOFFSET= Option

This feature is useful for graphing multiple response variables side by side on a
common axis. By default within an overlay-type layout, if multiple HIGHLOWPLOT
statements are used with different response variables, then the bars or lines for
matching category values are centered on the midpoints and superimposed on each
other. In those cases, you can make it easier to distinguish among superimposed
bars or lines by assigning a different width for them in each HIGHLOWPLOT
statement in the overlay.

If you prefer to avoid superimposed bars or lines, then you can assign a different
offset to each HIGHLOWPLOT statement. If desired, you can adjust the width of
the bar or line in conjunction with DISCRETEOFFSET= to create narrower bars that
require less width within the plot area.

The following example defines offsets for the lines in a stock report:
layout overlay / cycleattrs=true
yaxisopts=(label="Stock Value");
highlowplot x=month high=a_high low=a_low /
close=a_close open=a_open
legendlabel="A" name="ahighlow" lineattrs=(pattern=solid)
discreteoffset=-0.2 ;
highlowplot x=month high=b_high low=b_low /
close=b_close open=b_open
legendlabel="B" name="bhighlow" lineattrs=(pattern=solid)
discreteoffset=0 ;
highlowplot x=month high=c_high low=c_low /
close=c_close open=c_open
legendlabel="C" name="chighlow" lineattrs=(pattern=solid)
discreteoffset=+0.2 ;
discretelegend "ahighlow" "bhighlow" "chighlow" / title="Stock"
location=inside halign=right valign=top;
endlayout;
586 Chapter 6 / Plot Statements

Examples

Example 1: Vertical High-low Chart

The following vertical high-low chart was generated by “Example Program” on page
587:
 HIGHLOWPLOT Statement 587

Example Program
proc template;
define statgraph highlow;
begingraph;
layout overlay /
yaxisopts=(griddisplay=on label="Stock Value");
highlowplot x=date high=high low=low /
open=open close=close;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=highlow;

where date >= '01JAN03'd and stock="IBM";
run;

Example 2: Horizontal High-low Chart

The following horizontal high-low chart was generated by “Example Program” on
page 587 :

Example Program
data highlowbar;
length cap $ 12;
input drug $ 1-10 low high cap $;
datalines;
 588 Chapter 6 / Plot Statements

Drug A 10 20 NONE
Drug A 30 60 FILLEDARROW
Drug B 20 35 NONE
Drug B 50 75 FILLEDARROW
Drug C 30 90 FILLEDARROW
;
proc template;
define statgraph highlowbar;
begingraph;
entrytitle 'Medications Plot';
layout overlay /
xaxisopts=(
griddisplay=on
linearopts=(viewmin=0 viewmax=100)
display=(line ticks tickvalues))
yaxisopts=(
griddisplay=on
display=(line ticks tickvalues));
highlowplot y=drug high=high low=low /
group=drug outlineattrs=(pattern=solid)
datatransparency=0.4
type=bar barwidth=0.4
highcap=cap lowcap=cap;
endlayout;
endgraph;
end;

proc sgrender data=highlowbar template=highlowbar;

run;

HISTOGRAM Statement
Creates a univariate histogram computed from input data.

Requirement: If you include a histogram and a density plot in your graph and both plots specify
the same data column, you must either specify SCALE=DENSITY in the
HISTOGRAM statement or place the HISTOGRAM statement before the
DENSITYPLOT statement in your template code. Otherwise, the density plot
values might not scale properly to the histogram bins on the dependent axis. This
requirement applies even when the histogram and density plots are in different
cells in a multicell graph.
Note: When a histogram is placed on a log axis, binning is done on a linear scale, but the
bins are drawn on the log scale. As a result, the bins might have different widths
along the log axis.

Syntax
HISTOGRAM numeric-column | expression </options>;
 HISTOGRAM Statement 589

Summary of Optional Arguments

Appearance options
BINSTART=number
specifies the X coordinate of the first bin.
BINWIDTH=positive-number
specifies the bin width.
BOUNDARY=UPPER | LOWER
specifies how a boundary is counted when it lies on the endpoint of a bin.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fills and bar outlines.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT
specifies the bar fill type.
NBINS=positive-integer
specifies the number of bins.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the bar outlines.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower
endpoints, or upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
SCALE=PERCENT | COUNT | PROPORTION | DENSITY
specifies the scale for the Y axis.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
590 Chapter 6 / Plot Statements

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the bar labels.
DATALABELTYPE=NONE | AUTO | COUNT | DENSITY | PERCENT |
PROPORTION
specifies the statistic to display at the end of each bar.
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate bar segment or bar for each unique group value in the
specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.
WEIGHT=numeric-column | expression
specifies a column that contains a bin-width calculation a priori weight
for each observation of the input data object.

Required Argument
numeric-column | expression
specifies a column that contains numeric values, or a dynamic variable that
refers to such a column.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.
 HISTOGRAM Statement 591

TRUE
specifies that the ENDLABELS= option determines how the axis ticks and
value labels are displayed.

FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are
set by the ENDLABELS= option are ignored.

Default TRUE

Interactions This option is ignored when this histogram is not the primary plot.
For more information about primary plots, see “When Plots Share
Data and a Common Axis” on page 1046.

BINAXIS=TRUE is ignored when the histogram is placed on a log

axis.

When this option is set to TRUE, some X-axis options that are set
on the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “ENDLABELS=TRUE | FALSE” on page 594

“boolean ” on page xii for other Boolean values that you can use.

BINSTART=number
specifies the X coordinate of the first bin. This option can be used with the
BINWIDTH= and NBINS= options to specify the bins.

Default Determined by the system.

Interaction This option is ignored if XVALUES=LEFTPOINT and the specified

BINSTART value is greater than the minimum of your data.

Note If the BINSTART= value is less than the minimum data value, then
the lower end of the histogram might be padded with zero-height
bins in order to accommodate the BINSTART= value.

See XVALUES=

BINWIDTH=positive-number
specifies the bin width. This option can be used with the BINSTART= and
NBINS= options to specify the bins.

Default Determined by the system.

Restriction Starting with SAS 9.4M1, the number of bins is limited to

approximately 10,000. If the number of bins computed from the
data, the BINSTART= value, and the BINWIDTH= value exceeds the
limit, SAS computes a new bin-width value that yields
approximately the maximum number of bins. A warning of the
change is written to the SAS log in that case.

Interaction This option is ignored when the NBINS= option is also specified
(with or without the BINSTART= option) and the resulting data
592 Chapter 6 / Plot Statements

range does not completely span the unbinned input data range. In
that case, the NBINS= option is honored and the default bin width is
used instead.

Note When BINSTART=, NBINS=, or both are specified with this option
and the resulting data range is greater than the unbinned input data
range, the histogram might be padded with zero-height bins in order
to accommodate the BINSTART=, NBINS=, and BINWIDTH= option
values.

Tip Starting with SAS 9.4M4, the limit on the number of bins is specified
by the ODS GRAPHICS statement NBINSMAX= option. The default
value is 10,000. See NBINSMAX= on page 1573.

BOUNDARY=UPPER | LOWER
specifies how a boundary is counted when it lies on the endpoint of a bin. If this
option is set to UPPER, then the value is counted as one of the values in the
upper bin (the bin to the right). Otherwise, it is counted in the lower bin.

Default UPPER

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the bar labels.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphDataText style element

Interactions The DATALABEL= option must be specified for this option to have
any effect.

When text options are specified, any font properties that are not
specified (color, family, size, weight, and style) are derived from the
GraphDataText style element.

See “General Syntax for Attribute Options” on page 1599

“Text Options” on page 1606

DATALABELTYPE=NONE | AUTO | COUNT | DENSITY | PERCENT | PROPORTION

specifies the statistic to display at the end of each bar.

NONE
suppresses the data labels.

AUTO
uses the SCALE= option value. By default, SCALE=PERCENT.

See “SCALE=PERCENT | COUNT | PROPORTION | DENSITY” on page 601

 HISTOGRAM Statement 593

COUNT
DENSITY
PERCENT
PROPORTION
specifies that the count, density, percentage, or proportion statistic is to be
displayed at the end of each bar.

Default NONE

Interaction When DATALABELTYPE=AUTO, the SCALE= option determines the

statistic that is displayed at the end of each bar.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
histogram bars with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.
594 Chapter 6 / Plot Statements

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fills and bar outlines.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
bar fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines but a different transparency for the
bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the bar fill color

FILLPATTERN
displays the bar fill pattern

Note This option is valid starting in SAS 9.4M1.

OUTLINE
displays the bar outline

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Default The GraphHistogram:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= , FILLATTRS= , and FILLPATTERNATTRS=

options to control the appearance of the bars.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the
parent layout. For more information, see the PRIMARY= option.
 HISTOGRAM Statement 595

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.

If the TICKS= suboption is specified in the XAXISOPTS= option,

then this option is ignored.

See “boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Defaults In SAS 9.4M1 and earlier releases, the Color attribute of the
GraphDataDefault style element

Starting with SAS 9.4M2, the Color attribute of the

GraphDataDefault style element for non-grouped data or the
GraphData1–GraphDataN style elements for grouped data.

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

When a discrete attribute map is in effect for the histogram, the fill
colors that are specified by the attribute map take precedence over
COLOR. If the attribute map provides no color for a value, COLOR
prevails for that value.

Tip The DATATRANSPARENCY= option sets the transparency for bar

fills and bar outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphWall style element Color attribute.

Requirement FILLTYPE=COLORGRADIENT must be in effect. Otherwise, this

option is ignored.

See “color ” on page xii

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
596 Chapter 6 / Plot Statements

Note: This option is valid starting in SAS 9.4M1.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Restriction In SAS 9.4M4 and in earlier releases, the only styles that are
delivered by SAS that support fill patterns are JOURNAL2,
JOURNAL3, and MONOCHROMEPRINTER. If style-element is
specified and the active ODS style does not support fill patterns,
this option is ignored.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT

specifies the bar fill type.

Note: This feature applies to SAS 9.4M2 and to later releases.

SOLID
fills each bar with the color that is assigned to that bar.

ALPHAGRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This feature applies to SAS 9.4M5 and to later releases.

Tips The fill color is determined by a style element or by the FILLATTRS=

option COLOR= suboption. The initial transparency is determined by
the DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete attribute

map to set the initial transparency in the gradients for specific values.
 HISTOGRAM Statement 597

COLORGRADIENT
fills each bar with a color gradient that starts with the fill color that is
assigned to that bar and ends with the color that is specified in the
FILLENDCOLOR= option.

Note: This feature applies to SAS 9.4M5 and to later releases.

Alias GRADIENT

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tip Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

See “FILLENDCOLOR=color” on page 595

GRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This option applies to SAS 9.4M2 through SAS 9.4M4. Starting with
SAS 9.4M5, ALPHAGRADIENT replaces GRADIENT, and GRADIENT is
changed to an alias of COLORGRADIENT.

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tips The initial fill color is determined by a style element or by the

FILLATTRS= option COLOR= suboption. The initial transparency
is determined by the DATATRANSPARENCY= option or by the
FILLATTRS= option TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

Note The output for SAS programs written prior to SAS 9.4M5 that
specify FILLTYPE=GRADIENT might change starting with SAS
9.4M5. To restore the original appearance, specify
FILLTYPE=ALPHAGRADIENT.
598 Chapter 6 / Plot Statements

See DATASKIN= on page 593

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Notes If the value of numeric-column for a category is 0, that category is

represented as a 1-pixel-high bar. If the value of numeric-column is
missing for a category, a bin-width-wide space is reserved for that
category.

If n is the value of the numeric column for a given observation, then

that observation is used n times for the purposes of any statistical
computation. The marker for that observation is drawn n times,
which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a separate bar segment or bar for each unique group value in the
specified column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct bar or bar segment is created for each group value by varying the
visual attributes of the bar display features that are enabled by the current ODS
style or by the DISPLAY= option. The default group appearance for each display
feature is shown in the following table.

Display
Feature1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.
 HISTOGRAM Statement 599

Display
Feature1 Style Attributes That Control Default Group Appearance

If the active ODS style does not specify FillPattern, fill patterns
are not displayed by default.

Outline n if the outline is enabled with fill

color, ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is enabled with fill
pattern or if the outline is the only
display feature enabled,
ContrastColor and LineStyle
attributes of a GraphData1–
GraphDataN style element or the
GraphMissing style element

1 The default display features are determined by the DisplayOpts attribute of the
GraphHistogram style element.

Note If you specify a column in a SAS data set, the visual attributes for each
group value are assigned in data order. If you specify a column in a CAS
in-memory table, the visual attributes for each group value are assigned
in ascending order of the group column character values or of
unformatted numeric values.

Tips You can individually override the representations that are used to identify
the groups. For example, in some ODS styles, each distinct group value is
represented by a different line pattern for the bar outlines. In that case,
you can use the PATTERN= setting on the OUTLINEATTRS= option to
assign the same line pattern to all of the bar outlines.

By default, ODS Graphics limits the number of groups to 1000. When this
limit is exceeded, the GROUP= option is ignored. Use the GROUPMAX=
option in the ODS GRAPHICS statement to change the maximum number
of groups that can be processed.

See “Example 2: Histogram of Highway Mileage Grouped by Origin” on page

605

“DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

Default TRUE
600 Chapter 6 / Plot Statements

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the MISSING= system option changes the default missing
character, or a user-defined format is applied to the group value. In
those cases, the attributes of the missing group value are
determined by a GraphData1–GraphDataN style element instead of
by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

NBINS=positive-integer
specifies the number of bins. This option can be used with the BINSTART= and
BINWIDTH= options to specify the bins.

Default Determined by the system.

Ranges A positive integer in SAS 9.4 and earlier releases

2–10000 starting with SAS 9.4M1

Note When BINWIDTH=, BINSTART=, or both are specified with this option
and the resulting data range is greater than the unbinned input data
range, the histogram might be padded with zero-height bins in order to
accommodate the BINSTART=, BINWIDTH=, and NBINS= option
values.

Tip Starting with SAS 9.4M4, the limit on the number of bins is specified by
the ODS GRAPHICS statement NBINSMAX= option. The default value
is 10,000. See NBINSMAX= on page 1573.
 HISTOGRAM Statement 601

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the bar outlines.

Defaults For non-grouped data, the GraphOutlines style element.

In SAS 9.4M1 and earlier releases, for grouped data and filled bars,
the ContrastColor and LineStyle attributes of the GraphData1–
GraphDataN style elements, and the LineThickness attribute of the
GraphOutlines style element.

Starting with SAS 9.4M2, for grouped data and filled bars, the
ContrastColor attribute of the GraphData1–GraphDataN style
elements, and the LineThickness and LineStyle attributes of the
GraphOutlines style element.

Interaction For this option to have any effect, the outlines must be enabled by
the ODS style or by the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

SCALE=PERCENT | COUNT | PROPORTION | DENSITY

specifies the scale for the Y axis.

PERCENT
displays the percentages between 0 and 100 on the Y axis.
602 Chapter 6 / Plot Statements

COUNT
displays the frequency counts on the Y axis.

PROPORTION
displays the proportions between 0 and 1 on the Y axis.

DENSITY
displays the true density estimates on the Y axis.

Default PERCENT

Requirement If you need to include a histogram and a density plot in your graph
and both plots specify the same data column, place the
HISTOGRAM statement before the DENSITYPLOT statement, or
specify SCALE=DENSITY in the HISTOGRAM statement. Otherwise,
the density plot values might not scale properly to the histogram
bins on the dependent axis. This requirement applies even when the
histogram and density plot are in different cells in the graph.

Tip When SCALE=PERCENT, the response axis tick values do not

include a percent sign after each value. To add a percent sign after
each response axis tick value in that case, use
SCALE=PROPORTION, and then specify the options shown in the
following example for the response axis:
yaxisopts=(label="Percent" linearopts=(tickvalueformat=percent.))

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Note The X role represents the binned value. The Y role represents the
computed amount of X in the units specified by the SCALE=
option.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Percent")
 HISTOGRAM Statement 603

Default The column label or column name of the column assigned to the role.

Note The Y role represents the computed amount of X in the units specified
by the SCALE= option.

WEIGHT=numeric-column | expression
specifies a column that contains a bin-width calculation a priori weight for each
observation of the input data object.

Requirement The value must be nonnegative.

Note The values of the weight variable must be greater than zero. If the
weight value for an observation is zero, negative, or missing, that
observation is deleted from the analysis.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins. For example, if BINSTART=10 and BINWIDTH=10,
then using LEFTPOINTS would result in bins 10 - 20, 20 - 30, and so on. Using
RIGHTPOINTS would result in bins 0 - 10, 10 - 20, ...., and using MIDPOINTS
would result in bins 5 - 15, 15 - 25, ... .

Default MIDPOINTS

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The histogram’s default bin width is computed by using the number of observations
and the range of the data. When a curve is overlaid on the histogram, the
histogram’s bin width is used to scale the curve so that the area under the curve is
equal to the area of the histogram.
 604 Chapter 6 / Plot Statements

The X axis and Y axis are linear by default. You can change axis properties with the
XAXISOPTS= and YAXISOPTS= options of the layout statement.

Examples

Example 1: Histogram of Vehicle Weight

The following graph was generated by the “Example Program” on page 604:

Example Program
proc template;
define statgraph histogram;
begingraph;
entrytitle "Histogram of Vehicle Weights";
layout overlay /
xaxisopts=(label="Vehicle Weight (LBS)")
yaxisopts=(griddisplay=on);
histogram weight;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=histogram;

run;
 HISTOGRAMPARM Statement 605

Example 2: Histogram of Highway Mileage Grouped by Origin

The following graph was generated by the “Example Program” on page 605.

Example Program
proc template;
define statgraph histogram;
begingraph;
entrytitle "Highway Mileage Distribution by Origin";
layout overlay /;
histogram mpg_highway / name="histogram" group=origin;
discretelegend "histogram" / title="Origin:";
endlayout;
endgraph;
end;
run;

proc sgrender template=histogram data=sashelp.cars;

run;

HISTOGRAMPARM Statement
Creates a univariate histogram for specified values of bin midpoints and bin frequencies.

Restrictions: Only uniform width bins are supported.

 606 Chapter 6 / Plot Statements

Starting with SAS 9.4M4, the number of bins is limited to 10,000. If the number of
bins exceeds the limit, the histogram is not drawn. A warning message is written to
the SAS log in that case.
Tips: Starting with SAS 9.4M4, the limit on the number of bins is specified by the ODS
GRAPHICS statement NBINSMAX= option. The default value is 10,000. See
NBINSMAX= on page 1573.
Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
HISTOGRAMPARM X=numeric-column | expression
Y=non-negative-numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bars (outline and fill).
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the interior fill area of the bars.
FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT
specifies the bar fill type.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line properties of the bar outlines.
XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS
specifies whether the input X values represent midpoints, lower
endpoints, or upper endpoints of the bins.

Axes options
BINAXIS=TRUE | FALSE
 HISTOGRAMPARM Statement 607

specifies whether to use bins as the basis for the axis tick marks.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
histogram bin.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies a column for the bar labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the bar labels.
DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin
endpoints (TRUE) or at the bin midpoints (FALSE).
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.
608 Chapter 6 / Plot Statements

Required Arguments
X=numeric-column | expression
specifies the column for the midpoint values.

Requirement When you plot a SAS data set, you must sort the data in ascending
order of the specified X column before you plot the data. When
you plot a CAS in-memory table, you do not need to sort the data.
The data is automatically sorted in ascending order of the X
column formatted character values or of the unformatted numeric
values before the data is plotted.

Y=non-negative-numeric-column | expression
specifies the column for the frequency values. These values can be frequency
counts, percentages, or proportions between 0 and 1.

Requirement The Y values cannot be negative.

Note If the value of Y= is 0 for a category, that category is represented

as a 1-pixel-high bar. If the value of Y= is missing for a category, a
bin-width-wide space is reserved for that category.

Optional Arguments
BINAXIS=TRUE | FALSE
specifies whether to use bins as the basis for the axis tick marks.

TRUE
specifies that the ENDLABELS= option determines how the axis ticks and
value labels are displayed.

FALSE
specifies that standard axes are used. Bin boundaries and midpoints that are
set by the ENDLABELS= option are ignored.

Default TRUE

Interactions This option is ignored when this histogram is not the primary plot.
For more information about primary plots, see “When Plots Share
Data and a Common Axis” on page 1046.

BINAXIS=TRUE is ignored when the histogram is placed on a log

axis.

When this option is set to TRUE, some X-axis options that are set
on the parent layout might not apply, such as INTEGER=,
TICKVALUELIST=, TICKVALUESEQUENCE=, and
INCLUDERANGES=.

See “ENDLABELS=TRUE | FALSE” on page 614

“boolean ” on page xii for other Boolean values that you can use.
 HISTOGRAMPARM Statement 609

DATALABEL=column | expression
specifies a column for the bar labels. The labels appear at the top or end of each
bar, depending on the chart orientation.

Default No data labels are displayed

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the bar labels.

Default The GraphDataText style element

Interactions The DATALABEL= option must be specified for this option to have
any effect.

When text options are specified, any font properties that are not
specified (color, family, size, weight, and style) are derived from the
GraphDataText style element.

See “General Syntax for Attribute Options” on page 1599

“Text Options” on page 1606

DATALABELFITPOLICY=AUTO | NONE | ROTATE | SPLIT | SPLITALWAYS

specifies a policy for avoiding collisions among the bar labels when labels are
displayed.

AUTO
selects a collision avoidance policy based on the chart orientation and data
type.

For a numeric column with ORIENT=VERTICAL, AUTO rotates the labels if

they do not fit the midpoint spacing. For a character column, AUTO splits the
labels if they do not fit the midpoint spacing.

Note When ORIENT=HORIZONTAL, AUTO always draws the labels

horizontally.

Tip If character labels do not fit after splitting, then try using ROTATE
instead of AUTO.

See ORIENT= on page 618 for information about chart orientation.

NONE
does not attempt to fit bar labels that collide.

ROTATE
rotates the bar labels for vertical bars if the labels collide in the available
width.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

SPLIT
splits the label for vertical bars at a split character only if a split is needed at
that character in order to make the label fit the available space.

No split occurs at split characters that occur where a split is not needed. If
the label does not contain any of the specified split characters, then a split
610 Chapter 6 / Plot Statements

does not occur. In that case, if the label does not fit the available space, then
it might collide with the adjoining labels.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See DATALABELSPLITCHAR= for information about specifying the

split characters

SPLITALWAYS
splits the label for vertical bars at every occurrence of a split character.

If the label does not contain any of the specified split characters, then a split
does not occur.

Requirement The chart orientation must be vertical (ORIENT=VERTICAL).

See DATALABELSPLITCHAR= for information about specifying the

split characters

INSIDEPREFERRED
attempts to place the data labels in a horizontal bar chart inside the bars.

Note: This feature applies to SAS 9.4M5 and to later releases.

Restrictions This option applies to non-grouped or cluster-grouped

horizontal bar charts only.

Label splitting is not supported when a label is placed inside a

bar.

Requirement For a grouped horizontal bar chart, the

GROUPDISPLAY=CLUSTER option must be in effect.
Otherwise, the INSIDEPREFERRED policy is ignored.

Interaction If SEGMENTLABEL=TRUE is in effect, the INSIDEPREFERRED

policy is ignored, and the labels are placed outside.

Notes If any bar label does not fit inside its bar because the font
height is too high, all of the bar labels are placed outside.

If any bar label cannot fit inside the length of its bar, that label
is placed outside of its bar and is fit using the AUTO fit policy.

If the outside labels collide because the font height is greater

than the midpoint space, then all labels are dropped.

WARNING: The bar labels are suppressed. Use DATALABELFITPOLICY=NONE to

force the labels to be displayed.

Default AUTO

Requirement The DATALABEL= option must also be specified.

 HISTOGRAMPARM Statement 611

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split. When
multiple split characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in
the data label. In that case, all of the specified split characters together are
treated as a single split character.

When DATALABELFITPOLICY=SPLIT, data labels are split on a split character

only if a split is needed at that point in order to make the label fit. When
DATALABELFITPOLICY=SPLITALWAYS, the data labels are split
unconditionally on each occurrence of a split character. If the data label does
not contain any of the specified characters, then the label is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interactions The DATALABELFITPOLICY= option specifies the policy that is

used to manage the split behavior of the data labels.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops a split character from the data label when a split occurs at that
character.

Split characters at which a split does not occur are left in place. The
DATALABELFITPOLICY= option determines where the labels are split. When
DATALABELFITPOLICY=SPLIT, each label is split at a split character only
where a split is needed in order to make the label fit the available space. At
each split point, the split character is dropped, and the characters that
follow the split character, up to but not including the split character at the
next split point, are wrapped to the following line.
612 Chapter 6 / Plot Statements

When DATALABELFITPOLICY=SPLITALWAYS, each label is split at every

instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next
split character, are wrapped to the next line.

The following figure shows how label “Product*Group*1” is split when the
DATALABELSPLITCHARDROP=TRUE and DATALABELSPLITCHAR="*"
options are used with the SPLIT and SPLITALWAYS fit policies.

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is
needed at the second asterisk. The first asterisk is dropped, and Group*1
wraps to the next line. Notice that the second asterisk is not dropped in this
case. When DATALABELFITPOLICY=SPLITALWAYS, the label is split at
every occurrence of the asterisk. In this case, both asterisks are dropped, and
the characters that follow each asterisk wrap to the next line.

FALSE
includes the split characters in the data label.

The DATALABELFITPOLICY= option determines how the split characters are

displayed. When DATALABELFITPOLICY=SPLIT, each data label is split at a
split character only where a split is needed in order to make the label fit the
available space. A split might not occur at every split character in the label.
At each split point, the split character remains as the last character in the
current line. The characters that follow the split character, up to and
including the split character at the next split point, are then wrapped to the
following line. This process repeats until the entire data label is displayed.

When DATALABELFITPOLICY=SPLITALWAYS, each data label is split at

every instance of a split character in the label regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line.

The following figure shows how label “Product*Group*1” is split when the
DATALABELSPLITCHARDROP=FALSE and DATALABELSPLITCHAR="*"
options are used with the SPLIT and SPLITALWAYS fit policies.
 HISTOGRAMPARM Statement 613

In this example, when DATALABELFITPOLICY=SPLIT, the label is split at the

first occurrence of the asterisk in order to make the label fit. No split is
needed at the second asterisk. The characters that follow the first asterisk
wrap to the next line. When DATALABELFITPOLICY=SPLITALWAYS, the
label is split at every occurrence of the asterisk. Each asterisk remains as the
last character in the current line, and the characters that follow are wrapped
to the next line.

Default TRUE. A split character is dropped from thedata label when a

split occurs at that character.

Requirements The DATALABEL= option must also be specified.

The DATALABELFITPOLICY= option must specify SPLIT or

SPLITALWAYS.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
histogram bars with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.
614 Chapter 6 / Plot Statements

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

When FILLTYPE=GRADIENT is in effect, DATASKIN=SHEEN is

ignored. In that case, use one of the other skins.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bars (outline and fill).

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
bar fills. You can combine this option with FILLATTRS= to set one
transparency for the bar outlines but a different transparency for the
bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
a space-separated list of one or more of the following options, enclosed in
parentheses:

FILL
displays the bar fill color

FILLPATTERN
displays the bar fill pattern

Note This option is valid starting in SAS 9.4M1.

OUTLINE
displays the bar outline

STANDARD
specifies FILL and OUTLINE

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Default The GraphHistogram:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= , FILLATTRS= , and FILLPATTERNATTRS=

options to control the appearance of the bars.

ENDLABELS=TRUE | FALSE
specifies whether the axis ticks and value labels are drawn at the bin endpoints
(TRUE) or at the bin midpoints (FALSE).
 HISTOGRAMPARM Statement 615

Default FALSE.

Interactions This option is ignored if this plot is not the primary plot in the
parent layout. For more information, see the PRIMARY= option.

This option is ignored if BINAXIS= FALSE. By default,

BINAXIS=TRUE.

If the TICKS= suboption is specified in the XAXISOPTS= option,

then this option is ignored.

See “boolean ” on page xii for other Boolean values that you can use.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the interior fill area of the bars.

Default The GraphDataDefault style element

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

When a discrete attribute map is in effect for the histogram, the fill
colors that are specified by the attribute map take precedence over
COLOR. If the attribute map provides no color for a value, COLOR
prevails for that value.

Tip The DATATRANSPARENCY= option sets the transparency for bar

fills and bar outlines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLENDCOLOR=color
specifies the end color of the color gradient for the bar fill.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphWall style element Color attribute.

Requirement FILLTYPE=COLORGRADIENT must be in effect. Otherwise, this

option is ignored.

See “color ” on page xii

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
616 Chapter 6 / Plot Statements

Note: This option is valid starting in SAS 9.4M1.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

FILLTYPE=SOLID | ALPHAGRADIENT | COLORGRADIENT | GRADIENT

specifies the bar fill type.

Note: This feature applies to SAS 9.4M2 and to later releases.

SOLID
fills each bar with the color that is assigned to that bar.

ALPHAGRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This feature applies to SAS 9.4M5 and to later releases.

Tips The fill color is determined by a style element or by the FILLATTRS=

option COLOR= suboption. The initial transparency is determined by
the DATATRANSPARENCY= option or by the FILLATTRS= option
TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete attribute

map to set the initial transparency in the gradients for specific values.

COLORGRADIENT
fills each bar with a color gradient that starts with the fill color that is
assigned to that bar and ends with the color that is specified in the
FILLENDCOLOR= option.
 HISTOGRAMPARM Statement 617

Note: This feature applies to SAS 9.4M5 and to later releases.

Alias GRADIENT

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tip Use the DATATRANSPARENCY= option or the FILLATTRS=

option TRANSPARENCY= suboption to set the initial
transparency in the gradients.

See “FILLENDCOLOR=color” on page 615

GRADIENT
fills each bar with a color and a transparency gradient that starts at the bar
top with the initial transparency that is assigned to that bar, and ends with
full transparency at the bar baseline.

Note: This option applies to SAS 9.4M2 through SAS 9.4M4. Starting with
SAS 9.4M5, ALPHAGRADIENT replaces GRADIENT, and GRADIENT is
changed to an alias of COLORGRADIENT.

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Tips The initial fill color is determined by a style element or by the

FILLATTRS= option COLOR= suboption. The initial transparency
is determined by the DATATRANSPARENCY= option or by the
FILLATTRS= option TRANSPARENCY= suboption.

For grouped plots, use the FILLATTRS= option in a discrete

attribute map to set the initial transparency in the gradients for
specific values.

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

Note The output for SAS programs written prior to SAS 9.4M5 that
specify FILLTYPE=GRADIENT might change starting with SAS
9.4M5. To restore the original appearance, specify
FILLTYPE=ALPHAGRADIENT.

See DATASKIN= on page 613

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

618 Chapter 6 / Plot Statements

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis and the bars.

Default VERTICAL

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line properties of the bar outlines.

Default The GraphOutlines style element.

Interaction For this option to have any effect, the outlines must be enabled by
the ODS style or the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.
 HISTOGRAMPARM Statement 619

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, and DATALABEL.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
histogram bin. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the histogram
can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique HISTOGRAMPARM and user-
defined roles.

HISTOGRAMPARM roles include X, Y, and DATALABEL.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned
to the roles X and Y, as well as the column Pct, which is not
assigned to any pre-defined HISTOGRAMPARM role. The Pct
column must first be assigned a role:

ROLENAME=(TIP1=PCT)
TIP=(X Y TIP1)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X and Y.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.
620 Chapter 6 / Plot Statements

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

XVALUES=MIDPOINTS | LEFTPOINTS | RIGHTPOINTS

specifies whether the input X values represent midpoints, lower endpoints, or
upper endpoints of the bins.
 HISTOGRAMPARM Statement 621

Default MIDPOINTS

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
When a curve is overlaid on the histogram, the histogram bin width is used to scale
the curve so that the area under the curve is equal to the area of the histogram.

The X axis and Y axis are linear by default. You can change axis properties with the
XAXISOPTS= and YAXISOPTS= options of the LAYOUT OVERLAY statement.

Example: HISTOGRAMPARM Statement

The following graph was generated by the “Example Program” on page 622:
 622 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph histogramparm;
begingraph;
entrytitle "Histogram of Vehicle Weights";
layout overlay;
histogramparm x=midpoint y=frequency;
endlayout;
endgraph;
end;
run;

data bins;
input Midpoint Frequency;
label midpoint="Vehicle Weight (LBS)";
datalines;
2000 18
3000 208
4000 159
5000 36
6000 6
7000 1
;
run;

proc sgrender data=bins template=histogramparm;

run;

LINECHART Statement
Creates a line chart that is computed from input data.

Restrictions: GROUPDISPLAY= STACK has stringent requirements for the data. If the
requirements are not met for all the data, then the line chart is not drawn. If the
LINECHART statement is in a LAYOUT DATALATTICE or LAYOUT DATAPANEL
block, then the requirements are tested for all of the data, not for each individual
panel.
The category axis (the X axis when ORIENT=VERTICAL or the Y axis when
ORIENT=HORIZONTAL) is always discrete.
The response axis (the Y axis when ORIENT=VERTICAL or the X axis when
ORIENT=HORIZONTAL) is always numeric.
The LINECHART statement performs discrete binning for a numeric category
column only.
Note: Specifying only the CATEGORY= option creates a computed line chart with vertices
representing frequency counts or percents of unique CATEGORY values. For a non-
grouped chart, specifying both the CATEGORY= and RESPONSE= options creates a
computed line chart with vertices representing the summarized values of the
RESPONSE values that are categorized by unique CATEGORY values.
 LINECHART Statement 623

Tips: The line segments in the chart always join the categorical values in the order in
which they appear on the axis. By default, the vertices in the line chart appear in
the order in which the X values occur in the input data. To change the categorical
axis tick value order, use the SORTORDER= or TICKVALUELIST= suboption of the
DISCRETEOPTS= option for the X axis. For example:
XAXISOPTS=(DISCRETEOPTS=(SORTORDER=ASCENDINGFORMATTED))

The response axis of the line chart includes the zero value by default. You can use
the BASELINEINTERCEPT= to change the Y intercept.
By default, missing category values are ignored. To treat missing category values as
a category, include the INCLUDEMISSINGDISCRETE=TRUE option in the
BEGINGRAPH statement.
Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.
See: “LAYOUT DATALATTICE Statement” on page 51
“LAYOUT DATAPANEL Statement” on page 79
“Axis Options in Layouts” on page 1055
“BEGINGRAPH Statement” on page 25

Syntax
LINECHART CATEGORY=column | expression < /options > ;

LINECHART CATEGORY=column | expression

RESPONSE=numeric-column | expression < /options > ;

Summary of Optional Arguments

Appearance options
BREAK=TRUE | FALSE
breaks the line at missing values of the RESPONSE variable.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line,
marker, and fill colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the lines and markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, and vertex
labels, if displayed.
624 Chapter 6 / Plot Statements

DISPLAY=(display-options) | STANDARD | ALL

specifies which line-chart features to display.
FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled area.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
GROUPDISPLAY=OVERLAY | STACK
specifies how to display grouped lines.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and
line pattern) and fill attributes to one of the GraphData1–GraphDataN
style elements.
LINEATTRS=style-element | (line-options)
specifies the appearance of the line.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the data markers.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.
SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.

Axes options
BASELINEINTERCEPT=number | AUTO | AXISMIN | AXISMINEXTEND |
AXISMAX | AXISMAXEXTEND
specifies the Y-intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
line.
TIPFORMAT=(role-format-list)
 LINECHART Statement 625

specifies display formats for tip columns.

TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
VERTEXLABEL=TRUE | FALSE
specifies whether to label the vertices with their response value (or
statistic).
VERTEXLABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the vertex labels.
VERTEXLABELFORMAT=format
specifies the format used to display the vertex label response or statistic.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate line for each unique group value in the specified
column.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a line segment, marker, or fill
area is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
STAT=FREQ | PCT | PROPORTION | SUM | MEAN
specifies the statistic to be computed for the Y axis.

Required Argument
CATEGORY=column | expression
specifies the column or expression for the category values. Duplicated category
values are summarized into a unique value. All values are treated as discrete.
626 Chapter 6 / Plot Statements

Optional Arguments
BASELINEINTERCEPT=number | AUTO | AXISMIN | AXISMINEXTEND |
AXISMAX | AXISMAXEXTEND
specifies the Y-intercept for the baseline.

number
specifies the Y coordinate of the baseline.

This value is included in the data range that is reported by the line chart.

Interaction When number is specified, if necessary, the response axis data

range is extended to include the baseline intercept. When a
logarithmic response axis is requested and number is 0 or a
negative value, the response axis reverts to a linear axis. To
restore the log axis in that case, set BASELINEINTERCEPT= to a
positive value.

AUTO
bases the intercept on the response axis range.

The intercept is determined as follows:

n If the response axis range has both positive and negative values or
contains 0, then the intercept is 0.
n If the response axis range contains all positive values, then AUTO is
interpreted as AXISMINEXTEND.
n If the response axis range contains only negative values, then AUTO is
interpreted as AXISMAXEXTEND.

AXISMIN
places the baseline at the minimum value of the axis range.

AXISMINEXTEND
places the baseline at the start of the minimum offset in the wall.

This location corresponds to the bottom edge of the wall when there is no
inner margin plot and the axis is not reversed. If there is an inner margin plot
at the bottom, then the baseline is placed at the boundary of the inner
margin and the minimum offset.

AXISMAX
places the baseline at the maximum value of the axis range.

AXISMAXEXTEND
places the baseline at the start of the maximum offset in the wall.

This location corresponds to the top edge of the wall when there is no inner
margin plot, and the axis is not reversed. If there is an inner margin plot at
the top, then the baseline is placed at the boundary of the inner margin and
the maximum offset.

Default AUTO

Interactions When GROUPDISPLAY= STACK is in effect, this option is ignored,

and the plot is drawn as if BASELINEINTERCEPT=0.
 LINECHART Statement 627

When DISPLAY= includes FILL, the fill extends to the baseline that
is specified by the BASELINEINTERCEPT= option.

BREAK=TRUE | FALSE
breaks the line at missing values of the RESPONSE variable.

Default FALSE

Note If BREAK=FALSE, then missing values are skipped and a continuous

line is drawn.

See “boolean ” on page xii for other Boolean values that you can use.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line, marker,
and fill colors.

Note: This feature applies to SAS 9.4M3 and to later releases.

628 Chapter 6 / Plot Statements

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement The COLORRESPONSE values should remain constant for each

group value in a grouped plot and for the entire plot in an
ungrouped plot. If the COLORRESPONSE column has multiple
values for a single GROUP value or for a non-grouped plot,
unexpected results might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

When this option is specified without the GROUP= option, only a

single line is generated for the plot, and the line color is derived
from the COLORRESPONSE= value.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map. The line and marker colors in
that case are controlled by the ContrastColor attribute of the
GraphDataDefault style element.

When fill is not displayed, this option overrides suboption COLOR=

in the LINEATTRS= and MARKERATTRS= options, and varies the
line and marker colors according to the color gradient or the
attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorRamp style

element defines the fill color gradient, and the ThreeColorAltRamp
style element defines the line color gradient.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the lines and markers. The following figure
shows lines and CIRCLEFILLED markers with each of the skins applied.
 LINECHART Statement 629

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The skin appearance is based on the color that is in effect for the
lines and markers.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, and vertex labels,
if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
fills. You can combine this option with FILLATTRS= to set one
transparency for the lines but a different transparency for the fills. For
example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which line-chart features to display.

(display-options)
a space-separated list of one or more display options enclosed in
parentheses.

The list can include one or more of the following options:

630 Chapter 6 / Plot Statements

FILL
displays a filled area between the line and the baseline

Interaction When GROUPDISPLAY=STACK, DISPLAY=FILL fills between

adjacent group lines except for the first group, which fills to
the baseline.

LINE
displays line segments that join the vertices

MARKERS
displays markers at each vertex

STANDARD
specifies LINE only

ALL
specifies all features: FILL, LINE, and MARKERS

Default STANDARD

Tip Use the LINEATTRS=, MARKERATTRS=, and FILLATTRS= options to

control the appearance of the line segments, markers, and fill,
respectively.

FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled area.

Defaults For non-grouped data, the Color attribute of GraphDataDefault

style element.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements is used.

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the fill colors vary according to the gradient.

Note When style-element is specified, only the style element’s Color

attribute is used.

Tip The DATATRANSPARENCY= option sets the transparency for the

fills and the lines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the lines but a
different transparency for the fills. For example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599

“Fill Color Options” on page 1600

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
 LINECHART Statement 631

TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.

When this option is TRUE, the fill color and outline color for filled markers
are determined in the following ways:
n If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and
the marker outlined color is derived from the GraphData1–GraphDataN
style elements ContrastColor attribute.
n If the GROUP= option is not specified, then by default, the fill color is
derived from the GraphDataDefault style elements Color attribute, and
the marker outlined color is derived from the GraphOutline style
elements ContrastColor attribute.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set
this option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 631

MARKERFILLATTRS= on page 635

MARKEROUTLINEATTRS= on page 636

“boolean ” on page xii for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a separate line for each unique group value in the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value is represented in the plot by a different

combination of color, line pattern, and marker symbol. Lines and
markers vary according to the ContrastColor, LineStyle, and
MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements.

Interaction When both the GROUP= and COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.
632 Chapter 6 / Plot Statements

Tips The representations that are used to identify the groups can be
overridden individually. For example, each distinct group value is
represented by a different line pattern for the lines, but you can use
the PATTERN= suboption of the LINEATTRS= option to assign the
same line pattern to all lines.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRMAP Statement” on page 1505

GROUPDISPLAY=OVERLAY | STACK
specifies how to display grouped lines.

OVERLAY
displays group values overlaid on top of each other.

STACK
displays group values as stacked lines.

Default OVERLAY

Restriction When STACK is in effect, if any response value is negative or if any

crossing of the group value with the category is absent or is a
missing value, then the chart is not drawn and a warning message is
written to the SAS log.

Interaction When STACK is in effect, the BASELINEINTERCEPT= option is

treated as if it is set to zero.

Tip When the response axis is linear, STAT=MEAN or STAT=PCT, and

GROUPDISPLAY=STACK, the axis tick values might be displayed as
integer values. When the response axis is linear, STAT=MEAN or
STAT=PCT, and GROUPDISPLAY=OVERLAY, the axis tick values
might be displayed as decimal values. To keep the integer axis
values for both cases, you can specify the INTEGER=TRUE option
for the response axis. See INTEGER= on page 1085.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.
 LINECHART Statement 633

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.
634 Chapter 6 / Plot Statements

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) and fill attributes to one of the GraphData1–GraphDataN style
elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default If the RESPONSE= option is specified, then the response variable

label is used. Otherwise, the CATEGORY= variable label is used. If a
label is not assigned to the response variable or category variable,
then the variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | (line-options)
specifies the appearance of the line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the LineThickness attributes of the

GraphDataDefault style element, and the ContrastColor and
LineStyle attributes of the GraphData1–GraphDataN style elements.

Interaction When COLORRESPONSE= is in effect, the LINEATTRS= suboption

COLOR= is ignored, and the line fill colors vary according to the
gradient.
 LINECHART Statement 635

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s COLOR=

suboption is ignored. In that case, to specify the marker fill color,
use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that
case, all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data. When the SYMBOL= suboption is specified in
that case, all markers have the same symbol, and the symbol color
alone distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display

of the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current
ODS style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
636 Chapter 6 / Plot Statements

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element

For grouped data, the Color attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

Note When style-element is specified, only the style element’s Color

attribute is used.

See “General Syntax for Attribute Options” on page 1599

“Fill Color Options” on page 1600

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current
style or by the DATASKIN= option. In the latter case, the outline is
set by the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1599

“Line Options” on page 1602

NAME="string"
assigns a name to this plot statement for reference in other template
statements. This option is used mostly in the CONTINUOUSLEGEND on page
1291 statement in order to coordinate the use of colors and line patterns
between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.
 LINECHART Statement 637

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the Y axis.

Default VERTICAL

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY, RESPONSE, DATALABEL, and
GROUP.

Note The column values for a custom tip are not summarized using the
STAT= statistic. For a given category value, the custom tip column
values for that category value should be the same. Otherwise,
unexpected results might occur.

SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.
638 Chapter 6 / Plot Statements

Default FALSE. Straight line segments are used if the vertices are to be
connected.

Interaction This option is ignored when GROUPDISPLAY=STACK.

See “boolean ” on page xii for other Boolean values that you can use.

STAT=FREQ | PCT | PROPORTION | SUM | MEAN

specifies the statistic to be computed for the Y axis. For line charts with no
RESPONSE= column:

FREQ frequency count

PCT percentages between 0 and 100 inclusive
PROPORTION proportions between 0 and 1 inclusive

For line charts with a RESPONSE= column:

SUM
MEAN

Defaults SUM for line charts that specify the RESPONSE= argument.

FREQ for line charts that do not specify the RESPONSE= argument.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a line. If
this option is used, then the information specified replaces all of the information
that is displayed by default. You can specify roles for columns that do not
contribute to the line chart along with roles that do.

(role-list)
an ordered, space-separated list of unique LINECHART roles and user-
defined roles.

LINECHART roles include CATEGORY, RESPONSE, COLORRESPONSE,

DATALABEL, and GROUP.

Note The COLORRESPONSE role is valid starting with SAS 9.4M1.

Tip User-defined roles are defined with the ROLENAME= option.

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: CATEGORY, RESPONSE, and
GROUP.

Restriction Data tips are available only for graphs that are written to the ODS
HTML destination.

Requirement To generate data tips, include an ODS GRAPHICS ON statement

that specifies the IMAGEMAP option. See “ODS GRAPHICS
 LINECHART Statement 639

Statement” on page 1561 for information about the ODS

GRAPHICS statement.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note The RESPONSE role represents the computed values for the
CATEGORY role (and RESPONSE= role), based on the STAT=
option.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

Example To display data tips for the columns assigned to the roles X and Y
as well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.
640 Chapter 6 / Plot Statements

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a line segment, marker, or fill area is
selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each line that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable line segments, markers, and fill areas, you
must include an ODS GRAPHICS ON statement that specifies the
IMAGEMAP option, and you must write the output to the ODS
HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Notes For non-grouped data, the values of the column are expected to be
same for each unique X value. If they are not, then the results
might be unpredictable.

For grouped data, the values of the column are expected to be the
same for each unique X and GROUP combination.

Tips The URL value can be blank for some X values, meaning that no
action is taken when the line segments for those X values are
selected.

The URL value can be the same for different X values, meaning
that the same action is taken when the line segments for those X
values are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

VERTEXLABEL=TRUE | FALSE
specifies whether to label the vertices with their response value (or statistic).

Default FALSE

Interaction When GROUPDISPLAY=STACK is in effect, vertex labeling displays

the sum of the vertex responses per category.
 LINECHART Statement 641

Note By default, the vertex-label format is derived from the format that is
assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response-
column format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Tips You can modify the visual attributes for the label by using the
VERTEXLABELATTRS= option.

You can modify the text format by using the

VERTEXTLABELFORMAT= option.

See “boolean ” on page xii for other Boolean values that you can use.

VERTEXLABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the vertex labels.

Default The GraphDataText style element.

Requirement VERTEXLABEL=TRUE must be in effect for this option to have any

effect.

Interaction If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties are derived from the GraphDataText style
element.

Note When style-element is specified, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VERTEXLABELFORMAT=format
specifies the format used to display the vertex label response or statistic.

Default The vertex label format is derived from the format that is applied
to the response column or from BEST6 if no format is assigned.
The derived format retains the precision of the response-column
format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Requirement VERTEXLABEL=TRUE must be in effect for this option to have any

effect.

Note When a vertex-label format is specified with this option, the vertex
labels are formatted as specified by format. The specified format is
not automatically expanded to accommodate values that are too
wide.
642 Chapter 6 / Plot Statements

Tip If you want the vertex-label format to expand automatically for

the response-column values, specify the format for the response
column rather than in this option.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Optional Response Argument

RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Details
A line chart shows the relationship of one variable to another as trends in the data
over a period of time. The trends are shown by connecting the successive data
points with a line. Typically, a line chart is used to chart a response value against a
discrete categorical value where each value on the horizontal axis has only one
corresponding value on the vertical axis. A grouping variable can be used to show
multiple trends based on the group values.

In a line chart, the category axis (X) is always discrete, and the response axis (Y) is
always linear. The line segments in the chart always join the categorical values in
the order in which they appear on the axis. The vertices in the line chart appear in
the order in which the categorical values appear in the input data.
 LINECHART Statement 643

Examples

Example 1: Grouped Line Chart with Custom Line and Fill

Attributes
This example shows you how to generate a simple line chart by using the
LINECHART statement and how to use the statement options to customize the
plot.

The following graph was generated by the “Example Program” on page 643. It
shows the trend of the average closing price of the IBM, Intel, and Microsoft stocks
from 1995 through 2005.

Example Program
Here is the SAS code for this example.
/* Extract per-year data for 1995 through 2005 from SASHELP.STOCKS */
data linechartdata;
set sashelp.stocks(where=(year(date) between 1995 and 2005));
year=year(date);
label year="Year";
run;

/* Define the line chart template */

proc template;
define statgraph linechart;
begingraph;
entrytitle "Stock Index Performance: 1995 - 2005";
layout overlay /
644 Chapter 6 / Plot Statements

/* Add a grid */
xaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray));

/* Generate the line chart */

linechart category=year response=close / name="linechart"

/* Compute the mean statistic */

stat=mean

/* Group by stock to draw a line for each stock */

group=stock

/* Set the baseline at 0 */

baselineintercept=0

/* Display the lines and fill */

display=(line fill)

/* Specify the line attributes */

lineattrs=(thickness=2)

/* Specify the fill attributes */

fillattrs=(transparency=0.8);

/* Add a legend */
discretelegend "linechart";
endlayout;
endgraph;
end;
run;

/* Generate the chart */

proc sgrender data=linechartdata template=linechart;
run;

Example 2: Grouped Line Chart with Discrete Attribute Map

This example shows you how to create a more flexible template that enables you to
easily generate multiple line graphs that use different data. It also shows you how
to customize the plot appearance by using a discrete attribute map.

The following graph was generated by the “Example Program” on page 645. It
shows the trend of the average monthly closing price of the IBM, Intel, and
Microsoft stocks for 2001.
 LINECHART Statement 645

Example Program
Here is the SAS code for this example.
/* Define the line chart template */
proc template;
define statgraph linechart;
begingraph;
/* Create a dynamic variable for the year */
dynamic year;

/* Define the display attributes for each stock. Since

DISCRETEATTRMAP does not support the MARKERATTRS=
suboption SIZE=, the marker size is set separately. */
discreteattrmap name="stocks" / ignorecase=true ;
value "IBM" /
markerattrs=(color=blue symbol=trianglefilled)
lineattrs=(color=lightblue pattern=solid);
value "Intel" /
markerattrs=(color=red symbol=circlefilled)
lineattrs=(color=verylightred pattern=solid);
value "Microsoft" /
markerattrs=(color=orange symbol=squarefilled)
lineattrs=(color=verylightorange pattern=solid);
enddiscreteattrmap ;

/* Associate the attribute map with input data column Stock

and
assign the name STOCKATTRS to the named association */
discreteattrvar attrvar=stockattrs var=stock
attrmap="stocks" ;
 646 Chapter 6 / Plot Statements

entrytitle "Stock Index Performance in " year;

layout overlay /
/* Add a grid */
xaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(pattern=dot
color=lightgray));
linechart category=date response=close / name="linechart"
/* Compute the mean */
stat=mean

/* Group by stock using the specified attributes */

group=stockattrs

/* Display the lines and markers */

display=(line markers)

/* Set the marker size */

markerattrs=(size=6)

/* Show vertex labels and specify their attributes */

vertexlabel=true
vertexlabelattrs=(size=7pt)
vertexlabelformat=dollar4.0;

/* Add a legend */
discretelegend "linechart";
endlayout;
endgraph;
end;
run;

/* Create a macro that generates a line chart for a specific year */

%macro genchart(year=);
/* Generate the chart */
proc sgrender data=sashelp.stocks template=linechart;
where year(date)=&year; /* Get the data for the specified year */
format date monname3.; /* Format the date as 3-letter month */
dynamic year="&year;"; /* Pass the year to the template */
run;
%mend genchart;

/* Generate a chart for 2001 */

%genchart(year=2001);

LINEPARM Statement
Creates a straight line specified by a point and a slope.

Requirement: A LINEPARM statement can be used only within a 2-D layout (OVERLAY,
OVERLAYEQUATED, DATALATTICE, or DATAPANEL). Another plot statement that
 LINEPARM Statement 647

is derived from data values that provide boundaries for the axis area must be
included.

Syntax
LINEPARM X=number | numeric-column | expression
Y=number | numeric-column | expression
SLOPE=number | numeric-column | expression < /options >;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the data for the line is considered when the data
ranges are determined for the axes.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line.
EXTEND=TRUE | FALSE
specifies whether the line is to be drawn to the area bounded by the axes.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the line.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
CURVELABEL="string" | column | expression
specifies a label for the line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the line label.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN
specifies the position of the line label relative to the line end points.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the line label at the specified split characters.
648 Chapter 6 / Plot Statements

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the line label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the line label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the line label
block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a separate parameterized line plot for each unique group value of
the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=number | numeric-column | expression
specifies the X coordinate of a point. By default, if the specified value is outside
of the data range, then the data range is extended to include the specified
intercept. This behavior can be changed with the CLIP= option. If a numeric-
column is specified and the column contains missing values, then no line is
drawn for the missing values.

Values are displayed in the units of the data.

Y=number | numeric-column | expression

specifies the Y coordinate of a point. By default, if the specified value is outside
of the data range, then the data range is extended to include the specified
intercept. This behavior can be changed with the CLIP= option. If a numeric-
column is specified and the column contains missing values, then no line is
drawn for the missing values.

Values are displayed in the units of the data.

SLOPE=number | numeric-column | expression

specifies the slope of the line. Slope can be positive or negative. SLOPE=0
creates a line parallel to the X axis. SLOPE=. (a missing value) creates a line
parallel to the Y axis.
 LINEPARM Statement 649

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the data for the line is considered when the data ranges are
determined for the axes.

FALSE
specifies that the data for the line contributes to the data range for each
axis.

Each axis might be extended in order to force the display of the line. When
CLIP=FALSE, the SLOPE= option determines how the X= and Y= values
contribute to the axis data range in the following ways:
n If SLOPE=0, then only the Y= values contribute to the axis data range.

n If SLOPE=. (missing), then only the X= values contribute to the axis data
range.
n If SLOPE= is neither 0 nor missing, then the X= and Y= values contribute
to the axis data range.

TRUE
specifies that the data for the line is ignored when axis scales are being
established.

Each axis scale is determined by the other plots in the layout. In this case,
the line might not be displayed if its data range is not within the data ranges
of the other plots.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABEL="string" | column | expression

specifies a label for the line.

Default No line label is displayed

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 654

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the line label.

Defaults For non-grouped data, the GraphValueText style element.

650 Chapter 6 / Plot Statements

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

Tip When the GROUP= option is used, each distinct group value might
be represented by a different color. The line label that is associated
with the group is assigned the group color. This option can be used
to specify a single color for all line labels in a plot, without affecting
the line colors.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the line label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the LINEPARM is used in multi-cell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN

specifies the position of the line label relative to the line end points. This option
is used in conjunction with the CURVELABELLOCATION= option to determine
where the line label appears.

AUTO
automatically positions the line label near the line boundary along unused
axes whenever possible (typically Y2 and X2) in order to avoid collision with
tick values.
 LINEPARM Statement 651

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, upper
right).

MIN
forces the line label to appear near minimum line values (typically, lower
left).

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

MAX when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION= INSIDE is

specified.

Interaction For this option to take effect, the CURVELABEL= option must also
be specified.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the curve label might fall outside of the graph area. In
that case, the curve label might not be displayed or might be
positioned incorrectly.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the line label at the specified split characters. When a
line label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the line label can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the line label. In that case, all of the specified split characters
together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the line label

is split unconditionally at each occurrence of any of the specified split
652 Chapter 6 / Plot Statements

characters. If the line label does not contain any of the specified characters,
then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the line label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the line label text.

TRUE
drops the split characters from the line label text.

FALSE
includes the split characters in the line label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a line label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"
 LINEPARM Statement 653

Note: The horizontal line to the left of the label represents the maximum end of
the line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the line label block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the line for reference.
654 Chapter 6 / Plot Statements

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the line label.

EXTEND=TRUE | FALSE
specifies whether the line is to be drawn to the area bounded by the axes.

Default FALSE

Note If this option is not specified, then there can be a small gap between
the line and the axis. The gap is controlled by the axis offset. If the
offset is set to 0, then there is no gap.

See “boolean ” on page xii for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a separate parameterized line plot for each unique group value of the
specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of line color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. The line patterns
 LINEPARM Statement 655

vary according to the LineStyle attribute of the GraphData1–

GraphDataN and GraphMissing style elements.

Restriction When this option is used, the X, Y, and SLOPE arguments must
specify numeric columns.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors
and line patterns

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The LINEATTRS= option can be used to override the

representations that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the lines, letting the line color distinguish group
values. Likewise, LINEATTRS=(COLOR=BLACK) can be used to
assign the same color to all of the lines, letting the line pattern
distinguish group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.
656 Chapter 6 / Plot Statements

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.
 LINEPARM Statement 657

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The LINEPARM statement creates a straight line. You can generate a single line by
specifying a constant for each required argument. You can generate multiple lines
by specifying a numeric column for any or all required arguments. If any of the X= or
Y= columns contains a missing value, then no line is drawn. To request a vertical
line, specify SLOPE=. (specify a missing value as a constant or column value).

A LINEPARM statement can be used in any layout except GRIDDED or

OVERLAY3D layouts. The parent layout must include another plot statement that
is derived from data values that establish a data range for the axes. For example, it
can be used with a scatter plot or a histogram.

To draw vertical or horizontal reference lines, consider using the simpler

REFERENCELINE statement.
 658 Chapter 6 / Plot Statements

Example: LINEPARM Statement

The following graph was generated by the “Example Program” on page 658:

Example Program
The LINEPARM statement draws a line based on a point and the slope of the line
that passes through that point. You can use this statement to create a reference
line with any slope or, in this example, to draw a fit from a linear regression. Many
SAS/STAT procedures create output data sets containing a Y-intercept and slope
and coefficient for the linear regression equation.
proc template;
define statgraph lineparm;
begingraph;
entrytitle "Robust Fit of Height and Weight by Sex ";
layout overlay / xaxisopts=(offsetmax=0.35);
scatterplot x=height y=weight / group=sex
markercharacter=eval(substr(sex,1,1))
markercharacterattrs=(size=5pt) datatransparency=0.7;
lineparm x=0 y=intercept slope=slope /
name="Line" group=sex clip=true
curvelabel=eval("Weight = "||put(slope,5.3)||
" * Height + "||put(intercept,6.1))
curvelabellocation=inside
curvelabelattrs=(size=8pt);
discretelegend "Line";
endlayout;
endgraph;
 LOESSPLOT Statement 659

end;
run;

proc sort data=sashelp.heart(keep=height weight sex)

out=heart;
by sex;
run;

ods exclude all;

proc robustreg data=heart method=m
outest=stats(rename=(height=slope));
by sex;
model weight=height;
run;

data all;
merge heart stats(keep=intercept slope sex);
run;

ods select all;

proc sgrender data=all template=lineparm;
run;

LOESSPLOT Statement
Creates a fitted loess curve computed from input data.

Restriction: The LOESSPLOT statement supports only models of one independent and one
dependent variable.
Note: If the input data contains a large number of observations, then it might take several
minutes to generate the plot.
Tips: By default, the LOESSPLOT statement will process up to 5000 observations. If the
input data exceeds 5000 observations, then the plot is not generated. In that case,
you can use the ODS GRAPHICS statement LOESSMAXOBS= option to extend the
limit. See “Details” on page 671.
Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.

Syntax
LOESSPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-options> <options>>;
660 Chapter 6 / Plot Statements

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the loess curve.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the loess curve.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the loess curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the curve labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the curve label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the curve label relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label
block.
 LOESSPLOT Statement 661

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of curves from just the observations that
correspond to each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.

Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the loess curve.

Default No curve label is displayed

Interaction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the curve labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the curve label relative to the plot area.
662 Chapter 6 / Plot Statements

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the LOESSPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve
labels appear.

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the curve label relative to the curve line. This option is
used in conjunction with the CURVELABELLOCATION= option to determine
where the curve label appears.

AUTO
automatically positions the curve label near the curve boundary along
unused axes whenever possible (typically Y2 and X2) in order to avoid
collision with tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the curve label to appear near maximum curve values (typically,
upper right).

MIN
forces the curve label to appear near minimum curve values (typically, lower
left).

START
forces the curve label to appear near the beginning of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the curve label to appear near the end of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

 LOESSPLOT Statement 663

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the curve label might fall outside of the graph area. In
that case, the curve label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When
a curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the curve label. In that case, all of the specified split characters
together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve

label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters,
then the label is not split.
664 Chapter 6 / Plot Statements

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.

TRUE
drops the split characters from the curve label text.

FALSE
includes the split characters in the curve label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a curve label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"
 LOESSPLOT Statement 665

Note: The horizontal line to the left of the label represents the maximum end of
the curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the curve for reference.
666 Chapter 6 / Plot Statements

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the loess curve.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of curves from just the observations that correspond to
each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–
GraphDataN style elements.

Restriction The input data must be sorted by the GROUP= column.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors
and line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Tips The LINEATTRS= option can be used to override the

representations that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
 LOESSPLOT Statement 667

pattern to all of the loess curves, letting the line color distinguish
group values. Likewise, LINEATTRS=(COLOR=BLACK) can be used
to assign the same color to all of the curves, letting the line pattern
distinguish group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.
668 Chapter 6 / Plot Statements

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the loess curve.

Default The GraphFit style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.
 LOESSPLOT Statement 669

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles
are automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y
670 Chapter 6 / Plot Statements

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Loess Regression Options

ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLM="name"
produces confidence limits for a mean predicted value for each observation. The
confidence level is set by the ALPHA= option.

Interaction "name" is a unique name within the template that is case sensitive
and cannot contain spaces. It must be assigned in order for the
confidence limits to be computed. To display confidence limits, you
must use this name as the required argument of a MODELBAND
statement.

DEGREE=1 | 2
specifies the degree of the local polynomials to use for each local regression.
The valid values are 1 for local linear fitting or 2 for local quadratic fitting.

Default 1

INTERPOLATION=LINEAR | CUBIC
specifies the degree of the interpolating polynomials used for blending local
polynomial fits at the kd tree vertices.

CUBIC
cubic polynomials

LINEAR
linear polynomials

Default LINEAR

MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the loess
curve as well as confidence limits.

Default 201

SMOOTH=AUTO | positive-number
specifies a smoothing parameter value. The value represents the proportion of
observations that are used for local regression.

AUTO
automatically selects a smoothing parameter value.
 LOESSPLOT Statement 671

positive-number
specifies a smoothing parameter value.

Range (0,1]

See What is loess regression? on blogs.sas.com.

The LOESS Procedure in SAS/STAT User’s Guide for information

about the fitting methodology that the LOESSPLOT statement uses.

Default AUTO

REWEIGHT=NONE | positive-integer
specifies the number of iterative re-weighting steps to be done. Such iterations
are appropriate when there are outliers in the data or when the error distribution
is a symmetric long-tailed distribution.

Default NONE

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a
priori weights for a loess fit. The values of the weight column must be
nonnegative. If an observation’s weight is zero, negative, or missing, then the
observation is deleted from the analysis.

Details
The LOESSPLOT statement supports only statistical models of one independent
and one dependent variable. For more information about the fitting methodology,
see The LOESS Procedure in SAS/STAT User’s Guide.

In addition to the loess curve, the LOESSPLOT statement can compute confidence
levels for the fitted line. To display the confidence levels:

1 Use the CLM= option to declare a name for the confidence level of the mean .

2 Use a MODELBAND statement to refer this name. This statement draws a

confidence band from this information. See “MODELBAND Statement” on page
673 for information about how to control the appearance of the confidence
band.

By default, the LOESSPLOT statement will process up to 5000 observations. If the

input data contains more than 5000 observations, then the plot is not drawn and
the following note is written to the SAS log:

NOTE: The number of observations of the LOESS plot (nnnn) exceeds the
limit of 5000. Specify the LOESSMAXOBS option of the ODS GRAPHICS statement to
override the limit.

In that case, you can use the following statement to extend the limit:
ods graphics / loessmaxobs=nnnn
 672 Chapter 6 / Plot Statements

where nnnn is the new limit.

Note: When the input data contains a large number of observations, it might take
several minutes to generate the plot.

For more information about the LOESSMAXOBS= option, see “ODS GRAPHICS
Statement” on page 1561.

Example: LOESSPLOT Statement

The following graph was generated by the “Example Program” on page 672:

Example Program
proc template;
define statgraph loessplot;
begingraph;
entrytitle "Loess Fit Plot";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
loessplot x=weight y=mpg_highway / name="fitline"
alpha=0.05 legendlabel="Loess Fit";
discretelegend "fitline";
endlayout;
endgraph;
 MODELBAND Statement 673

end;
run;

proc sgrender data=sashelp.cars template=loessplot;

run;

MODELBAND Statement
Creates a band showing confidence limits for an associated smoother plot.

Requirement: A MODELBAND statement must be associated with a smoother statement

(LOESSPLOT, REGRESSIONPLOT, or PBSPLINEPLOT) that specifies a fitted model
and a type of confidence level to compute.
Interaction: Starting with SAS 9.4M2, a confidence band that depicts confidence limits for
individual predicted values (CLI) for a weighted spline plot or regression plot is
displayed as a high-low chart instead of a band.
Note: Grouping in the MODELBAND statement is controlled by the smoother statement.
To create a grouped model-band plot, specify the GROUP= option in the smoother
statement.

Syntax
MODELBAND "confidence-name" </options>;

Summary of Optional Arguments

Appearance options
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and the band
outline.
DISPLAY=(display-options) | STANDARD | ALL
specifies which band features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled modelband area.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the modelband outlines.

Axes options
XAXIS=X | X2
674 Chapter 6 / Plot Statements

specifies whether data are mapped to the primary X (bottom) axis or to

the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the upper and lower band labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.
CURVELABELLOWER="string"
specifies a label for the lower band limit.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the band label relative to the band line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve labels at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve labels can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label
blocks.
CURVELABELUPPER="string"
specifies a label for the upper band limit.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Argument
"confidence-name"
specifies the case-sensitive name assigned to a confidence option in a smoother
plot statement.
 MODELBAND Statement 675

Requirement confidence-name must have been assigned to the CLM= or CLI=

option on a smoother plot statement such as LOESSPLOT,
REGRESSIONPLOT, or PBSPLINEPLOT.

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the
ODS GRAPHICS statement.

OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS GRAPHICS
statement.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the upper and lower band labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

If the smoother statement’s GROUP= option is specified, then this

option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOWER="string"
specifies a label for the lower band limit.

Default No curve label is displayed for the lower band

Interaction If the smoother statement’s GROUP= option is specified, then this

option is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELUPPER="string"
specifies a label for the upper band limit.
676 Chapter 6 / Plot Statements

Default No curve label is displayed for the upper band

Interaction If the smoother statement’s GROUP= option is specified, then this

option is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the band labels relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the MODELBAND is used in multi-

cell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the band
labels appear.

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the band label relative to the band line.

AUTO
automatically positions the band labels near the band boundary along
unused axes whenever possible (typically Y2 and X2).

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the band label to appear near maximum band values (typically, upper
right)

MIN
forces the band label to appear near minimum band values (typically, lower
left)

START
forces the band label to appear near the beginning of the band.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

 MODELBAND Statement 677

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the band label to appear near the end of the band.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restrictions The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified.

The START and END settings are ignored if

CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABELLOWER= or

CURVELABELUPPER= option must also be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the band
label appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the band label might fall outside of the graph area. In
that case, the band label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve labels at the specified split characters.
When a curve label is split, the label is split on each occurrence of the specified
split characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.
678 Chapter 6 / Plot Statements

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve labels can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the curve label. In that case, all of the specified split characters
together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve

labels are split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters,
then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve labels or are
dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label blocks.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.

TRUE
drops the split characters from the curve label text.

FALSE
includes the split characters in the curve label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.
 MODELBAND Statement 679

The following figure shows an example of a curve label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the series line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve labels.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label blocks.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.
680 Chapter 6 / Plot Statements

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the band for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the band fill and the band outline.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the curve label.

Tip The FILLATTRS= option can be used to set transparency for just the
filled band area. You can combine this option with FILLATTRS= to set
one transparency for the band outline but a different transparency for
the band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD | ALL

specifies which band features to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays the band fill color

OUTLINE
displays the band outline

STANDARD
specifies FILL only

ALL
specifies all features: FILL and OUTLINE
 MODELBAND Statement 681

Default The GraphBand:DisplayOpts style reference.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the band.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled modelband area.

Defaults For non-grouped data, the GraphConfidence:Color style reference.

For grouped data, the GraphData1:Color–GraphDataN:Color style

references.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Note Grouping in the MODELBAND statement is controlled by the

smoother statement. To create a grouped model-band plot, specify
the GROUP= option in the smoother statement.

Tip The DATATRANSPARENCY= option sets the transparency for band

outline and the band fill. You can combine this option with
DATATRANSPARENCY= to set one transparency for the band
outline but a different transparency for the band fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interaction For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Note Grouping in the MODELBAND statement is controlled by the

smoother statement. To create a grouped model-band plot, specify
the GROUP= option in the smoother statement.

Tip In order to change the fill pattern for each group value in ODS styles
such as HTMLBlue that use color-priority attribute rotation, specify
ATTRPRIORITY=NONE in an ODS GRAPHICS statement or in the
BEGINGRAPH statement in your graph template. For more
682 Chapter 6 / Plot Statements

information, see “Attribute Rotation Patterns” in SAS Graph

Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction The smoother statement’s GROUP= option overrides this option.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the modelband outlines.

Defaults For non-grouped data, the GraphConfidence style element.

For grouped data, the GraphData1: ContrastColor–

GraphDataN:ContrastColor style references.

Interaction If DISPLAY=(FILL), then this option has no effect.

Note Grouping in the MODELBAND statement is controlled by the

smoother statement. To create a grouped model-band plot, specify
the GROUP= option in the smoother statement.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.
 MODELBAND Statement 683

(role-format-list)
a space-separated list of role-name = format pairs. The role-names X,
LIMITLOWER, LIMITUPPER, GROUP, and INDEX are available to indicate
which data tip values to format.

Example TIPFORMAT=(LIMITUPPER=5.3 LIMITLOWER=5.3)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs. The role-names X,
LIMITLOWER, LIMITUPPER, GROUP, and INDEX are available to indicate
which data tip values to label.

Example TIP=(X)
TIPLABEL=(X="Type")

Default The column label or column name of the column assigned to the role.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Requirement The setting for this option should be the same as for the smoother
statement referenced by the confidence-name .

Interaction The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Requirement The setting for this option should be the same as for the smoother
statement referenced by the confidence-name .

Interaction The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.
 684 Chapter 6 / Plot Statements

Example: MODELBAND Statement

The following graph was generated by the “Example Program” on page 684:

Example Program
proc template;
define statgraph modelband;
begingraph / attrpriority=none;
entrytitle "Spline Fit with Confidence Bands";
layout overlay / cycleattrs=true;
modelband "cliband" / name="confband1" display=all
legendlabel="90% CLI" datatransparency=0.4;
modelband "clmband" / name="confband2" display=all
legendlabel="90% CLM" datatransparency=0.4;
scatterplot x=weight y=mpg_highway /
markerattrs=(size=4pt symbol=circle)
datatransparency=0.3;
pbsplineplot x=weight y=mpg_highway / name="fitline"
clm="clmband" cli="cliband"
alpha=0.1 legendlabel="Spline Fit"
lineattrs=(pattern=solid thickness=2px);
discretelegend "fitline" "confband1" "confband2";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=modelband;

 MOSAICPLOTPARM Statement 685

run;

MOSAICPLOTPARM Statement
Creates a mosaic plot from pre-summarized categorical data.

Restriction: You can use the MOSAICPLOTPARM statement in GRIDDED, LATTICE, and
REGION layouts only.
Note: On z/OS hosts, the mosaic plot categories might appear in an order other than data
order.

Syntax
MOSAICPLOTPARM CATEGORY=(column-list)
COUNT=non-negative-numeric-column | expression < /options >;

Summary of Optional Arguments

Appearance options
COLORGROUP=column | discrete-attr-var
specifies the category column to use for discrete fill colors for the tiles.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies a numeric column to use to map tile fill colors to a continuous
gradient.
DATATRANSPARENCY=number
specifies the degree of the transparency of the tile fill, outlines, and the
values that are located inside the tiles if those values are displayed.
DISPLAY=(display-options ) | STANDARD | ALL
specifies which mosaic-plot features to display.
FILLATTRS=style-element | (fill-options )
specifies the appearance of the tile fill areas.
GUTTER=dimension | (dimension-list )
specifies the gutter (gap) between the splits.
OUTLINEATTRS=style-element | (line-options )
specifies the appearance of the tile outlines.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
SQUARED=TRUE | FALSE
specifies that a square aspect ratio be used for the plot area.
686 Chapter 6 / Plot Statements

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list ) | NONE
specifies the information that is displayed when the cursor is positioned
over a tile.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
INSIDEVALUEATTRS=style-element | style-element (text-options ) | (text-
options )
specifies the color and font attributes of the category values when they
are located inside a tile.
LABELATTRS=style-element | style-element (text-options ) | (text-options )
specifies the color and font attributes of the category labels.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
VALUEATTRS=style-element | style-element (text-options ) | (text-options )
specifies the color and font attributes of the category values that are
located outside of the tiles.
VALUELOCATION=AUTO | INSIDE | OUTSIDE
specifies the location of the category column values in a two-way plot.
XVALUEFITPOLICY=ROTATE | NONE
specifies a policy for avoiding collisions along the width of the plot
among category values that are outside of the tiles.
YVALUEFITPOLICY=NONE | ROTATEALWAYS
specifies a policy for avoiding collisions along the height of the plot
among category values that are outside of the tiles.

ODS options
URL=string-column
specifies the URL of an HTML page to display when a tile is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
CATEGORY=(column-list)
specifies a list of columns of category (classification) values.

Restriction No more than three columns can be specified.

 MOSAICPLOTPARM Statement 687

Note When a SAS data set is plotted, the category values are arranged in
data order. When a CAS in-memory table is plotted, the category
values are arranged in ascending order of the formatted character
values or the unformatted numeric values of the category column or
columns.

COUNT=non-negative-numeric-column | expression
specifies the column of counts (pre-summarized) for each of the category value
combinations.

Restriction The column values cannot be negative.

Tip You need to provide only the category crossings with nonzero
counts.

Optional Arguments
COLORGROUP=column | discrete-attr-var
specifies the category column to use for discrete fill colors for the tiles.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Each unique value of this column is mapped to the Color attribute of the
GraphData1–GraphDataN style elements that are in effect. If a discrete attribute
map variable is specified, the color mapping from its associated
DISCRETEATTRMAP statement is used.

Restriction This column or the associated column in the discrete attribute map
variable must be one of the columns in the category column list.

Interactions This option is ignored if the COLORRESPONSE= option is specified.

This option overrides the FILLATTRS= option.

Note If you specify a column in a SAS data set, colors are mapped to
group values in data order. If you specify a column in a CAS in-
memory table, colors are mapped to group values in ascending
order of the group column character values or unformatted numeric
values.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

688 Chapter 6 / Plot Statements

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies a numeric column to use to map tile fill colors to a continuous
gradient.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Interactions When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.
 MOSAICPLOTPARM Statement 689

When only the outlines are displayed, this option overrides

suboption COLOR= in the OUTLINEATTRS= option and varies the
outline color according to the color gradient or the attribute map.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATATRANSPARENCY=number
specifies the degree of the transparency of the tile fill, outlines, and the values
that are located inside the tiles if those values are displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

DISPLAY=(display-options ) | STANDARD | ALL

specifies which mosaic-plot features to display.

display-options
a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

FILL
displays filled tiles

OUTLINE
displays the tile outline

LABELS
displays the category column labels

TICKS
displays the category ticks

VALUES
displays the category values

STANDARD
specifies FILL, OUTLINE, LABELS, and VALUES

ALL
specifies all features: FILL, OUTLINE, LABELS, TICKS, and VALUES

Default STANDARD

Interactions If neither FILL nor OUTLINE are present in the display-options list,
then filled and outlined tiles are displayed.

If YVALUELOCATION=INSIDE or if YVALUELOCATION=AUTO and

is effectively set to INSIDE, then the axis ticks are not displayed
even if the display of the ticks is specified for the axis.
690 Chapter 6 / Plot Statements

FILLATTRS=style-element | (fill-options )
specifies the appearance of the tile fill areas. See “General Syntax for Attribute
Options” on page 1599 for the syntax on using a style element. See “Fill Color
Options” on page 1600 for the available fill options.

Default The GraphDataDefault style element

Interaction When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is ignored,
and the fill colors vary according to the gradient.

Note When style-element is specified, only the style element’s Color

attribute is used.

GUTTER=dimension | (dimension-list )
specifies the gutter (gap) between the splits. The splits occur in the following
way:
n When a single dimension is specified, the dimension applies to the gap for
the last split, which has the smallest gap. From the next-to-last last split to
the first split, the gutter is doubled on each split. A single dimension has the
effect of setting a minimum gap for the plot. The following figure shows an
example in which GUTTER=10.

Note: The plot contains equally sized tiles for demonstration purposes.

In this case, the gaps are 30 pixels for the first split, 20 pixels for the second
split, and 10 pixels for the third split (minimum gap).
n When a list of dimension values is used, the values apply to each split in the
order in which they are specified. The following figure shows an example in
which GUTTER=(10 20 30).
 MOSAICPLOTPARM Statement 691

In this case, the gaps are 10 pixels for the first split, 20 pixels for the second
split, and 30 pixels for the third split.

Default 3px. Dimensions with no units are assumed to be in pixels.

See “dimension” on page xiii

INSIDEVALUEATTRS=style-element | style-element (text-options ) | (text-

options )
specifies the color and font attributes of the category values when they are
located inside a tile. See “General Syntax for Attribute Options” on page 1599for
the syntax on using a style element. See “Text Options” on page 1606 for the
available text options.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the non-specified properties are derived from the GraphValueText
style element.

Note When style-element is specified, only the element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

Tip You can use the VALUEATTRS= option to change the text attributes
for the values that are located outside the plot area.

LABELATTRS=style-element | style-element (text-options ) | (text-options )

specifies the color and font attributes of the category labels. See “General
Syntax for Attribute Options” on page 1599for the syntax on using a style
element. See “Text Options” on page 1606 for the available text options.

Default The GraphLabelText style element.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the non-specified properties are derived from the GraphLabelText
style element.
692 Chapter 6 / Plot Statements

Note When style-element is specified, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the COLORGROUP= option is in effect, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | (line-options )
specifies the appearance of the tile outlines. See “General Syntax for Attribute
Options” on page 1599for the syntax on using a style element. See “Line Options”
on page 1602 for the available line options.

Default The GraphOutlines style element.

Interaction When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the outline colors vary according to the gradient.

Note When style-element is used, only the style element’s ContrastColor,

LineStyle, and LineThickness are used.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.
 MOSAICPLOTPARM Statement 693

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles. The predefined roles are CATEGORY1–
CATEGORYn (in the order in which they are specified in the
CATEGORY= option), COUNT, COLORGROUP, and
COLORRESPONSE.

SQUARED=TRUE | FALSE
specifies that a square aspect ratio be used for the plot area.

Default FALSE

Restriction This option applies to multi-way plots only.

Tip Setting this option to TRUE makes the height of the plot the same as
its width, which can make it easier to compare the proportions.

See “boolean ” on page xii for other Boolean values that you can use.

TIP=(role-list ) | NONE
specifies the information that is displayed when the cursor is positioned over a
tile. If this option is used, then all of the information that is displayed by default
is replaced. Roles for columns that do not contribute to the bar chart can be
specified along with roles that do.

(role-list)
an ordered, space-separated list of unique MOSAICPLOTPARM roles and
user-defined roles.

The MOSAICPLOTPARM roles include COUNT, COLORGROUP, and

COLORRESPONSE. User-defined roles are defined with the ROLENAME=
option.

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns that are assigned to the category columns and
COUNT roles are automatically included in the data tip
information.
694 Chapter 6 / Plot Statements

Requirement To generate data tips, you must include an ODS GRAPHICS ON

statement with the IMAGEMAP option specified. You must also
write the graphs to the ODS HTML destination.

Interaction The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

Example To display data tips for the columns that are assigned to the roles
X and Y as well as the user-defined role TIP1:
ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

Note This option is no longer supported starting with SAS 9.4M4. To

format tip values in SAS 9.4M4 and in later releases, use a
FORMAT statement in the SGRENDER procedure step. See
“FORMAT Statement” in SAS ODS Graphics: Procedures Guide.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)
 MOSAICPLOTPARM Statement 695

URL=string-column
specifies the URL of an HTML page to display when a tile is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each tile that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable tiles, you must include an ODS GRAPHICS

ON statement that specifies the IMAGEMAP option. You must also
write the graphs to the ODS HTML destination.

Tips The URL value can be blank for some tiles, which means that no
action is taken when those tiles are selected. The URL value can be
the same for different tiles, which means that the same action is
taken when those tiles are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

VALUEATTRS=style-element | style-element (text-options ) | (text-options )

specifies the color and font attributes of the category values that are located
outside of the tiles. See “General Syntax for Attribute Options” on page 1599for
the syntax on using a style element. See “Text Options” on page 1606 for the
available text options.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
the font properties such as color, family, size, weight, and style, then
the non-specified properties are derived from the GraphValueText
style element.

Note When style-element is used, only the style element’s COLOR,

FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT
attributes are used.

Tip You can use the INSIDEVALUEATTRS= option to change text

attributes for values that are located inside the tiles.

VALUELOCATION=AUTO | INSIDE | OUTSIDE

specifies the location of the category column values in a two-way plot.

AUTO
locates the values for the second category in the CATEGORY= list that are
outside of the tiles.

If the first category in the CATEGORY= list has any missing crossings with
the second category, or if any of the second category values collide, then the
values are located inside the tiles per the default fit policy.
696 Chapter 6 / Plot Statements

See “CATEGORY=(column-list)” on page 686

INSIDE
for each category value, locates the values inside the largest tile.

OUTSIDE
locates the values outside of the tiles, in the plot area.

Default AUTO

Restriction This option applies to two-way plots only.

XVALUEFITPOLICY=ROTATE | NONE
specifies a policy for avoiding collisions along the width of the plot among
category values that are outside of the tiles.

ROTATE
rotates the values if any of the values collide.

NONE
does not attempt to fit values that collide.

Default ROTATE

YVALUEFITPOLICY=NONE | ROTATEALWAYS
specifies a policy for avoiding collisions along the height of the plot among
category values that are outside of the tiles.

NONE
does not attempt to fit values that collide.

ROTATEALWAYS
rotates the values regardless of whether any of the values collide.

Default NONE

Interaction This option is effective only when VALUELOCATION=OUTSIDE.

Details
A mosaic plot displays relative frequencies for categorical variables. Each crossing
of the categorical values is represented by a tile. The area of each tile is
proportional to the frequency of that crossing. The plot is the result of an iterative
process. The first iteration splits the plot area into tiles along the width according
to the relative frequency of the first category column values. Subsequent iterations
split the tiles from the previous iteration in the direction orthogonal to the previous
split by using the relative frequencies of each category column's values. By default,
the gap (or gutter) for each split gets progressively smaller, with a minimum gap of
3 pixels. You can use the GUTTER= option to specify a different gap for each split.

The following figure provides an example of a three-way mosaic plot, which has
three categories.
 MOSAICPLOTPARM Statement 697

Note: The plot contains equally sized tiles for demonstration purposes.

In the example plot, the first split is along the width for CATEGORY 1. The second
split is along the height for CATEGORY 2. Finally, the third split is along the width
for CATEGORY 3. Notice how the gaps between the tiles get progressively smaller
from the first split to the last split.

Example: MOSAICPLOTPARM Statement

The following graph was generated by the “Example Program” on page 698:
 698 Chapter 6 / Plot Statements

Example Program
/* Summarize the SASHELP.CARS data for ORIGIN and TYPE */
proc summary data=sashelp.cars nway;
class origin type;
var mpg_highway;
output out=mileage mean=avgMpg N=count / noinherit;
run;

/* Generate the plot */

proc template;
define statgraph mosaicPlotParm;
begingraph;
layout region;
mosaicPlotParm category=(type origin) count=count /
name="mosaic" colorresponse=avgMpg;
continuouslegend "mosaic" / title="Average Miles Per Gallon"
pad=(left=5);
endlayout;
endgraph;
end;
run;

proc sgrender data=mileage template=mosaicPlotParm;

run;
 NEEDLEPLOT Statement 699

NEEDLEPLOT Statement
Creates a plot of observations as points connected to a baseline by vertical line segments.

Syntax
NEEDLEPLOT X=column | expression
Y=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint
spacing or bin width.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the needle plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the needle lines, markers, and
data labels, if displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which needle features to display.
DISPLAYBASELINE=ON | OFF | AUTO
specifies whether the baseline is displayed.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and
line pattern) to one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the needle lines for the data points.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the data markers.

Axes options
BASELINEINTERCEPT=number | RELATIVE
specifies the Y-intercept for the baseline.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
700 Chapter 6 / Plot Statements

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
needle line or marker.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column
specifies labels at the data points.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end of the needle
lines and markers, if displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all needle lines and markers from discrete X
values when graphing multiple response variables side by side on a
common axis.
GROUP=column | discrete-attr-var | expression
 NEEDLEPLOT Statement 701

creates a distinct set of needle lines, markers, and data labels for each
unique group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped needle lines are overlaid or clustered around
the category midpoints on a discrete axis or around the intervals on an
interval axis.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a needle or marker is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies a column or expression for the X values.

Y=numeric-column | expression
specifies a numeric column or numeric expression for the Y values.

Optional Arguments
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The COLOR= option in the LINEATTRS= option, if specified. Otherwise,

the GraphDataDefault:ContrastColor style reference.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s Color,

LineStyle, and LineThickness attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

BASELINEINTERCEPT=number | RELATIVE
specifies the Y-intercept for the baseline. Prior to SAS 9.4M5, the baseline is
always displayed in the chart, whether for a specified value or for the default
702 Chapter 6 / Plot Statements

value. Starting with SAS 9.4M5, the baseline display is controlled by the
DISPLAYBASELINE= option, which is ON by default.

number
specifies the Y-intercept value to use for the baseline.

Interactions When number is specified, the response axis data range is

extended if necessary to include the baseline intercept. When a
logarithmic response axis is requested and number is 0 or a
negative value, the response axis reverts to a linear axis. To
restore the log axis in that case, set BASELINEINTERCEPT= to a
positive value.

When DISPLAYBASELINE=AUTO is in effect, the baseline is not

displayed if the baseline intercept is equal to the minimum or
maximum value of the response-axis range.

Tips The baseline does not add a tick or a tick value to the axis. To
label the baseline, you can overlay a REFERENCELINE
statement with the same Y value and use its CURVELABEL
option.

The appearance of the baseline is controlled by the

BASELINEATTRS= option.

To suppress the baseline, prior to SAS 9.4M5, use the

BASELINEATTRS= option to set the line thickness to 0. Starting
with SAS 9.4M5, specify DISPLAYBASELINE=OFF.

RELATIVE
places the baseline at the Y-axis tick mark closest to the minimum of the
range for the needle data points.

Default 0

CLUSTERWIDTH=number
specifies the width of the group clusters as a fraction of the midpoint spacing or
bin width.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Requirement For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

Interaction When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in effect,
the size of the markers in each cluster might be reduced to no less
than 5 pixels in order to display the cluster within the smallest
effective midpoint space. If you need larger markers in that case,
use the MARKERATTRS= option to specify a larger marker size.
 NEEDLEPLOT Statement 703

DATALABEL=column
specifies labels at the data points.

Default No data labels are displayed

Note The position of the labels is adjusted to prevent the labels from
overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to have any effect, the DATALABEL= option must
also be specified.

The LINEATTRS= on page 712 COLOR= suboption overrides the

default data label color for grouped plots.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to
avoid overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end of the needle lines
and markers, if displayed.

AUTO
uses a data-label collision avoidance algorithm to position the labels.

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.
704 Chapter 6 / Plot Statements

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters. When
set to TRUE, the data label is split unconditionally at each occurrence of any of
the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

 NEEDLEPLOT Statement 705

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.
706 Chapter 6 / Plot Statements

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the needle plot lines. The following figure
shows a needle plot with each of the skins applied.
 NEEDLEPLOT Statement 707

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data skin
is not applied to that plot. In that case, use the DATASKINMAX=
option in your ODS GRAPHICS statement to increase the maximum
limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the needle lines, markers, and data
labels, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEOFFSET=number
specifies an amount to offset all needle lines and markers from discrete X
values when graphing multiple response variables side by side on a common
axis.

Default 0 (no offset, all needle lines and markers are centered on the
discrete X values)

Range -0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right. If the parent layout’s
axis options set REVERSE=TRUE, then the offset direction is also
reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.
708 Chapter 6 / Plot Statements

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=(display-options) | STANDARD | ALL

specifies which needle features to display.

(display-options)
a space-separated list of one or more options enclosed in parentheses.

Currently, only the MARKERS option is supported, which displays needle

lines with markers.

STANDARD
displays needle lines without markers

ALL
displays needle lines with markers

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

DISPLAYBASELINE=ON | OFF | AUTO

specifies whether the baseline is displayed.

Note: This feature applies to SAS 9.4M5 and to later releases.

ON
always displays the baseline.

OFF
does not display the baseline.

AUTO
displays the baseline if the baseline intercept is within the response-axis
range, excluding the minimum and maximum axis values.

This is the typical case when the needle plot contains both positive and
negative needles, and the default baseline intercept of 0 is used. If the
baseline intercept is equal to the minimum or maximum value of the
response-axis range, the baseline is not displayed.

Default ON

Tip By default, the baseline intercept is 0. Use the BASELINEINTERCEPT=

option to change the baseline intercept.

GROUP=column | discrete-attr-var | expression

creates a distinct set of needle lines, markers, and data labels for each unique
group value of the specified column.
 NEEDLEPLOT Statement 709

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol.
These vary according to the ContrastColor, LineStyle, and
MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Notes The representations that are used to identify the groups can be
overridden. For example, each distinct group value might be
represented by a different line pattern, but the
LINEATTRS=(PATTERN=pattern) option could be used to assign
the same line pattern to all of the plot’s line patterns, letting line
color indicate group values. Likewise, LINEATTRS=(COLOR=color)
could be used to assign the same color to all lines, letting line
pattern indicate group values.

If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tip By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped needle lines are overlaid or clustered around the
category midpoints on a discrete axis or around the intervals on an interval axis.

OVERLAY
centers the needle lines for matching category values on the midpoints.

The needle lines in each set of group values are superimposed on each other.
710 Chapter 6 / Plot Statements

CLUSTER
clusters the needle lines for matching category values around the midpoints.

Each cluster of group values is centered at the midpoint for the category.

Default OVERLAY

Interaction For this option to take effect, the GROUP= option must also be
specified.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
 NEEDLEPLOT Statement 711

formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
712 Chapter 6 / Plot Statements

then a modulo operation remaps that index value to a number

less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the needle lines for the data points.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the LineThickness
attribute of the GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions This option’s COLOR= suboption overrides the default behavior for
grouped data. When the COLOR= suboption is specified in that
case, all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data. When the SYMBOL= suboption is specified in
that case, all markers have the same symbol, and the symbol color
alone distinguishes the markers.
 NEEDLEPLOT Statement 713

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display

of the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current
ODS style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.
714 Chapter 6 / Plot Statements

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X. Y, DATALABEL, INDEX, and GROUP.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a needle
line or marker. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the needle
plot can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique NEEDLEPLOT and user-defined
roles.

NEEDLEPLOT roles include X, Y, DATALABEL, and GROUP.

User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned
to the roles X and Y as well as the column Obs, which is not
assigned to any pre-defined NEEDLEPLOT role. The Obs column
must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X, Y, DATALABEL, and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.
 NEEDLEPLOT Statement 715

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a needle or marker is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each needle that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable needle lines, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.
716 Chapter 6 / Plot Statements

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding needle or marker is
selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the needle or marker is selected for
those X and Y pairs.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
In the NEEDLEPLOT statement, the X column can specify character or numeric
values. The Y column must specify numeric values. For character columns, the X
axis is always of TYPE=DISCRETE. For numeric columns, the X axis is of
TYPE=LINEAR by default.

The Y axis is of TYPE=LINEAR by default.

Example: NEEDLEPLOT Statement

The following graph was generated by the “Example Program” on page 717:
 PBSPLINEPLOT Statement 717

Example Program
proc template;
define statgraph needleplot;
begingraph;
entrytitle "IBM Stock Trend";
layout overlay;
needleplot x=date y=close /
baselineintercept=80 lineattrs=(color=blue);
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.stocks template=needleplot;

where stock="IBM" and date > "31dec1999"d;
run;

PBSPLINEPLOT Statement
Creates a fitted penalized B-spline curve computed from input data.

Restriction: The PBSPLINEPLOT statement supports only models of one independent and one
dependent variable.
Tip: Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
718 Chapter 6 / Plot Statements

ODS GRAPHICS statement. For information about the BEGINGRAPH statement

SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.

Syntax
PBSPLINEPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-options> <options>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the curve and curve label.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes of the spline curve.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the spline curve.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the spline curve labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the spline curve label relative to the plot area.
 PBSPLINEPLOT Statement 719

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the spline curve label relative to the curve line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if
needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the curve label
block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of curves from just the observations that
correspond to each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.

Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the spline curve.

Default No curve label is displayed

Interaction This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the spline curve labels.
720 Chapter 6 / Plot Statements

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the spline curve label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the PBSPLINEPLOT is used in

multicell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the curve
labels appear.

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the spline curve label relative to the curve line.

AUTO
automatically positions the spline curve label near the curve boundary along
unused axes whenever possible (typically Y2 and X2) in order to avoid
collision with tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the spline curve label to appear near maximum curve values
(typically, upper right)

MIN
forces the spline curve label to appear near minimum curve values (typically,
lower left)

START
forces the spline curve label to appear near the beginning of the curve.
 PBSPLINEPLOT Statement 721

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

END
forces the spline curve label to appear near the end of the curve.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the curve line has a spiral
shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the spline
curve label appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the spline curve label might fall outside of the graph
area. In that case, the spline curve label might not be displayed or
might be positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the curve label at the specified split characters. When
a curve label is split, the label is split on each occurrence of the specified split
characters.

Default FALSE. The curve label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.
722 Chapter 6 / Plot Statements

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the curve label can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the curve label. In that case, all of the specified split characters
together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the curve

label is split unconditionally at each occurrence of any of the specified split
characters. If the curve label does not contain any of the specified characters,
then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the curve label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the curve label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the curve label text.

TRUE
drops the split characters from the curve label text.

FALSE
includes the split characters in the curve label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.
 PBSPLINEPLOT Statement 723

The following figure shows an example of a curve label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the curve for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the curve label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the curve label block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.
724 Chapter 6 / Plot Statements

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the curve for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATATRANSPARENCY=number
specifies the degree of the transparency of the curve and curve label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of curves from just the observations that correspond to
each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Line colors vary
according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–
GraphDataN and GraphMissing style elements.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned
line cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000.
If the limit is exceeded, the plot is not drawn and a note is written
to the SAS log stating that the limit has been exceeded. In that
 PBSPLINEPLOT Statement 725

case, increase LINEPATTERNOBSMAX= to the value suggested in

the note or change the line pattern to SOLID.

Requirement You must sort the data in ascending order by the specified group
column before you plot the data. If the plot data is a CAS in-
memory table, you must download the in-memory table to your
SAS client, and then sort it locally before you plot it.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors
and line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a
column in a CAS in-memory table, the visual attributes for each
group value are assigned in ascending order of the group column
character values or of unformatted numeric values.

Tips The LINEATTRS= option can be used to override the

representations that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the loess curves, letting the line color distinguish
group values. Likewise, LINEATTRS=(COLOR=BLACK) can be used
to assign the same color to all of the curves, letting the line pattern
distinguish group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use
the GROUPMAX= option in the ODS GRAPHICS statement to
change the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.
726 Chapter 6 / Plot Statements

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes of the spline curve.

Default The GraphFit style element.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned line
cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000. If
the limit is exceeded, the plot is not drawn and a note is written to
the SAS log stating that the limit has been exceeded. In that case,
increase LINEPATTERNOBSMAX= to the value suggested in the
note or change the line pattern to SOLID.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.
 PBSPLINEPLOT Statement 727

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.
728 Chapter 6 / Plot Statements

Note The columns assigned to the X, Y, and GROUP (if assigned) roles
are automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

PBSPLINE Regression Options

ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < positive-number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLI="name"
produces confidence limits for individual predicted values for each observation.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the
confidence limits to be computed. To display confidence limits, you
 PBSPLINEPLOT Statement 729

must use this name as the required argument of a MODELBAND

statement. See the example in the section “Example:
PBSPLINEPLOT Statement” on page 731 .

CLM="name"
produces confidence limits for a mean predicted value for each observation.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the
confidence limits to be computed. To display confidence limits, you
must use this name as the required argument of a MODELBAND
statement. See the example in the section “Example:
PBSPLINEPLOT Statement” on page 731 .

DEGREE=non-negative-integer
specifies the degree of B-spline.

Default 3

Ranges 0–174 in SAS 9.4M1 and earlier releases.

0–10 starting with SAS 9.4M2.

Restriction Starting with SAS 9.4M2, DEGREE= and NKNOTS= cannot be set to
0 simultaneously. When both are set to 0, an error results.

FREQ=numeric-column
specifies a column in the input data set that represents the frequency of
occurrence of the current observation, essentially treating the data set as if
each observation appeared n times, where n is the value of the FREQ column for
the observation. Noninteger values of the FREQ column are truncated to the
largest integer less than the FREQ value. The observation is used in the analysis
only if the value of the FREQ column is greater than or equal to 1.

MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the spline
curve as well as any confidence limits.

Default 201

NKNOTS=non-negative-integer
specifies the number of evenly spaced internal knots. By default, a large number
of knots (100) is specified, which allows for an extreme lack of smoothness in
the results. However, the final function is typically much smoother due to the
penalty. When SMOOTH=0 is specified, you should typically ask for many fewer
knots than the default, since there is no penalty for lack of smoothness. For
example, ten or fewer knots is usually enough to follow the functional form
found in most data.

Default 100

Restriction Starting with SAS 9.4M2, MKNOTS= and DEGREE= cannot be set to
0 simultaneously. When both are set to 0, an error results.
730 Chapter 6 / Plot Statements

See example “Penalized B-Splines” in the TRANSREG procedure

description in SAS/STAT User's Guide.

SMOOTH=AUTO | non-negative-number
specifies a smoothing parameter value.

AUTO
automatically selects a smoothing parameter value that minimizes a lack-of-
smoothness penalty.

non-negative-number
specifies a smoothing parameter value.

Range [0,1]

Tip Specify SMOOTH=0 for an ordinary B-spline fit.

Default AUTO

See The TRANSREG Procedure in SAS/STAT User’s Guide for information

about the fitting methodology that the PBSPLINEPLOT statement
uses.

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a
priori weights for a penalized B-spline fit. If an observation’s weight is zero,
negative, or missing, then the observation is deleted from the analysis.

Interaction Starting with SAS 9.4M2, when the CLI= option is used with this
option, the confidence band for individual predicted values is
displayed as a high-low chart instead of a band.

Details
The PBSPLINEPLOT statement supports only models of one independent and one
dependent variable. For more information about the fitting methodology, see The
TRANSREG Procedure in SAS/STAT User’s Guide.

In addition to the penalized B-spline, the PBSPLINEPLOT statement can compute

confidence levels for the fitted line. To display the confidence levels,

1 use the CLI= or CLM= option to declare a name for the confidence level

2 use a MODELBAND statement to refer to this name. This statement draws a

confidence band from this information. See “MODELBAND Statement” on page
673 for information about how to control the appearance of the confidence
band.
 PBSPLINEPLOT Statement 731

Example: PBSPLINEPLOT Statement

The following graph was generated by the “Example Program” on page 731:

Example Program
proc template;
define statgraph pbsplineplot;
begingraph;
entrytitle "Spline Fit";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
pbsplineplot x=weight y=mpg_highway / name="fitline"
alpha=0.05 legendlabel="Spline Fit";
discretelegend "fitline";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=pbsplineplot;

run;
 732 Chapter 6 / Plot Statements

PIECHART Statement
Creates a pie chart that is computed from input data.

Requirement: The PIECHART statement must be placed in a LAYOUT REGION, LAYOUT

GRIDDED, or LAYOUT LATTICE block. It cannot be placed in an overlay-type layout
such as LAYOUT OVERLAY or LAYOUT OVERLAYEQUATED because a pie chart
does not have axes.
Note: The PIECHART statement does not honor the ODS GRAPHICS options
DISCRETEMAX=, GROUPMAX=, and LABELMAX=.

Syntax
PIECHART CATEGORY=column | discrete-attr-var | expression </options>;

PIECHART CATEGORY=column | discrete-attr-var | expression

RESPONSE=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
CATEGORYDIRECTION=COUNTERCLOCKWISE | CLOCKWISE
specifies whether to display the pie slices in counterclockwise or
clockwise sequence.
CENTERFIRSTSLICE=TRUE | FALSE
specifies whether the first pie slice is centered on the starting angle or
starts on the starting angle.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled pie slices.
DATATRANSPARENCY=number
specifies the degree of the transparency of all pie slices, outlines, and
text.
DISPLAY=(display-options) | STANDARD
specifies which pie-slice options to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled pie slices.
OTHERSLICEOPTS=(other-slice-options)
specifies the properties of the Other slice.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the properties of the pie and slice outlines.
START=degrees
 PIECHART Statement 733

specifies which degree between 0 and 360 serves as the starting position
for the first pie slice.

Data tip options

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
pie slice.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Grouping options
GROUP=column | expression
creates a separate concentric annulus (or stacked cylinders) for each
unique group value of the specified column.
GROUPGAP=dimension
specifies a dimension for the optional gap that can be displayed between
each annulus of a grouped pie.
GROUPLABELOPTS=(grouplabel-options)
specifies text attributes, location, and other options for displaying group
labels.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the Group column are included in the
pie.
OTHERSLICE=TRUE | FALSE
specifies whether to consolidate smaller pie slices into a single slice that
represents “other” values that are in the data, or whether to display those
smaller slices as separate pie slices.

Label options
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the slice labels.
DATALABELCONTENT=ALL | STANDARD | NONE | (content-options)
specifies the information to display in the slice labels.
DATALABELLOCATION=AUTO | INSIDE | OUTSIDE | CALLOUT
specifies whether to display the slice labels within the pie slices or
outside of the pie circumference.
LABELFITPOLICY=NONE | DROP
specifies the label fitting policy to be used if a particular label does not
fit within the pie slice.

ODS options
URL=string-column
specifies an HTML page to display when a pie slice is selected.

Plot reference options

NAME="string"
734 Chapter 6 / Plot Statements

assigns a name to this plot statement for reference in other template

statements.

Statistics options
STAT=FREQ | PCT | SUM | MEAN
specifies the statistic to be computed.

Required Argument
CATEGORY=column | discrete-attr-var | expression
specifies the column for the category values. Duplicated values of CATEGORY
are summarized into a unique value. All values are treated as discrete.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

See “DISCRETEATTRVAR Statement” on page 1516

Optional Arguments
CATEGORYDIRECTION=COUNTERCLOCKWISE | CLOCKWISE
specifies whether to display the pie slices in counterclockwise or clockwise
sequence.

Default COUNTERCLOCKWISE

Tip The START= option controls the starting angle for the first pie slice.

CENTERFIRSTSLICE=TRUE | FALSE
specifies whether the first pie slice is centered on the starting angle or starts on
the starting angle.

Note: This option is valid starting in SAS 9.4M1.

The following figure shows the effect of this option on a pie chart in which Asia
is the first category slice, the starting angle is 0 degrees, and the category
direction is counterclockwise.
 PIECHART Statement 735

Default FALSE

Tips Use the START= option to change the starting angle.

Use the CATEGORYDIRECTION= option to change the category

direction.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the slice labels.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELCONTENT=ALL | STANDARD | NONE | (content-options)

specifies the information to display in the slice labels.

ALL
displays all available information

STANDARD
equivalent to specifying the two content-options CATEGORY and
RESPONSE

NONE
does not display slice labels

(content-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

CATEGORY
displays the CATEGORY value

PERCENT
displays the PERCENT value when STAT=FREQ or STAT=PCT. When
STAT=MEAN or STAT=SUM, it displays nothing
736 Chapter 6 / Plot Statements

RESPONSE
displays the statistic that is requested in the STAT= option.

Defaults When STAT=PCT, the default is (CATEGORY PERCENT).

Otherwise, the default is STANDARD.

Note The position of the labels is adjusted to prevent the labels from
overlapping.

DATALABELLOCATION=AUTO | INSIDE | OUTSIDE | CALLOUT

specifies whether to display the slice labels within the pie slices or outside of
the pie circumference.

AUTO
automatically selects either INSIDE, OUTSIDE, or CALLOUT to optimize the
label position

INSIDE
locates the slice labels inside the pie slices.

Note: If a particular label does not fit within the pie slice, then the fit policy
takes effect (set by the LABELFITPOLICY= option).

OUTSIDE
locates the slice labels outside of the pie circumference.

CALLOUT
locates the slice labels outside of the pie circumference and draws a line
from the label to its slice.

Default AUTO

Note When OUTSIDE or CALLOUT is specified and there is not enough space
to display one or more of the labels, the labels are suppressed.
 PIECHART Statement 737

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled pie slices. The following figure
shows pies with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, DISPLAY= FILL must be in
effect. Otherwise, this option is ignored.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

When a data skin is applied, all slice outlines are set by the skin,
and the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of all pie slices, outlines, and text.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the pie
slices. The OTHERSLICEOPTS= option can be used to specify
transparency for the “other” slice. You can combine this option with
FILLATTRS= and with OTHERSLICEOPTS= to set one transparency for
738 Chapter 6 / Plot Statements

the outlines and text but a different transparency for the pie slices.
Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

DISPLAY=(display-options) | STANDARD
specifies which pie-slice options to display.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

OUTLINE
displays the pie-slice outline.

The default outline properties are set by the GraphOutline style element.

FILL
displays the pie-slice fill color.

The default colors are set by the Color attribute of the GraphData1–
GraphDataN style elements. The fill color of the “other” slice (if shown) is
from the Color attribute of the GraphOther style element. If FILL is not
specified, then an opaque pie is drawn using the background color of the
containing layout.

STANDARD
specifies OUTLINE and FILL

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the pie slices.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled pie slices. Prior to SAS 9.4M3, this option
specifies the color and transparency of all of the pie slices, excluding the Other
slice. Starting with SAS 9.4M3, this option specifies the transparency for the
Other slice as well.

Default The GraphDataDefault:Color style reference.

Interaction For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Tips The FILLATTRS= suboption of the OTHERSLICEOPTS= option

specifies the color and transparency of the Other slice.

The DATATRANSPARENCY= option sets the transparency for all pie

slices, outlines, and text. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
and text but a different transparency for the pie slices. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
 PIECHART Statement 739

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element value.

“Fill Color Options” on page 1600 for available fill-options values.

GROUP=column | expression
creates a separate concentric annulus (or stacked cylinders) for each unique
group value of the specified column. The grouped rings are displayed in data
order.

Restriction The PIECHART GROUP= option does not honor the ODS
GRAPHICS statement GROUPMAX= option.

Interactions When this option is used, the unique column values are found and
then the slice colors are taken from the GraphData1–GraphDataN
style elements.

Missing values in the data can affect the group order. You can use
the INCLUDEMISSINGGROUP= option to manage missing group
values. In addition, you can use
INCLUDEMISSINGDISCRETE=TRUE in the BEGINGRAPH
statement to create pie slices for missing CATEGORY values.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips This option creates only rings of pies. To create a grid of pies,
specify the PIECHART statement within a LAYOUT LATTICE,
LAYOUT DATALATTICE, or LAYOUT DATAPANEL statements.
740 Chapter 6 / Plot Statements

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See FILLATTRS=

GROUPGAP=

GROUPLABELOPTS=

GROUPGAP=dimension
specifies a dimension for the optional gap that can be displayed between each
annulus of a grouped pie.

Default 0

Restriction For this option to take effect, the GROUP= option must also be
specified.

Interaction If the specified dimension is too large for the area that is available to
the pie chart, then the results might be unexpected.

Note The size of the inner pie remains the same regardless of the
GROUPGAP= value.

See “dimension” on page xiii

GROUPLABELOPTS=(grouplabel-options)
specifies text attributes, location, and other options for displaying group labels.
The following grouplabel-options are available. One or more options can be
specified as space-separated name = value pairs.

LABEL=AUTO | NONE | "string"

specifies a descriptive label for the Group column

AUTO
specifies the column label of the GROUP= column or the column name of
the GROUP= column, if no column label exists.

NONE
specifies that no label is displayed

"string"
specifies a string to use as the label

Default AUTO

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the group label.

Default The GraphLabelText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.
 PIECHART Statement 741

“Text Options” on page 1606 for available text-options.

LOCATION=RIGHT | LEFT
specifies whether the block of text for group labeling appears to the right or
left of the pie.

Default RIGHT

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the group values.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

Default The group label and values are shown as a block of text to the right
or left of the pie. Slice labels are moved to the inside of the pie
slices. A line is drawn from each group value to its annulus (or
cylinder).

Restriction For this option to take effect, the GROUP= option must also be
specified.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the Group column are included in the pie.

Default TRUE

Restriction For this option to take effect, the GROUP= option must also be
specified.

See “boolean ” on page xii for other Boolean values that you can use.

LABELFITPOLICY=NONE | DROP
specifies the label fitting policy to be used if a particular label does not fit
within the pie slice.

NONE
draws each label regardless of whether it fits within the slice region.

DROP
drops labels that do not fit within the slice region, but draws labels that do
fit.

Default NONE

Tip This option determines how labels are managed when

DATALABELLOCATION= INSIDE and a particular label does not fit
within the pie slice.
742 Chapter 6 / Plot Statements

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

OTHERSLICE=TRUE | FALSE
specifies whether to consolidate smaller pie slices into a single slice that
represents “other” values that are in the data, or whether to display those
smaller slices as separate pie slices. If this option is set to FALSE, then all
unique category values appear as slices. If this option is set to TRUE, then some
of the smaller slices might be combined into a single slice, referred to as the
Other slice.

Default TRUE

Tip To set the properties of the “other” slice, use the OTHERSLICEOPTS=
option.

See “boolean ” on page xii for other Boolean values that you can use.

OTHERSLICEOPTS=(other-slice-options)
specifies the properties of the Other slice. Example:
piechart category=region / name="p"
datalabelcontent=(percent) datalabellocation=inside
otherslice=true
othersliceopts=(type=percent percent=11 label="Other Regions") ;
 PIECHART Statement 743

The following other-slice-options values are available. You can specify one or
more options as space-separated name = value pairs.

TYPE=PERCENT | MAXSLICES
specifies which method to use to determine the size of the Other slice.

PERCENT
uses the percentage that is set by the PERCENT= suboption.

MAXSLICES
uses the count that is set by the MAXSLICES= suboption.

Default PERCENT

MAXSLICES=positive-integer
specifies the maximum number of category values to represent with pie
slices.

Any remaining values are consolidated into the Other slice.

Default 10

Interactions For this option to have any effect, TYPE=MAXSLICES must also
be specified among the suboptions for OTHERSLICEOPTS=.

The slices are counted in the order in which they are displayed.
This order is affected by the CATEGORYDIRECTION= option.

PERCENT=percent-of-total
collects all category values with response values less than or equal to the
specified percent-of-total value into the Other slice.

Default 4. Any original slice that represents 4% or less of the total is put
in the Other category.

Range 0–100

Interaction For this option to have any effect, TYPE=PERCENT must also be
specified among the suboptions for OTHERSLICEOPTS=.

LABEL="string"
specifies a label for the Other slice.

Default "OTHER"

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the area fill for the Other slice.

This option does not affect the appearance of the area fill for the remaining
slices.

Default The GraphOther style element.

Interactions Suboption TRANSPARENCY= in this option overrides the

DATATRANSPARENCY= option for the other slice only.
744 Chapter 6 / Plot Statements

Starting with SAS 9.4M3, suboption TRANSPARENCY= in this

option overrides suboption TRANSPARENCY= in option
FILLATTRS= in the PIECHART statement only for the Other
slice.

See “General Syntax for Attribute Options” on page 1599 for the
syntax for using a style-element value.

“Fill Color Options” on page 1600 for available fill-options

values.

Interaction This option is ignored if OTHERSLICE=FALSE.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the properties of the pie and slice outlines.

Default The GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

START=degrees
specifies which degree between 0 and 360 serves as the starting position for
the first pie slice. A value of 0 degrees corresponds to the three o'clock position.
Degrees can be either positive or negative. Positive values move the starting
position counterclockwise, and negative values move the starting position
clockwise. From the starting point specified by this option, the slices are drawn
in the direction specified by the CATEGORYDIRECTION= option.

Default 0

Range 0 to 360

STAT=FREQ | PCT | SUM | MEAN

specifies the statistic to be computed. For pie charts with no RESPONSE=
column:

FREQ frequency count

PCT percentages between 0 and 100

For pie charts with a RESPONSE= column:

SUM
MEAN

Defaults SUM for pie charts that specify the RESPONSE= argument
 PIECHART Statement 745

FREQ for pie charts that do not specify the RESPONSE= argument

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a pie
slice. If this option is used, then it replaces all of the information that is
displayed by default.

(role-list)
an ordered, space-separated list of unique PIECHART roles.

PIECHART roles include CATEGORY, RESPONSE, and GROUP. The

RESPONSE role represents the computed statistic for the CATEGORY value,
based on the statistic that is set by the STAT= option.

Example The following example displays data tips for the columns assigned
to the roles CATEGORY and RESPONSE.
TIP=(CATEGORY RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: CATEGORY and RESPONSE .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.
746 Chapter 6 / Plot Statements

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when a pie slice is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each pie slice that is to have an active link.

Example http://www.sas.com

Restriction A generated Other slice does not have a URL. See OTHERSLICE=.

Requirement To generate a plot with selectable pie slices, you must include an
ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Interaction This option has no effect when TIP=NONE.

Tips The URL value can be blank for some pie slices, meaning that no
action is taken when the corresponding slice is selected.

The URL value can be the same for any CATEGORY and
RESPONSE pairs. In that case, the same action is taken when the
pie slices for those pairs are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

Optional Response Argument

RESPONSE=numeric-column | expression
specifies response values that are read from a numeric column or an expression.

Details
The input data for the PIECHART statement is raw, unsummarized input data. The
PIECHART performs discrete binning for the Category column and calculates
 PIECHART Statement 747

appropriate summarization statistics (sum, mean, and so on) based on the setting
for the STAT= option.

The fill color of each pie slice is derived from the Color attribute of the
GraphData1–GraphDataN style elements as described in “Attribute Rotation
Patterns” in SAS Graph Template Language: User’s Guide. The default order of the
pie slices depends on the data type of the CATEGORY values:
n For numeric data, the slices appear in the ascending order of the unformatted
data values.
n For discrete data, the slices appear in data order.

You can use the START= and CATEGORYDIRECTION= options to control the pie
slice positions and display order.

By default, the pie slices are labeled with the CATEGORY and RESPONSE values,
which are displayed inside the slices. You can use the DATALABELLOCATION= and
DATALABELCONTENT= options to control where the pie slices are labeled and the
label content.

By default, if two or more slices take up less than 4% of a pie, then an “other” slice
is created by consolidating those small slices. To change the default criteria, use
the OTHERSLICE= and OTHERSLICEOPTS= options. The calculated “other” slice is
displayed as the last slice in the pie, and as the last legend entry for the pie. If a
category value is the same as the “other” slice label, then two slices might be
displayed with the same label ("Other" by default) and different fill attributes. In
that case, both slices are represented in the pie legend.

To create a pie slice for missing CATEGORY values, specify

INCLUDEMISSINGDISCRETE=TRUE in the BEGINGRAPH statement. The fill color
of the missing category slice is assigned the fill color from the GraphMissing style
element except when a user-defined format is applied to the category value. In that
case, the missing category slice is assigned the fill color from a GraphData1–
GraphdDataN style element in data order instead.

Note: The PIECHART statement does not honor the MISSING= system option.
Regardless of the MISSING= system option value, unless a user-defined format is
applied to the value, the default missing-numeric-value character (.) is used to
depict missing numeric values.

Example: PIECHART Statement

The following graph was generated by the “Example Program” on page 748 :
 748 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph simplepie;
begingraph;
entrytitle "Car Models by Origin";
layout region;
piechart category=origin / datalabellocation=outside;
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.cars
template=simplepie;
run;

POLYGONPLOT Statement
Draws a polygon from data that is stored in a data set.

Restriction: The POLYGONPLOT statement does not support CAS in-memory tables. If the
polygon plot data is stored in a CAS table, the polygon plot is not drawn.
Note: This statement is valid starting in SAS 9.4M1.
Tip: Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
 POLYGONPLOT Statement 749

BEGINGRAPH statement or in an ODS GRAPHICS statement. For information

about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.

Syntax
POLYGONPLOT X=column | expression Y=column | expression
ID=column | expression< /options(s) >

Summary of Optional Arguments

Appearance options
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.
BACKLIGHT=number | AUTO
specifies a back-light effect for the polygon label text.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variable to use to
determine the polygon colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of filled polygons.
DATATRANSPARENCY=number
specifies the degree of the transparency of the polygon fill, outline, and
label, when these attributes are displayed.
DISPLAY=(display-options ) | STANDARD | ALL
specifies which polygon features to display.
FILLATTRS=style-element | style-element (fill-options ) | (fill-options )
specifies the appearance of the filled polygon areas.
FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.
OUTLINEATTRS=style-element | style-element (line-options ) | (line-options )
specifies the appearance of the polygon outline.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
ROTATE=numeric-column | numeric-constant | expression
specifies the angle of rotation for the polygon, measured in degrees.
ROTATELABEL=AUTO | NONE | VERTICAL
specifies the rotation of the polygon label with respect to the rotation of
the polygon.

Axes options
750 Chapter 6 / Plot Statements

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over
the polygon.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
LABEL=column | expression
specifies the label for the polygon.
LABELATTRS=style-element | style-element (text-options ) | (text-options )
specifies the color and font attributes of the polygon label.
LABELLOCATION=INSIDEBBOX | OUTSIDEBBOX | OUTSIDE
specifies the location of the polygon label.
LABELPOSITION=CENTER | XMIN | XMAX | YMIN | YMAX
specifies the position of the polygon label with respect to the label
location.
LABELSPLIT=TRUE | FALSE
specifies whether to split the polygon label at the specified split
characters.
LABELSPLITCHAR="character-list"
specifies one or more characters on which the polygon label can be split
if needed.
LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the polygon label.
LABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the polygon label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
 POLYGONPLOT Statement 751

creates a separate and visually distinctive polygon for each unique

grouping value.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.
XOFFSET=numeric-column | expression
specifies an individual offset from the discrete X value on each vertex of
the polygon.
YOFFSET=numeric-column | expression
specifies an individual offset from the discrete Y value on each vertex of
the polygon.

ODS options
URL=string-column
specifies an HTML page that is displayed when the polygon is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.

Note A missing value in the X column signals the start of the data for a hole in
the polygon. See “Drawing Holes in a Polygon” on page 773.

Y=column | expression
specifies the column for the Y values.

Note A missing value in the Y column signals the start of the data for a hole in
the polygon. See “Drawing Holes in a Polygon” on page 773.

ID=column | expression
specifies the column that contains the ID value that is associated with each
polygon.

Restriction Only unformatted values in the ID= column are used.

Requirements All of the observations for a single polygon must have the same
ID value.

When multiple polygons are defined in the same data set, all of
the observations for a given ID must be defined contiguously.
Interspersing the ID observations in the data set can produce
unexpected results.

Note Observations that have a missing value in the ID column are

ignored by the POLYGONPLOT statement.
752 Chapter 6 / Plot Statements

Optional Arguments
ANTIALIAS=AUTO | OFF
specifies whether anti-aliasing is turned off for this plot.

Note: This feature applies to SAS 9.4M2 and to later releases.

AUTO
specifies that anti-aliasing is controlled by the ANTIALIAS= option in the
ODS GRAPHICS statement.

OFF
specifies that anti-aliasing is always disabled for this plot.

Default AUTO

Interaction This option overrides the ANTIALIAS= option in the ODS GRAPHICS
statement.

BACKLIGHT=number | AUTO
specifies a back-light effect for the polygon label text.

Note: This feature applies to SAS 9.4M2 and to later releases.

The effect is applied only to the polygon label text.

number
specifies the degree of the back-light effect.

Range 0–1, where 0 specifies no effect and 1 specifies maximum effect

AUTO
the system selects an appropriate level for the back-light effect.

If the GROUP= or COLORRESPONSE= option is in effect, BACKLIGHT=0.75.

Otherwise, BACKLIGHT=0.5.

The following figure shows the effect on a polygon label located inside the
polygon bounding box.

The back light is based on text color. For dark colors, a contrasting white back-
light effect is used. For lighter colors, a contrasting black back-light effect is
used. The following figure shows the two back-light types when BACKLIGHT=1.
 POLYGONPLOT Statement 753

Default 0 (no back-light effect)

Restriction Vector graphics output cannot be generated when the back-light

effect is applied. If you request vector graphics output and enable
the back-light effect, an image is generated instead.

Interaction The LABEL= option must be specified for this option to have any
effect.

Note If the background color is white, the white backlight effect for dark
text colors is not visible. Conversely, if the background color is
black, the black backlight effect for light text colors is not visible.

Tip The BACKLIGHT= option is most effective when the text color has a
low level of contrast with the background or when the background is
cluttered.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Defaults For a filled polygon, the ThreeColorRamp style element

For an unfilled polygon, the ThreeColorAltRamp style element

754 Chapter 6 / Plot Statements

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variable to use to
determine the polygon colors.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement The COLORRESPONSE= value should remain constant for the

same ID value. Otherwise, unexpected results might occur.

Interactions Prior to SAS 9.4M5, when the GROUP= option is specified with the
COLORRESPONSE= option, the GROUP= option is ignored.
Starting with SAS 9.4M5, the COLORRESPONSE= option controls
the color of the group fill patterns when the fill patterns are
displayed.

When fill is displayed, this option overrides suboption COLOR= in

the FILLATTRS= option and varies the fill color according to the
color gradient or the attribute map.

When only the outline is displayed, this option overrides suboption

COLOR= in the OUTLINEATTRS= option and varies the outline
color according to the color gradient or the attribute map.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of filled polygons. The following figure shows a
polygon with each of the skins applied.
 POLYGONPLOT Statement 755

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the DISPLAY= option must
include FILL.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The appearance of the data skin is based on the FILLATTRS= color.

When a data skin is applied, all polygon outlines are set by the skin
and the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the polygon fill, outline, and label,
when these attributes are displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

Tip You can use the FILLATTRS= option to set transparency for just the
filled polygon areas. You can combine this option with FILLATTRS=
to set one transparency for the polygon outline and label and a
different transparency for the polygon fill. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
756 Chapter 6 / Plot Statements

DISPLAY=(display-options ) | STANDARD | ALL

specifies which polygon features to display.

(display-options)
specifies a space-separated list of one or more display options enclosed in
parentheses.

The list can include one or more of the following options:

FILL
displays the polygon fill

FILLPATTERN
displays a pattern-filled polygon

Note This feature applies to SAS 9.4M5 and to later releases.

Tip By default, this option has no effect if the active ODS style does
not specify a fill pattern. To display a fill pattern in that case,
select an ODS style that supports fill patterns or use the
FILLPATTERNATTRS= option to specify a fill pattern for the
polygon.

OUTLINE
displays the polygon outline

STANDARD
specifies OUTLINE only

ALL
specifies all features: FILL, FILLPATTERN, and OUTLINE

Defaults In SAS 9.4M4 and in earlier releases, STANDARD

Starting in SAS 9.4M5, the GraphPolygon:DisplayOpts style reference,

if defined in the active ODS style. Otherwise, STANDARD.

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the polygon outline and fill color. Starting with SAS
9.4M5, use the FILLPATTERNATTRS= on page 757 option to control
the appearance of the polygon fill pattern.

FILLATTRS=style-element | style-element (fill-options ) | (fill-options )

specifies the appearance of the filled polygon areas.

Defaults For non-grouped data, the GraphDataDefault:Color style reference.

For grouped data, the Color attribute of GraphData1–GraphDataN

style elements.

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or by the DISPLAY= option.

When COLORRESPONSE= is in effect and the DISPLAY= option

enables FILL display, the FILLATTRS= suboption COLOR= is
ignored, and the polygon fill colors vary according to the gradient.
 POLYGONPLOT Statement 757

Tip The DATATRANSPARENCY= option sets the transparency for the

polygon fill and outline. You can combine this option with
DATATRANSPARENCY= to set one transparency for the outlines
but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLPATTERNATTRS=style-element | (fill-pattern-options)
specifies the appearance of the pattern-filled areas.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For non-grouped data, the ContrastColor and FillPattern attributes

of GraphDataDefault.

For grouped data, the ContrastColor and FillPattern attributes of

GraphData1–GraphDataN.

Interactions For this option to take effect, the DISPLAY= option must include
FILLPATTERN among the display options.

Starting with SAS 9.4M5, when COLORRESPONSE= is in effect and

the DISPLAY= option enables FILLPATTERN display, the
FILLPATTERNATTRS= suboption COLOR= is ignored, and the fill-
pattern colors vary according to the gradient.

Tip In order to change the fill pattern for each group value in ODS
styles such as HTMLBlue that use color-priority attribute rotation,
specify ATTRPRIORITY=NONE in an ODS GRAPHICS statement or
in the BEGINGRAPH statement in your graph template. For more
information, see “Attribute Rotation Patterns” in SAS Graph
Template Language: User’s Guide.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Pattern Options” on page 1601 for available fill-pattern-options.

GROUP=column | discrete-attr-var | expression

creates a separate and visually distinctive polygon for each unique grouping
value.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.
758 Chapter 6 / Plot Statements

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

A distinct polygon is created for each group value by varying the visual
attributes of the polygon display features. The display features are controlled
by the current ODS style or by the DISPLAY= option. The default group
appearance for each display feature is shown in the following table.

Display
Feature 1 Style Attributes That Control Default Group Appearance

Fill color Color attribute of a GraphData1–GraphDataN style element or

the GraphMissing style element.

Fill pattern 2 FillPattern and ContrastColor attributes of a GraphData1–

GraphDataN style element or the GraphMissing style element.
The ContrastColor attribute controls the color of the fill pattern.

Outline n if the outline is enabled with fill

color, fill pattern, or both,
ContrastColor attribute of a
GraphData1–GraphDataN style
element or the GraphMissing style
element with a solid line style
n if the outline is the only display
feature enabled, ContrastColor and
LineStyle attributes of a
GraphData1–GraphDataN style
element or the GraphMissing style
element

1 Fill color and the outline are displayed by default.

2 Fill patterns are valid for polygon plots starting with SAS 9.4M5.

Requirement The group value must remain constant for the same ID value.
Otherwise, the results are unpredictable.

Interactions This option is ignored if the COLORRESPONSE= option is also

used.

If a discrete attribute map variable is specified, then the colors and

outline patterns are mapped according to the associated
DISCRETEATTRMAP statement. See “DISCRETEATTRMAP
Statement” on page 1505. Otherwise, the colors and outline
patterns are mapped according to data order.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note The group values should remain constant for the same ID value.
 POLYGONPLOT Statement 759

Tip By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use
the GROUPMAX= option in the ODS GRAPHICS statement to
change the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Restriction This option is ignored and missing group values are not displayed
when ID= and GROUP= specify the same values.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

LABEL=column | expression
specifies the label for the polygon.

Default No label is displayed for the polygon

Note The label text should be the same for all of the observations for a
polygon ID. When different labels are specified for the same ID, the
label that is specified in the first observation for that ID is used.

Tips The default label text color is based on the use of polygon fill and
outline colors, and on whether the GROUP= or COLORRESPONSE=
option is specified. To change the label text color and font, use the
LABELATTRS= option.

Use the LABELLOCATION= and LABELPOSITION= options to change

the location of the polygon label.

For long labels, use the LABELSPLIT=, LABELSPLITCHAR=,

LABELSPLITCHARTDROP=, and LABELSPLITJUSTIFY= options to split
the label into multiple lines.

LABELATTRS=style-element | style-element (text-options ) | (text-options )

specifies the color and font attributes of the polygon label.

Defaults For non-grouped data, the GraphDataText style element, unless

COLORRESPONSE= is specified. When COLORRESPONSE= is
specified, the label color is determined by the outline color when
760 Chapter 6 / Plot Statements

outlines are displayed, or by the GraphDataText style element Color

attribute when outlines are not displayed. All other text attributes
are derived from the GraphDataText style element.

For grouped data, the GraphData1–GraphDataN style elements. The

label color is determined by the contrast color attribute.

Interaction If one or more text options are specified and they do not include all
of the font properties (such as color, family, size, weight, and style),
then the non-specified properties are derived from the
GraphLabelText style element or from a GraphData1–GraphDataN
style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELLOCATION=INSIDEBBOX | OUTSIDEBBOX | OUTSIDE

specifies the location of the polygon label.

INSIDEBBOX
locates the label inside the bounding box of the polygon.

OUTSIDEBBOX
locates the label outside the bounding box of the polygon but inside the plot
area.

OUTSIDE
locates the label outside the plot area.

Default INSIDEBBOX

Restriction OUTSIDE cannot be used when the POLYGONPLOT statement is

used in multi-cell layouts such as LATTICE, DATAPANEL, or
DATALATTICE in which the axes might be external to the grid.

Interaction The LABEL= option must be specified for this option to have any
effect.

Tip The label's exact position is relative to the polygon's X and Y data
ranges and is determined by the combination of this option and the
LABELPOSITION= option.

LABELPOSITION=CENTER | XMIN | XMAX | YMIN | YMAX

specifies the position of the polygon label with respect to the label location.

CENTER
centers the label in the polygon’s bounding box

XMIN
positions the label at the polygon's minimum X value and centers it in the Y-
value range

XMAX
positions the label at the maximum X value and centers it in the Y-value
range
 POLYGONPLOT Statement 761

YMIN
positions the label at the minimum Y value and centers it in the X-value
range

YMAX
positions the label at the maximum Y value and centers it in the X-value
range

The following figure shows the label positions for each of the label locations
that are specified by the LABELLOCATION= option.

Defaults CENTER when LABELLOCATION=INSIDEBBOX

YMAX when LABELLOCATION=OUTSIDEBBOX or

LABELLOCATION=OUTSIDE

Restriction CENTER is valid only when LABELLOCATION=INSIDEBBOX.

Interaction When LABELLOCATION=OUTSIDE, increasing label length might

cause the available plot area to decrease.

Tip When LABELLOCATION=OUTSIDE, the polygon label might collide

with the axis tick values on the orthogonal axis. In that case, if the
secondary orthogonal axis is not being used, specify the opposite
end of the axis. Otherwise, change LABELLOCATION= to
INSIDEBBOX or OUTSIDEBBOX.
762 Chapter 6 / Plot Statements

LABELSPLIT=TRUE | FALSE
specifies whether to split the polygon label at the specified split characters.
When this option is set to TRUE, the polygon label is split unconditionally at
each occurrence of any of the specified split characters.

Default FALSE. The polygon label is not split.

Requirement The LABEL= option must also be specified.

Interaction The LABELSPLITCHAR= option specifies one or more characters

on which splits can occur.

See “boolean ” on page xii for other Boolean values that you can use.

LABELSPLITCHAR="character-list"
specifies one or more characters on which the polygon label can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the polygon label. In that case, all of the specified split
characters together are treated as a single split character.

When LABEL= is specified and LABELSPLIT=TRUE, the polygon label is split

unconditionally at each occurrence of any of the specified split characters. If the
polygon label does not contain any of the specified characters, then the label is
not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
labelsplitchar="abc"

The LABEL= option and the LABELSPLIT=TRUE option must also

be specified.

Interaction The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the polygon label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the polygon label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the polygon label.
 POLYGONPLOT Statement 763

TRUE
drops the split characters from the polygon label.

FALSE
includes the split characters in the polygon label.

When LABELSPLIT=TRUE and LABELSPLITCHARDROP=FALSE, each split

character remains as the last character in the current line. The characters
that follow the split character, up to and including the next split character,
are then wrapped to the next line.

The following figure shows an example of polygon label Product*Group*A split

on * when LABELSPLITCHARDROP=TRUE and
LABELSPLITCHARDROP=FALSE.

Default TRUE. The split characters are dropped from the polygon label.

Requirement The LABEL= option and the LABELSPLIT=TRUE option must also
be specified.

Interaction The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

LABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the polygon label blocks.

AUTO
justifies the labels based on the LABELPOSITION= option.

LABELPOSITION= Value Justification

XMIN RIGHT

XMAX LEFT

CENTER, YMAX, or YMIN CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which LABELPOSITION=YMAX.

764 Chapter 6 / Plot Statements

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because LABELPOSITION=YMAX, AUTO centers the lines of text.

The text box is anchored the same way that the unsplit text is anchored.

Default AUTO

Requirement The LABEL= option and the LABELSPLIT=TRUE option must also
be specified.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options ) | (line-options )

specifies the appearance of the polygon outline.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, unfilled polygons use both the

CONTRASTCOLOR and PATTERN attributes of the GraphData1–
GraphDataN style elements. Filled polygons use only the
ContrastColor attribute.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or by the DISPLAY= option.
 POLYGONPLOT Statement 765

If the DATASKIN= option applies a data skin, then this option is

ignored.

When the COLORRESPONSE= and DISPLAY=(OUTLINE) options

are in effect, the OUTLINEATTRS= suboption COLOR= is ignored,
and the polygon outline colors vary according to the gradient.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)
766 Chapter 6 / Plot Statements

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles ID, COLORRESPONSE, GROUP, LABEL, and
URL.

ROTATE=numeric-column | numeric-constant | expression

specifies the angle of rotation for the polygon, measured in degrees. Positive
angles rotate the image counter clockwise, and negative angles rotate the image
clockwise. The angle specification can exceed 360 degrees in absolute value.

Default 0. No rotation is performed.

Interaction When this option is specified, the LABELLOCATION= and

LABELPOSITION= options are overridden with
LABELLOCATION=INSIDEBOX and LABELPOSITION=CENTER.

Notes The rotation angle is measured in screen coordinates.

A missing value in the rotation data is treated as 0.

Rotating a polygon does not change data ranges that are reported to
the axes. As a result, clipping might occur in some cases.

ROTATELABEL=AUTO | NONE | VERTICAL

specifies the rotation of the polygon label with respect to the rotation of the
polygon.

AUTO
rotates the label with the rotation of the polygon.

NONE
does not rotate the label with the rotation of the polygon.

The label position remains fixed regardless of the polygon rotation.

VERTICAL
rotates the label to a vertical position.

Restriction VERTICAL is valid only when the polygon is not rotated

(ROTATE=0). When the polygon is rotated,
ROTATELABEL=VERTICAL is ignored, and the default (AUTO) is
used instead.

Default AUTO

Interaction The LABEL= option must be specified for this option to have any
effect.

TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over the
polygon. If you use this option, it replaces all of the information that is displayed
by default. You can specify roles for columns that do not contribute to the
polygon plot along with roles that do.
 POLYGONPLOT Statement 767

(role-list)
an ordered, space-separated list of unique POLYGONPLOT and user-defined
roles.

POLYGONPLOT roles include ID, COLORRESPONSE, GROUP, LABEL, and

URL.

Tip User-defined roles are defined with the ROLENAME= option.

Example The following example displays the columns that are assigned to
the roles ID and URL, and the columns XOffset and YOffset in the
data tips. The XOffset and YOffset columns are not assigned to
any predefined POLYGONPLOT role, so they must first be assigned
a role:

ROLENAME=(TIP1=XOFFSET TIP2=YOFFSET)
TIP=(ID TIP1 TIP2 URL)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns that are assigned to the following roles are
automatically included in the data tip information: ID,
COLORRESPONSE or GROUP, LABEL, and URL.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)
768 Chapter 6 / Plot Statements

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

URL=string-column
specifies an HTML page that is displayed when the polygon is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each polygon that is to have an active link.

Example http://www.sas.com

Requirement To generate a plot with selectable polygons, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note The URL values should remain constant for the same ID value.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.
 POLYGONPLOT Statement 769

XOFFSET=numeric-column | expression
specifies an individual offset from the discrete X value on each vertex of the
polygon.

Default 0 (all polygon vertices are centered on the discrete X values)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right on discrete X values. If
option REVERSE=TRUE is specified in the layout's X-axis options,
then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Interaction This option overrides the DISCRETEOFFSET= option.

Tip Setting the discrete offset for the plots does not affect the axis
minimum offset and maximum offset. In some cases, setting a
discrete offset can cause clipping at each end of the axis. In those
cases, use the OFFSETMIN= and OFFSETMAX= axis options to
increase the axis minimum and maximum offsets in order to
accommodate the discrete offset.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YOFFSET=numeric-column | expression
specifies an individual offset from the discrete Y value on each vertex of the
polygon.

Default 0 (all polygon vertices are centered on the discrete Y values)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is up on discrete Y values. If option
REVERSE=TRUE is specified in the layout's Y-axis options, then the
offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Interaction This option overrides the DISCRETEOFFSET= option.

Tip Setting the discrete offset for the plots does not affect the axis
minimum offset and maximum offset. In some cases, setting a
discrete offset can cause clipping at each end of the axis. In those
cases, use the OFFSETMIN= and OFFSETMAX= axis options to
increase the axis minimum and maximum offsets in order to
accommodate the discrete offset.
770 Chapter 6 / Plot Statements

Details

Overview
The POLYGONPLOT statement draws a polygon from a series of X-Y value pairs
that are stored in a SAS data set. The first X-Y value pair defines the starting point
of the polygon. The next X-Y pair in the data specifies the coordinates of the first
vertex. A line segment is drawn from the starting point to the first vertex. For
subsequent X-Y pairs, a line segment is drawn from the previous vertex to the
current vertex. This pattern repeats until all of the segments have been drawn. If
the last segment does not close the polygon, then the POLYGONPLOT statement
automatically draws a segment from the last vertex back to the starting point in
order to close the polygon.

Using the POLYGONPLOT statement, you can draw any data-driven shape on your
graph, which enables you to highlight data features, outline data boundaries, and so
on. Unlike DRAW statements in a BEGINPLOYGON block, the following conditions
apply to the POLYGONPLOT statement:
n you need to modify only the polygon data in the graph data set to modify the
polygon. You do not need to modify the template code. (See “BEGINPOLYGON
Statement” on page 1405.)
n you can draw polygons between plots. The BEGINPOLYGON block and DRAW
statements can draw polygons only on top of or behind the graph.

Requirements for the Polygon Data Set

In the simplest case of a single polygon, your data set must provide an X, Y, and ID
column that stores the X-Y values and the ID for your polygon. The X-Y values in
the first data-set observation must specify the starting point of your polygon. The
X-Y values in the subsequent observations must provide the coordinates of each
vertex in the order in which the polygon is to be drawn. There should be no gaps in
the data. If your last observation does not close the polygon, then the
POLYGONPLOT statement automatically draws a segment from your last vertex
back to the starting point in order to close the polygon.

If you want to draw multiple polygons, then your ID column must specify a unique
identifier value for each polygon. The identifier value associates the observations in
the data set with a specific polygon. All of the observations for each individual
polygon must be grouped together by ID and must be arranged in the order in which
the polygon segments are to be drawn.

The following table shows additional columns that you can use to customize your
polygon plot.

Column1 Type Description

COLORRESPONSE Numeric Specifies the numeric column or range

(p. 754) attribute map variable that is used to
determine the polygon colors.
 POLYGONPLOT Statement 771

Column1 Type Description

GROUP (p. 757) Numeric or Creates a separate polygon color, outline

character pattern, or both for each unique grouping
that is specified.

LABEL (p. 759) Numeric or Specifies the label for the polygon.
character

URL (p. 768) Character Specifies an HTML page that is displayed

when the polygon is selected.

XOFFSET (p. 769) Numeric Specifies an individual offset from the

discrete X value on each vertex of the
polygon.

YOFFSET (p. 769) Numeric Specifies an individual offset from the

discrete Y value on each vertex of the
polygon.

1 You can specify any valid column name for these columns in your data set.

Drawing a Single Polygon

For a single polygon, the polygon data set contains an X column and Y column that
define the polygon vertices, and an ID column that specifies a constant value. The
polygon segments are drawn in the order in which they occur in the data. If the
polygon overlaps any graphics elements that were drawn earlier, those elements
are obscured. In that case, you can use transparency to enable the underlying
graphics elements to show through.

Here is example data for a simple four-sided polygon that is identified as P1 and
that starts at point X=40, Y=100.
Polygon Data

Obs id x y

1 P1 40 100
2 P1 20 220
3 P1 160 200
4 P1 180 80
5 P1 40 100

The following figure shows how the polygon is drawn. Grid lines are provided to
help you locate the polygon vertices in the output.
772 Chapter 6 / Plot Statements

The polygon starting point is X=40, Y=100 (shown in red). From the starting point,
the segments are drawn in data order. Data order is in a clockwise direction, as
indicated by the gray arrow. Although the last observation (X=40, Y=100) is
provided in this example, it is not required. If the last observation is not provided in
the data, then the POLYGONPLOT statement draws the last segment automatically
in order to close the polygon.

For an example, see “Example 1: Drawing a Simple Polygon That Highlights Data” on
page 775.

Drawing Multiple Polygons

For multiple polygons, the POLYGONPLOT data ID column specifies a unique
identifier for all of the observations that are associated with each polygon. The X
and Y columns specify the polygon vertices. The polygons are overlaid on the graph
in the order in which they occur in the data. For overlapping polygons, each polygon
obscures part or all of the polygons and graphics elements that were drawn before
it. In that case, you can use transparency to enable the underlying polygons and
graphics elements to show through.

Here is example data for three separate polygons.

Polygon Data

Obs id x y label

1 1 0 0 ID=1
2 1 20 0
3 1 20 30
4 1 0 30
5 1 0 0
6 2 30 0 ID=2
7 2 50 0
8 2 40 30
9 2 30 0
10 3 60 0 ID=3
 POLYGONPLOT Statement 773

11 3 80 5
12 3 80 15
13 3 70 30
14 3 60 30
15 3 60 0

In addition to the ID column, X column, and Y column, the Label column is added to
label the polygons in the output. Notice that the observations for each ID value are
grouped together in the data set. The observations for each ID must occur
contiguously in the data. Otherwise, unexpected results might occur.

The following figure shows how the polygons are drawn from this data. Grid lines
are provided to help you locate the polygon vertices in the output.

The polygons are drawn in the order in which they appear in the data: rectangle
(ID=1), triangle (ID=2), and polygon (ID=3). The red dot on each shape indicates the
starting point for that shape. The gray arrow indicates the direction in which the
segments are drawn for each shape.

Drawing Holes in a Polygon

Using the POLYGONPLOT statement, you can draw one or more holes inside a
polygon. To create data for a polygon with one or more holes:

1 Specify the X and Y values for the outer polygon.

2 To start the data for a hole, add an observation that has missing X and Y values.
The missing X and Y values signal the POLYGONPLOT statement that the
observations that follow define the data for a hole.

3 Specify the X and Y values for the hole polygon.

4 Repeat Steps 2 and 3 for each additional hole.

Here is example data for a simple polygon that has two holes.
Polygon Data
774 Chapter 6 / Plot Statements

Obs id x y

1 1 1 1
2 1 9 1
3 1 7 5
4 1 9 9
5 1 1 9
6 1 3 5
7 1 1 1
8 1 . .
9 1 3 8
10 1 7 8
11 1 6 6
12 1 4 6
13 1 3 8
14 1 . .
15 1 3 2
16 1 7 2
17 1 6 4
18 1 4 4
19 1 3 2

Observations 1–7 specify the data for the outer polygon. In observation 8, the X and
Y values are missing, which indicates the start of the data for the first hole.
Observations 9–13 define the data for the first hole polygon. Observation 14
indicates the start of the data for the second hole, which is defined by observations
15–19.

The following figure shows how the polygon is drawn from this data. Grid lines are
provided to help you locate the polygon vertices in the output.

The outer polygon is drawn first, starting at point X=1, Y=1. The segments are drawn
in data order, which is in a counterclockwise direction as indicated by the gray
arrow. The first hole is drawn next, starting at point X=3, Y=8. Its segments are
 POLYGONPLOT Statement 775

drawn in a clockwise direction. The second hole is drawn last, starting at point X=3,
Y=2. Its segments are drawn in a counterclockwise direction.

Examples

Example 1: Drawing a Simple Polygon That Highlights Data

This example shows you how to use the POLYGONPLOT statement to draw a filled
polygon that highlights data in an iris petal dimension scatter plot. The polygon
surrounds the markers for the Setosa species in order to highlight the data for that
species. This example is a modified version of the example in “Example:
BEGINPOLYGON Statement” on page 1412. This version uses the POLYGONPLOT
statement instead of a BEGINPOLYGON block so that you can compare the two
methods. The following figure shows the output for this example.

Here is the SAS code for this example.

/* Generate the data for the polygon */
data polydata;
input polyID polyX polyY label $8-40;
datalines;
1 9 2 Setosa
1 13 5
1 16 7
1 17 6
1 20 5
1 20 1
1 17 1
1 15 0
1 14 0
 776 Chapter 6 / Plot Statements

1 11 0
;
run;

/* Concatenate the SASHELP.IRIS and polygon data into

data set IRIS */
data iris;
set sashelp.iris polydata;
run;

/* Create the template for the graph */

proc template;
define statgraph discretelegend;
begingraph / drawspace=datavalue;
entrytitle "Iris Petal Dimensions";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth / name="s"
group=species includemissinggroup=false;
ellipse x=petallength y=petalwidth / type=predicted alpha=.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth / type=predicted alpha=.05
name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
polygonplot x=polyX y=polyY id=polyID / display=(fill)
fillattrs=(color=yellow transparency=0.75)
label=label labellocation=outsidebbox labelposition=ymax;
discretelegend "s" / title="Species:";
discretelegend "p80" "p95" /across=1 autoalign=(topleft)
location=inside;
endlayout;
entryfootnote halign=left "Fisher's Iris Data" ;
endgraph;
end;
run;

/* Generate the graph */

proc sgrender data=iris template=discretelegend;
run;

Details
To draw a single polygon, the data set must provide an X column, a Y column, and
an ID column. The data specifies the polygon vertices around the Setosa data in a
clockwise direction. A Label column is added to provide a label for the polygon in
the plot output. Concatenation of the Polydata and Sashelp.Iris data sets results in
missing values for the SCATTERPLOT statement grouping variable in the Iris data
set. By default, the SCATTERPLOT statement includes missing group values. To
exclude the missing group values, the INCLUDEMISSINGGROUP=FALSE option is
added to the SCATTERPLOT statement.

In the POLYGONPLOT statement, the DISPLAY= option specifies the polygon fill
only. The FILLATTRS= option specifies the fill color as yellow and a fill
transparency of 0.75. Rather than using draw statements to draw an annotation for
the polygon, this example uses the POLYGONPLOT statement label feature to
 POLYGONPLOT Statement 777

label the polygon. The LABEL= option specifies the column in the data set that
contains the polygon label text. The LABELLOCATION= and LABELPOSITION=
options place the polygon label outside of and above the polygon’s bounding box.

To draw the polygon, the POLYGONPLOT statement starts at X=9, Y=2, and draws
a segment between each vertex in data order. The last vertex, X=11, Y=0, does not
close the polygon. To close the polygon, the POLYGONPLOT statement draws a
segment between X=11, Y=0 and X=9, Y=2 automatically.

If you want to highlight the Versicolor data instead of the Setosa data, then you
need only modify the data in the Polydata data set to draw a polygon around the
Versicolor data instead. You do not have to make any changes to the template
code.

Example 2: Drawing a Geographical Boundary

This example shows you how to use the POLYGONPLOT statement to draw an
outline of Wake County in North Carolina. The polygon data is extracted from data
set Maps.Counties using state ID 37 for North Carolina and county ID 183 for Wake
County. The GPROJECT procedure is then used to project the data before it is
plotted with the POLYGONPLOT statement. The Maps.Counties data set is
installed with SAS/GRAPH. If SAS/GRAPH is not installed at your site, you can
download data set Maps.Counties from SAS Maps Online. For more information,
see “Map Data Sets, Map Preparation Procedures, and Tools Provided by SAS” in
SAS/GRAPH and Base SAS: Mapping Reference.

The following figure shows the output for this example.

Here is the SAS code for this example.

/* Extract the map data for Wake County, NC. */
data wakecounproj;
set maps.counties(where=(state=37 and county=183));
label="Wake County, NC";
run;
 778 Chapter 6 / Plot Statements

/* Project the data */

proc gproject data=wakecounproj out=wakeco dupok;
id state county;
run;

/* Draw the county border */

proc template;
define statgraph wakecounty;
begingraph;
layout overlay / walldisplay=(fill)
xaxisopts=(display=none offsetmin=0.05 offsetmax=0.05)
yaxisopts=(display=none offsetmin=0.05);
polygonplot x=x y=y id=state / display=(outline fill)
fillattrs=(color=cxd1d1d1)
outlineattrs=(color=darkgreen thickness=2px)
label=label labelattrs=(weight=bold color=darkgreen
size=16pt);
endlayout;
endgraph;
end;
run;

proc sgrender data=wakeco template=wakecounty;

run;

REFERENCELINE Statement
Creates a horizontal or vertical reference line.

Requirement: A REFERENCELINE statement can be used within a 2-D layout (OVERLAY,

OVERLAYEQUATED, DATALATTICE, or DATAPANEL) only.
Note: Specifying the X= option creates a line perpendicular to the X axis at an X-
intercept. Specifying the Y= option creates a line perpendicular to the Y axis at a Y-
intercept.

Syntax
REFERENCELINE X=x-axis-value | column | expression </options>;

REFERENCELINE Y=y-axis-value | column | expression </options>;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the reference line data is to be considered when
determining the data range for the axis.
 REFERENCELINE Statement 779

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the reference lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the reference line.
DISCRETELINETHICKNESS=number
specifies the line thickness as a fraction of the midpoint spacing.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the reference line.
USEDISCRETETHICKNESS=TRUE | FALSE
specifies that the line thickness should be based on a fraction of the
midpoint spacing that is set by the DISCRETELINETHICKNESS= option.

Axes options
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Label options
CURVELABEL="string" | column | expression
specifies a label for the reference line or lines.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the reference line label(s).
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the reference line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN
specifies the position of the reference line label relative to the reference
line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the reference line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the reference line label can be
split if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the reference line
label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the reference line
label block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
780 Chapter 6 / Plot Statements

specifies an amount to offset all reference lines from the specified values
when the X or Y axis is discrete.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
You must use either the X= or the Y= argument in the REFERENCELINE statement.
X=x-axis-value | column | expression
specifies the X intercept of the reference line or lines.

Requirements If X is not specified, then Y must be specified.

Values must be the same type as the data type of the X axis. For
example, you should use numeric SAS date or time values (or SAS
date/time constants) for a time axis.

Unformatted numeric values do not map to a formatted discrete

axis. When the X axis is a discrete axis, the X-axis value must be
the formatted value that appears on the X axis. If a column is
specified for the values in that case, then the specified column
must use the same format that is used for the X axis.

Note When a character value is specified, leading blanks are honored

and trailing blanks are ignored when the specified value is
compared with the data values.

Tips By default, if the value that is specified for the X= argument is

outside of the data range, then the data range is extended to
include the specified intercept. You can change this behavior with
the CLIP= option.

You can use the COLN() or COLC() function in an EVAL()

expression to specify multiple reference lines on the X axis. See
“GTL Functions Used with the EVAL Function” on page 1547.

Y=y-axis-value | column | expression

specifies the Y intercept of the reference line or lines.

Requirements If Y is not specified, then X must be specified.

Values must be the same type as the data type of the Y axis.

Unformatted numeric values do not map to a formatted discrete

axis. When the Y axis is a discrete axis, Y-axis value must be the
formatted value that appears on the Y axis. If a column is
specified for the values in that case, then the specified column
must use the same format that is used for the Y axis.
 REFERENCELINE Statement 781

Note When a character value is specified, leading blanks are honored

and trailing blanks are ignored when the specified value is
compared with the data values.

Tips By default, if the value that is specified for the Y= argument is

outside of the data range, then the data range is extended to
include the specified intercept. You can change this behavior with
the CLIP= option.

You can use the COLN() or COLC() function in an EVAL()

expression to specify multiple reference lines on the Y axis. See
“GTL Functions Used with the EVAL Function” on page 1547.

Optional Arguments
CLIP=TRUE | FALSE
specifies whether the reference line data is to be considered when determining
the data range for the axis.

FALSE
specifies that the reference line values are to be considered when the axis
range is determined. The reference lines are drawn as follows based on the
axis type:
n For a discrete axis, the reference line values that are not already on the
axis are added to the end of the axis data list. When applicable, the axis
values are then sorted:
o If the axis values are numeric values, then they are sorted ordinally.
o If the axis values are character values and a sorting option is applied to
the axis, then they are sorted as specified by the sorting option.
Reference lines are then drawn at the specified locations.
n For a linear, log, or time axis, a new axis data list is created by performing
a mathematical union of the data values and the reference line values.
The reference lines are then drawn at the locations specified.

TRUE
specifies that the reference line values are not to be considered when the
axis range is determined. The reference lines are drawn as follows based on
the axis type:
n For a discrete axis, if the reference line value exactly matches a value on
the axis, a reference line is drawn at that location. Otherwise, the
reference line is not drawn.

Note: If the axis values are formatted, then the reference line value must
exactly match the formatted axis value in order for the line to be drawn.

n For a linear, log, or time axis, if the reference line value is within the axis
data range, then the reference line is drawn at the specified location.
Otherwise, the reference line is not drawn.
782 Chapter 6 / Plot Statements

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABEL="string" | column | expression

specifies a label for the reference line or lines.

Requirement If you use the COLN() or COLC() function in the X= or Y= option to

specify multiple reference line intercepts, then you must use the
COLC() function in the CURVELABEL= option to specify exactly
one label for each reference line intercept. Otherwise, this option is
ignored. See “GTL Functions Used with the EVAL Function” on
page 1547.

Interactions If the X or Y argument specifies a value, then use "string".

If the X or Y argument specifies a column, then use a column to

define the label for each value.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the reference line label(s).

Default The GraphValueText style element.

Interaction For this option to take effect, the CURVELABEL= option must also
be used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the reference line label relative to the plot area.

INSIDE locates the labels inside the plot area

OUTSIDE locates the labels outside the plot area

Default OUTSIDE

Restriction OUTSIDE cannot be used when the REFERENCELINE is used in

multicell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.
 REFERENCELINE Statement 783

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN

specifies the position of the reference line label relative to the reference line.

AUTO
automatically positions the line label near the line boundary along unused
axes whenever possible (typically Y2 and X2) in order to avoid collision with
tick values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, the top
or right).

MIN
forces the line label to appear near minimum line values (typically, the
bottom or left).

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

MAX when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION= INSIDE is

specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the line label
appears.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the reference line label might fall outside of the graph
area. In that case, the reference line label might not be displayed or
might be positioned incorrectly.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the reference line label at the specified split
characters. When a reference line label is split, the label is split on each
occurrence of the specified split characters.

Default FALSE. The reference line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.
784 Chapter 6 / Plot Statements

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the reference line label can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the reference line label. In that case, all of the specified split
characters together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the

reference line label is split unconditionally at each occurrence of any of the
specified split characters. If the reference line label does not contain any of the
specified characters, then the label is not split.

"character-list"
a list of one or more characters with no spaces between them enclosed in
quotation marks.

Example To specify the split characters a, b, and c:

curvelabelsplitchar="abc"

Default A blank space

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the reference line label or are
dropped.

Notes When multiple characters are specified, the order of the characters
in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the reference line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the reference line label
text.

TRUE
drops the split characters from the reference line label text.

FALSE
includes the split characters in the reference line label text. When
CURVELABELSPLIT=TRUE and CURVELABELSPLITCHARDROP=FALSE,
each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.
 REFERENCELINE Statement 785

The following figure shows an example of a Y-axis reference line label with the
following specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the reference line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the reference line
label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the reference line label
block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown
in the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

786 Chapter 6 / Plot Statements

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the reference line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the reference lines. The following figure
shows a blue reference line with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
 REFERENCELINE Statement 787

DATASKINMAX= option in your ODS GRAPHICS statement to

increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the LINEATTRS= color.

DATATRANSPARENCY=number
specifies the degree of the transparency of the reference line.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option does not affect the reference line label.

DISCRETELINETHICKNESS=number
specifies the line thickness as a fraction of the midpoint spacing.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1 (100%)

Range 0–1

Requirement The axis must be discrete and USEDISCRETETHICKNESS=TRUE

must be in effect. Otherwise, this option is ignored.

DISCRETEOFFSET=number
specifies an amount to offset all reference lines from the specified values when
the X or Y axis is discrete.

Default 0 (no offset, all reference lines are centered on discrete X or Y

values)

Range -0.5 to +0.5 where 0.5 represents half the distance between discrete
ticks. A positive offset is to the right for a vertical reference line and
up for a horizontal reference line. If the layout's axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.
788 Chapter 6 / Plot Statements

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the reference line.

Default The GraphReference style element.

Interaction This option is ignored when USEDISCRETETHICKNESS=TRUE is in

effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

.“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

USEDISCRETETHICKNESS=TRUE | FALSE
specifies that the line thickness should be based on a fraction of the midpoint
spacing that is set by the DISCRETELINETHICKNESS= option.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement The axis type must be discrete. Otherwise, this option is ignored.

Interaction When this option is set to TRUE, the LINEATTRS THICKNESS=

option is ignored.

Tips Use USEDISCRETETHICKNESS=TRUE with

DISCRETELINETHICKNESS=1, for example, to place custom bands
behind the midpoint values.

Use the LINEATTRS COLOR= option to specify the color.

Use DATATRANSPARENCY= to specify the transparency.

See DISCRETELINETHICKNESS= on page 787

 REFERENCELINE Statement 789

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Restriction Another plot that establishes a data range for the designed axis
must be included.

Interactions This option is ignored if the Y= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
Reference lines are always drawn perpendicular to the axes. They are drawn from
one axis boundary to the companion boundary (X to X2 or Y to Y2). Axis offsets do
not apply to reference lines.

A REFERENCELINE statement can be used only within 2-D overlay-type layouts

(OVERLAY, OVERLAYEQUATED, or PROTOTYPE). A standalone plot statement
that provides a sufficient data range for determining axis extents must be included
in the layout. For example, a REFERENCELINE statement can be used with a
scatter plot or a histogram.

If a column is used to generate multiple reference lines, then the column type
(numeric or string) must agree with the type of data presented on the axis.
 790 Chapter 6 / Plot Statements

Examples

Example 1: Specifying a Single Reference Line

This example shows you how to draw a reference line using the REFERENCELINE
statement. The follow figure shows the output for this example.

Example Program
Here is the SAS code for this example.
/* Create the template for the graph */
proc template;
define statgraph referenceline;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
referenceline x=3 /
lineattrs=(color=blue) curvelabel="X=3";
endlayout;
endgraph;
end;
run;

/* Generate the plot data */

data test;
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
 REFERENCELINE Statement 791

output;
end;
run;

/* Generate the graph */

proc sgrender data=test template=referenceline;
run;

Example 2: Specifying Reference Lines Using Data Columns

This example shows you how to specify the reference-line intercept values for
multiple reference lines using columns in the data set for the graph. The following
figure shows the output for this example.

Output 6.1 Multiple Reference Lines in a Graph

Example Program
Here is the SAS code for this example.
/* Define the template for the graph */
proc template;
define statgraph referenceline;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
referenceline y=yR / curvelabel=label
lineattrs=(color=gray pattern=dot);
 792 Chapter 6 / Plot Statements

endlayout;
endgraph;
end;
run;

/* Generate the graph data */

data graphdata;
/* Plot data */
do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
/* Reference line data */
x=.; y=.; yR=14; label='Y=14'; output;
x=.; y=.; yR=5; label='Y=5'; output;
run;

/* Generate the graph */

proc sgrender data=graphdata template=referenceline;
run;

Example 3: Specifying Reference Lines Using the COLN and

COLC Functions
This example shows you how to use the COLN() and COLC() functions to specify
multiple intercept values directly in the REFERENCELINE statement. This approach
is an alternative to including the reference line data in the plot data.The output is
shown in “Example 2: Specifying Reference Lines Using Data Columns” on page 791.

Example Program
Here is the SAS code for this example.
/* Create the template for the graph */
proc template;
define statgraph referencelines;
begingraph;
entrytitle "Line of Symmetry";
layout overlay / yaxisopts=(linearopts=(viewmin=0));
seriesplot x=x y=y;
/* Use COLN() to specify the intercept values */
referenceline y=eval(coln(14, 5)) /
/* Use COLC() to specify a label for each reference line */
curvelabel=eval(colc("Y=14", "Y=5"))
lineattrs=(color=gray pattern=dot);
endlayout;
endgraph;
end;
run;

/* Generate the plot data */

data test;
 REGRESSIONPLOT Statement 793

do X=0 to 8 by 0.25;
Y=(x-3)*(x-3) + 5;
output;
end;
run;

/* Generate the graph */

proc sgrender data=test template=referencelines;
run;

REGRESSIONPLOT Statement
Creates a fitted regression line or curve computed from input data.

Restriction: The REGRESSIONPLOT statement supports only models of one independent and
one dependent variable.
Tip: Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.

Syntax
REGRESSIONPLOT X=numeric-column | expression
Y=numeric-column | expression </<regression-options> <options>>;

Summary of Optional Arguments

Appearance options
DATATRANSPARENCY=number
specifies the degree of the transparency of the regression line and line
label.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the regression line.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
794 Chapter 6 / Plot Statements

specifies whether data are mapped to the primary X (bottom) axis or to

the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string"
specifies a label for the regression line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the regression line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the regression line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the regression line label relative to the
regression line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the regression line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the regression line label can be
split if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the regression line
label text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the regression line
label block.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of regression lines from just the observations that
correspond to each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
 REGRESSIONPLOT Statement 795

assigns a name to this plot statement for reference in other template

statements.

Required Arguments
X=numeric-column | expression
specifies the column for the X values.

Y=numeric-column | expression
specifies the column for the Y values.

Optional Arguments
CURVELABEL="string"
specifies a label for the regression line.

Defaults No regression line label is displayed

This option is not valid when the GROUP= option is specified.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the regression line labels.

Default The GraphValueText style element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

If the GROUP= option is specified, then this option is ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the regression line label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the REGRESSIONPLOT is used in

multi-cell layouts such as LATTICE, DATAPANEL, or
DATALATTICE, where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.
796 Chapter 6 / Plot Statements

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the line labels
appear.

See “Location and Position of Curve Labels” on page 213

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the regression line label relative to the regression line.

AUTO
automatically positions the line label near the line boundary along unused
axes whenever possible (typically Y2 and X2) in order to avoid collision with
tick values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the line label to appear near maximum line values (typically, upper
right).

MIN
forces the line label to appear near minimum line values (typically, lower
left).

START
forces the line label to appear near the beginning of the regression line.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the regression line has a
spiral shape.

END
forces the line label to appear near the end of the regression line.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the regression line has a
spiral shape.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the line label
appears.
 REGRESSIONPLOT Statement 797

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the label might fall outside of the graph area. In that
case, the regression-line label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

See “Location and Position of Curve Labels” on page 213

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the regression line label at the specified split
characters. When a regression line label is split, the label is split on each
occurrence of the specified split characters.

Default FALSE. The regression line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the regression line label can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the regression line label. In that case, all of the specified split
characters together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the

regression line label is split unconditionally at each occurrence of any of the
specified split characters. If the regression line label does not contain any of the
specified characters, then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.
798 Chapter 6 / Plot Statements

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the regression line label or are
dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the regression line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the regression line label
text.

TRUE
drops the split characters from the regression line label text.

FALSE
includes the split characters in the regression line label text. When
CURVELABELSPLIT=TRUE and CURVELABELSPLITCHARDROP=FALSE,
each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a regression line label with the
following specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the regression line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.
 REGRESSIONPLOT Statement 799

Default TRUE. The split characters are dropped from the regression line
label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the regression line label
block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option, as shown
in the following table.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the regression line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

800 Chapter 6 / Plot Statements

DATATRANSPARENCY=number
specifies the degree of the transparency of the regression line and line label.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

GROUP=column | discrete-attr-var | expression

creates a distinct set of regression lines from just the observations that
correspond to each unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of line color and line pattern. Line colors
vary according to the ContrastColor attribute of the GraphData1–
GraphDataN and GraphMissing style elements. Line patterns vary
according to the LineStyle attribute of the GraphData1–
GraphDataN style elements.

Requirement You must sort the data in ascending order by the specified group
column before you plot the data. If the plot data is a CAS in-
memory table, you must download the in-memory table to your
SAS client, and then sort it locally before you plot it.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line colors
and line patterns.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a
column in a CAS in-memory table, the visual attributes for each
group value are assigned in ascending order of the group column
character values or of unformatted numeric values.

Tips The LINEATTRS= option can be used to override the

representations that are used to identify the groups. For example,
LINEATTRS=(PATTERN=SOLID) can be used to assign the same
pattern to all of the lines, letting the line color distinguish group
values. Likewise, LINEATTRS=(COLOR=BLACK) can be used to
assign the same color to all of the lines, letting the line pattern
distinguish group values.
 REGRESSIONPLOT Statement 801

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use
the GROUPMAX= option in the ODS GRAPHICS statement to
change the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.
802 Chapter 6 / Plot Statements

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the regression line.

Default The GraphFit style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.
 REGRESSIONPLOT Statement 803

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIPFORMAT=(Y=6.2)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles
are automatically included in the data tip information.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIPLABEL=(Y="Curve")

Default The column label or column name of the column assigned to the role.

Note The columns assigned to the X, Y, and GROUP (if assigned) roles are
automatically included in the data tip information.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.
804 Chapter 6 / Plot Statements

Regression Options
ALPHA=positive-number
specifies the confidence level to compute.

Default 0.05

Range 0 < positive-number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLI="name"
produces confidence limits for individual predicted values for each observation.
The confidence level is set by the ALPHA= option.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the
confidence limits to be computed. To display confidence limits, you
must use this name as the required argument of a MODELBAND
statement. See the example in “Example: REGRESSIONPLOT
Statement” on page 805 .

CLM="name"
produces confidence limits for a mean predicted value for each observation. The
confidence level is set by the ALPHA= option.

Interaction name is a unique name within the template that is case sensitive and
cannot contain spaces. It must be assigned in order for the
confidence limits to be computed. To display confidence limits, you
must use this name as the required argument of a MODELBAND
statement. See the example in “Example: REGRESSIONPLOT
Statement” on page 805 .

DEGREE=non-negative-integer
specifies the degree of the polynomial.

The default value, DEGREE=1, produces a linear fit, DEGREE=2 produces a

quadratic fit, DEGREE=3 produces a cubic fit, and so on.

The value of the DEGREE=d option corresponds to one of the following

TRANSREG procedure specifications for the independent variable: SPLINE(X /
DEGREE=d) or PBSPLINE(X / DEGREE=d LAMBDA=0).

Default 1

Ranges 0–174 in SAS 9.4M1 and earlier releases.

1–10 starting with SAS 9.4M2.

FREQ=numeric-column
specifies a column in the input data set that represents the frequency of
occurrence of the current observation, essentially treating the data set as if
each observation appeared n times, where n is the value of the FREQ column for
the observation. Noninteger values of the FREQ column are truncated to the
 REGRESSIONPLOT Statement 805

largest integer less than the FREQ value. The observation is used in the analysis
only if the value of the FREQ column is greater than or equal to 1.

MAXPOINTS=positive-integer
specifies the maximum number of predicted points generated for the regression
curve as well as any confidence limits.

Default 201

WEIGHT=numeric-column
specifies a column in the input data set that contains values to be used as a
priori weights for a regression fit. If an observation’s weight is zero, negative, or
missing, then the observation is deleted from the analysis.

Interaction Starting with SAS 9.4M2, when the CLI= option is used with this
option, the confidence band for individual predicted values is
displayed as a high-low chart instead of a band.

Details
The REGRESSIONPLOT statement creates a fitted regression line or curve
computed from input data. It supports only models of one independent and one
dependent variable. Information about the fitting methodology that the
REGRESSIONPLOT statement uses can be found in The TRANSREG Procedure in
SAS/STAT User’s Guide.

In addition to the regression line, the REGRESSIONPLOT statement can compute

confidence levels for the fitted line. To display the confidence levels:

1 Use the CLI= or CLM= regression option(s) to declare a name for each
confidence level.

2 Use MODELBAND statements to refer to the name(s) and draw a confidence

band(s) from this information.

Example: REGRESSIONPLOT Statement

The following graph was generated by the “Example Program” on page 806:
 806 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph regressionplot;
begingraph;
entrytitle "Regression Fit Plot";
layout overlay;
scatterplot x=weight y=mpg_highway /
datatransparency=0.7;
regressionplot x=weight y=mpg_highway /
name="fitline"
alpha=0.05 legendlabel="Regression Fit";
discretelegend "fitline";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=regressionplot;

run;

SCATTERPLOT Statement
Creates a scatter plot of input data.
 SCATTERPLOT Statement 807

Syntax
SCATTERPLOT X=column | expression
Y=column | expression < /options >;

Summary of Optional Arguments

Appearance options
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction
of the midpoint spacing. On an interval axis, specifies the width of the
group clusters as a fraction of the minimum interval between adjacent
data values.
COLORMODEL=style-element | (color-list )
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with SAS 9.4M2, specifies the column or range attribute map
variable to use to determine the marker colors.
CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)
specifies whether space requirements for this plot contribute to the
calculation of the axis offsets.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the plot markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the markers, data labels, and
error bars, when displayed.
DISCRETEMARKERSIZE=number
specifies the size of a marker as a fraction of the tick spacing.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars that are associated with the data
points.
ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the error-bar caps.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one
of the GraphData1–GraphDataN style elements.
LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are
displayed in the plot.
808 Chapter 6 / Plot Statements

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.
MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of
marker symbols.
MARKERCHARACTERATTRS=style-element | style-element (text-options) |
(text-options)
specifies the color and font attributes of the marker character specified
on the MARKERCHARACTER= option.
MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |
TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters.
MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression
in SAS 9.4M1 and earlier releases, specifies the column or range attribute
map variable that is used to determine the marker colors.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
MARKERSIZEMAX=dimension
for SAS 9.4M1 and for earlier releases, specifies a drawing size for the
largest marker when the marker size represents response values.
MARKERSIZEMIN=dimension
for SAS 9.4M1 and for earlier releases, specifies a drawing size for the
smallest marker when the marker size represents response values.
MARKERSIZERESPONSE=numeric-column | expression
for SAS 9.4M1 and for earlier releases, specifies a column that is used to
map the drawing size of the markers.
OUTLINEDMARKERCHARACTERS=TRUE | FALSE
specifies whether the characters that are used as marker symbols are
outlined in order to enhance their appearance in the graph.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
SIZEMAX=dimension
starting with SAS 9.4M2, specifies a drawing size for the largest marker
when the marker size represents response values.
SIZEMIN=dimension
starting with SAS 9.4M2, specifies a drawing size for the smallest marker
when the marker size represents response values.
SIZERESPONSE=numeric-column | expression
starting with SAS 9.4M2, specifies a column that is used to map the
drawing size of the markers.
SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the
scatter plot is rendered.
USEDISCRETESIZE=TRUE | FALSE
 SCATTERPLOT Statement 809

specifies that the marker size should be based on fraction of the midpoint
spacing that is set by the DISCRETEMARKERSIZE= option.
XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error
bars are drawn from the markers to the endpoints.
XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars.
YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars.
YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over
the scatter points.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies a column for marker labels. The label positions are adjusted to
prevent them from overlapping.
DATALABELATTRS=style-element | style-element (text-options ) | (text-
options )
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.
DATALABELSPLIT=TRUE | FALSE
810 Chapter 6 / Plot Statements

specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all markers from discrete X values, or
discrete Y values, or both.
GROUP=column | discrete-attr-var | expression
creates a separate marker type for each unique group value of the
specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.
JITTER=NONE | AUTO
specifies whether to jitter data markers.
JITTEROPTS=(jitter-options)
specifies options for managing jittering.

ODS options
URL=string-column
specifies an HTML page to display when a point is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.

Required Arguments
X=column | expression
specifies the column for the X values.
 SCATTERPLOT Statement 811

Y=column | expression
specifies the column for the Y values.

Optional Arguments
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO
uses the discrete axis for clustering groups when only one axis is discrete.
Uses the X axis for clustering if both axes are discrete or interval.

X
Y
uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters
as a fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in

effect for interval data, the size of the markers in each cluster might
be reduced to no less than 5 pixels in order to display the cluster
within the smallest effective midpoint space. If you need larger
812 Chapter 6 / Plot Statements

markers in that case, use the MARKERATTRS= option to specify a

larger marker size.

COLORMODEL=style-element | (color-list )
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.

style-element
specifies the name of a style element.

The style element should contain these style attributes:

STARTCOLOR specifies the color for the smallest data value of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
NEUTRALCOLOR specifies the color for the midpoint of the range of
the COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
ENDCOLOR specifies the color for the highest data value of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.

(color-list)
specifies a space-separated list of colors to use in the color ramp.

You can use style attribute references such as GraphData3:Color, color

names, or RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color.
The list can contain a mix of style attribute references, color names, and
color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= or

MARKERCOLORGRADIENT= option must also be specified.

Tip The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with SAS 9.4M2, specifies the column or range attribute map variable to
use to determine the marker colors.

Note: Starting with SAS 9.4M2, the COLORRESPONSE= option replaces the
MARKERCOLORGRADIENT= option. The syntax and functionality are the same.
The MARKERCOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.
 SCATTERPLOT Statement 813

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

You can use this option to add a second response variable to an analysis. For
example, in an analysis of weight by height, you can specify an age column by
using the COLORRESPONSE= or MARKERCOLORGRADIENT= option so that
the change in the gradient color of the markers reflects the change in age.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the MARKERATTRS=

option and in the MARKERCHARACTERATTRS= option and varies
814 Chapter 6 / Plot Statements

the marker color according to the color gradient or the attribute

map.

Note You can use MARKERCOLORGRADIENT= as an alternative to

COLORRESPONSE=. However, if you use
MARKERCOLORGRADIENT=, be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize
MARKERCOLORGRADIENT as the color role, not
COLORRESPONSE.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Starting with SAS 9.4M2, when this option is in effect and error
bars are displayed, the error bars derive their color from the
markers. To set a fixed color for the error bars, use the
ERRORBARATTRS= option.

If the MARKERCHARACTER= option is also specified, then the

gradients that would be applied to the markers are applied to the
text strings.

CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)

specifies whether space requirements for this plot contribute to the calculation
of the axis offsets.

Note: This feature applies to SAS 9.4M2 and to later releases.

This plot’s layout container queries each of its plots for a preferred offset and
includes all of the offsets in the axis offset calculations. If the DATALABEL= or
MARKERCHARACTER= option is specified for this plot, this plot might request
a preferred offset that prevents the clipping of any data labels or marker
character strings that appear at the ends of the axes. The requested offset is
based on the longest string. If the label or marker character lengths vary
significantly, the result is wasted space when the shorter strings appear near the
ends of the axes. In that case, you can use the CONTRIBUTEOFFSETS= option
to modify or eliminate this plot’s contribution to the offset calculations in order
to reclaim that space.

ALL
the space requirements for this plot are contributed to the axis offset
calculations.

NONE
the space requirements for this plot are not contributed to the axis offset
calculations.

(axis-offset-list)
a space-delimited list of specific contributions that this plot makes to the
axis offset calculations.

The list is one or more of the following values, enclosed in parentheses:

 SCATTERPLOT Statement 815

XMAX
the space requirements for this plot are contributed to the X-axis offset
calculation for the maximum end.

XMIN
the space requirements for this plot are contributed to the X-axis offset
calculation for the minimum end.

YMAX
the space requirements for this plot are contributed to the Y-axis offset
calculation for the maximum end.

YMIN
the space requirements for this plot are contributed to the Y-axis offset
calculation for the minimum end.

Default ALL

Interaction Offsets that are set in the layout axis options are always honored,
regardless of the setting on this option.

Note This option does not affect offset requests from other plots.

DATALABEL=column | expression
specifies a column for marker labels. The label positions are adjusted to prevent
them from overlapping.

Default No data labels are displayed

Interactions If a numeric column is specified and the column has no format, then
a BEST6 format is applied.

This option is ignored if the MARKERCHARACTER= option is used,

which displays labels instead of the markers.

Tip Use this option to display labels for the markers. The position of
the labels is adjusted to prevent the labels from overlapping. If you
want labels displayed instead of markers, then use the
MARKERCHARACTER= option.
816 Chapter 6 / Plot Statements

DATALABELATTRS=style-element | style-element (text-options ) | (text-options )

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also
be specified.

This option is ignored if the MARKERCHARACTER= option is

specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to
avoid overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.

AUTO
uses a data-label collision avoidance algorithm to position the labels.

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.
 SCATTERPLOT Statement 817

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

Interactions This option has no effect if DATALABELPOSITION=AUTO or if

DATALABELSPLIT=FALSE.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels. The split
characters are specified by the DATALABELSPLITCHAR= option.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
818 Chapter 6 / Plot Statements

characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Interactions This option has no effect unless DATALABELSPLIT=TRUE.

The DATALABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

 SCATTERPLOT Statement 819

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Interaction This option has no effect if DATALABELPOSITION=AUTO or if

DATALABELSPLIT=FALSE.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the plot markers. The following figure shows
large CIRCLEFILLED markers with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.
820 Chapter 6 / Plot Statements

When a data skin is applied, all marker outlines are set by the skin,
and the outline settings are ignored from the ODS style or from
MARKERATTRS= option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the markers, data labels, and error
bars, when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

DISCRETEMARKERSIZE=number
specifies the size of a marker as a fraction of the tick spacing.

Default 0.5

Range 0 to 1

Requirement For this option to take effect, at least one of the axes must be
discrete.

Interactions If both of the axes are discrete, then the marker size is a fraction of
the smaller tick spacing.

If the X axis is a numeric interval axis and the

GROUPDISPLAY=CLUSTER option is in effect, then the marker
size is a fraction of the interval between the two closest but not
identical points in the X direction.

For this option to take effect, the USEDISCRETESIZE= option must

be set to TRUE (the default is FALSE).

DISCRETEOFFSET=number
specifies an amount to offset all markers from discrete X values, or discrete Y
values, or both. This feature is useful for graphing multiple response variables
side by side on a common axis. By default within an overlay-type layout, if a
SCATTERPLOT is used with other plots with a discrete axis, then the markers
are centered on the discrete X values, or discrete Y values, or both. Depending
on the data, the markers might be superimposed over other graph data. The
following code fragment shows the default positioning when a SCATTERPLOT
is used with a BOXPLOT:
layout overlay / cycleattrs=true
xaxisopts=(type=discrete);

scatterplot x=age y=weight;

boxplot x=age y=weight;

endlayout;
 SCATTERPLOT Statement 821

To avoid superimposed plots, you can assign a different offset to each plot
statement:
layout overlay / cycleattrs=true
xaxisopts=(type=discrete);
scatterplot x=age y=weight /
discreteoffset=0.2;
boxplot x=age y=weight /
discreteoffset=-0.2;
endlayout;

Default 0 (no offset, all markers are centered on the discrete X values, or
discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right on discrete X values
and up on discrete Y values. If the layout’s axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.
822 Chapter 6 / Plot Statements

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data
points.

Defaults For ungrouped data, the GraphError style element attributes

ContrastColor, LineStyle, and LineThickness.

For grouped data, the LineStyle and LineThickness attributes from

the GraphError style element, and the ContrastColor attribute from
the GraphData1–GraphDataN style elements.

Interaction For this option to take effect, error bars must be displayed by the
XERRORLOWER= , XERRORUPPER= , YERRORLOWER= , or
YERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the error-bar caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1

Interaction This option is ignored when ERRORBARCAPSHAPE=NONE is in

effect.

Tips Use the ERRORBARCAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the ERRORBARATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the error-bar caps:

errorbarcapscale=2

Reduce the size of the error-bar caps by 50%:

errorbarcapscale=0.5

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.

Defaults SERIF in SAS 9.4M5 and earlier releases.

 SCATTERPLOT Statement 823

Starting with SAS 9.4M2, GraphError:CapStyle style reference. If

attribute CapStyle is not defined in the active style, then SERIF is
the default value.

Interaction Starting with SAS 9.4M5, when ERRORBARCAPSCALE= is in effect,

the display width for SERIF caps is the cap's default width multiplied
by the value specified in ERRORBARCAPSCALE=.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.

TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.

When this option is TRUE, the fill color and outline color for filled markers
are determined in the following ways:
n If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and
the marker outlined color is derived from the GraphData1–GraphDataN
style elements ContrastColor attribute.
n If the GROUP= option is not specified, then by default, the fill color is
derived from the GraphDataDefault style elements Color attribute, and
the marker outlined color is derived from the GraphOutline style
elements ContrastColor attribute.
n If the COLORRESPONSE= MARKERCOLORGRADIENT= option is
specified, then the marker fill is drawn by using the mapped color that is
computed from the value of the COLORRESPONSE= or
MARKERCOLORGRADIENT= option for that observation. The marker
outline is drawn by using the MARKEROUTLINEATTRS= specification.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set
this option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 824

MARKERFILLATTRS= on page 835

MARKEROUTLINEATTRS= on page 835

COLORRESPONSE= on page 812

“boolean ” on page xii for other Boolean values that you can use.
824 Chapter 6 / Plot Statements

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
computation. The marker for that observation is drawn n times,
which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a separate marker type for each unique group value of the specified
column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color and marker symbol. Markers vary
according to the ContrastColor and MarkerSymbol attributes of the
GraphData1–GraphDataN and GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of markers and
colors.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

The marker size is set by the MARKERATTRS= option.

The MARKERCHARACTER= , COLORRESPONSE= , and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

The SIZERESPONSE= and MARKERSIZERESPONSE= options

override this option’s SIZE= setting.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.
 SCATTERPLOT Statement 825

Tips The representations that are used to identify the groups can be
overridden. For example, each distinct group value is represented
by a different marker symbol, but the
MARKERATTRS=(SYMBOL=marker) option could be used to assign
the same symbol to all of the plot’s marker symbols, letting marker
color indicate group values. Likewise,
MARKERATTRS=(COLOR=color) could be used to assign the same
color to all markers, letting marker symbol indicate group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.

OVERLAY
draws markers for a given group value at the exact coordinate.

Depending on the data, markers at a given coordinate might overlap.

CLUSTER
draws markers for a given group value adjacent to each other.

Default OVERLAY

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.
826 Chapter 6 / Plot Statements

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.
 SCATTERPLOT Statement 827

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interactions For this option to take effect, the GROUP= option must also be
specified.

The MARKERCHARACTER=, COLORRESPONSE=, and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

JITTER=NONE | AUTO
specifies whether to jitter data markers.

NONE
does not offset the data markers.

AUTO
the system determines the best way to display the data markers, based on
the settings specified by the JITTEROPTS= option.

The following figures show a side-by-side comparison of a BOXPLOT and

SCATTERPLOT in which the JITTER=NONE and JITTER=AUTO options are
specified in the SCATTERPLOT statements. The first figure shows the case in
which the Y axis is linear and the X axis is discrete.
828 Chapter 6 / Plot Statements

In this case, when jittering is disabled (JITTER=NONE), markers that represent

the same Y value are overlaid, which results in some markers being obscured.
When jittering is enabled (JITTER=AUTO), one-dimensional systematic jittering
occurs along the X axis. The markers that represent approximately the same Y
value are offset along the X axis from the midpoint of that value in order to
make all of the markers visible.

The next figure shows the case in which both axes are linear.

In this case, when jittering is disabled (JITTER=NONE), markers that represent

the same X and Y bin value are overlaid as in the previous case, which results in
some markers being obscured. However, unlike the previous case, when jittering
is enabled (JITTER=AUTO), two-dimensional random jittering occurs along the
X and the Y axes. The markers are offset randomly along both the X and Y axes
in order to make all of the markers visible.

Default NONE

Restrictions In SAS 9.4, jittering is not supported when

GROUPDISPLAY=CLUSTER. Starting with SAS 9.4M1, this
restriction is removed.
 SCATTERPLOT Statement 829

One-dimensional systematic jittering is not supported when the

MARKERCHARACTER= or the MARKERSIZERESPONSE= option is
in effect. Random jittering is supported in those cases when both
the X and Y axes are interval axes.

Interactions When jittering is enabled, the FREQ= option is ignored.

If this option is set to AUTO and both axes are discrete, then one-
dimensional systematic jittering occurs along the X axis.

Notes Jittering changes the default axis offsets, but it does not change
the axis data range.

When jittering is requested on a discrete axis and a large amount of

data is plotted, the jittering process can become resource-
intensive. In that case, it might take longer to render the plot.

You should not overlay a jittered scatter plot and a box plot with
outliers because their jittering offsets differ. A scatterplot's points
are jittered when the numeric values are close enough whereas a
box plot's points are jittered only when the numeric values are
identical.

JITTEROPTS=(jitter-options)
specifies options for managing jittering. The jitter options can be one or more of
the following values, separated by a space:

AXIS=AUTO | X | Y | BOTH
specifies the axis to use for jittering the data markers.

AUTO
the system determines the axis.

The axis is determined as follows:

n If the X axis is discrete, then one-dimensional systematic jittering is
applied along the X axis.
n If the Y axis is discrete and the X axis is interval, then one-dimensional
systematic jittering is applied along the Y axis.
n If the X and Y axes are interval, then random jittering is applied along
both the X and Y axes.

X
jittering is on the X axis.

If the X axis is discrete, then 1-dimensional systematic jittering is applied.

Otherwise, 1-dimensional random jittering is applied.

Note If both the X and Y axes are discrete, then specifying BOTH is
equivalent to specifying X.

Y
jittering is applied on the Y axis.

If the Y axis is discrete, then 1-dimensional systematic jittering is applied.

Otherwise, 1-dimensional random jittering is applied.
830 Chapter 6 / Plot Statements

BOTH
specifies that random jittering is applied on both the X and Y axes.

This option applies only when both the X and Y axes are interval.

Tip When the X or Y axis is discrete, specifying BOTH is equivalent to

specifying AUTO.

Default AUTO

Restriction When the MARKERCHARACTER= or the

MARKERSIZERESPONSE= option is in effect, 1-dimensional
systematic jittering is not supported. Random jittering is
supported in those cases when both the X and Y axes are interval
axes.

UNIFORM=TRUE | FALSE
specifies whether the jitter offsets are uniform across all category values or
are varied as needed.

Note: This feature applies to SAS 9.4M5 and to later releases.

When the markers for a category value do not fit the jittering space, the
jittering offsets are reduced in order to fit the markers for that category
value within the jittering space.

TRUE
uses a uniform jittering offset across all category values. The offset used
is the minimum offset that is required across all of the category values.

FALSE
varies the jittering offset only when the markers for a category value do
not fit the jittering space. The offset is varied as needed to fit the
markers for each case individually.

Default FALSE

Restriction This option applies to discrete-type jittering only.

Note When UNIFORM=TRUE is specified for a scatter plot within a

LAYOUT DATALATTICE or LAYOUT DATAPANEL, the jittering is
uniform across all potential category and group combinations.
 SCATTERPLOT Statement 831

WIDTH=positive-number
specifies the width of the jittering space as a fraction of either the midpoint
spacing or of the minimal interval width.

Defaults 0.85 for 1-dimensional systematic jittering

0.4 for random jittering on one or both axes.

Note For a discrete axis, this option takes effect only if the jittered
markers still overlap. Setting WIDTH= to a higher value increases
the jittering space up to the specified width or to a width sufficient
to eliminate overlap if it is less than the specified width. In the
latter case, increasing the value further has no effect.

Tip The specified number can be greater than 1.

Interaction This option is ignored when JITTER=NONE.

LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are displayed in
the plot. The MARKERCHARACTER= option specifies the column that provides
the marker strings that are to be used in place of marker symbols.

Default FALSE

Interactions This option effects marker strings only when the

MARKERCHARACTER= option is specified.

This option effects data labels only when DATALABEL= is specified

and DATALABELPOSITION= is not AUTO.

Tip Stripping the blanks from the numeric value strings helps center
each string relative to its data point. Stripping is useful when you
want to overlay the data values near or inside the markers for a
plot.

See “boolean ” on page xii for other Boolean values that you can use.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

832 Chapter 6 / Plot Statements

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If the COLORRESPONSE= or MARKERCOLORGRADIENT= option

is specified, then this option’s COLOR= suboption is ignored.

If the MARKERCHARACTER= option is specified, then this option’s

SYMBOL= and WEIGHT= suboptions are ignored.

If the MARKERSIZERESPONSE= option is specified, then this

option’s WEIGHT= suboption is ignored.

If FILLEDOUTLINEDMARKERS=TRUE, then this option’s COLOR=

suboption is ignored. In that case, to specify the marker fill color,
use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that
case, all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data. When the SYMBOL= suboption is specified in
that case, all markers have the same symbol, and the symbol color
alone distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of marker
symbols.
 SCATTERPLOT Statement 833

Each string is centered horizontally and vertically at the data point. The data
point positions are not adjusted to prevent text overlap.

Interactions This option overrides the DATALABEL= option.

If the GROUP= option is also used, then color is displayed for a

DISCRETE legend, but the character is not displayed in the legend.
For text strings, the same colors are applied to the text strings as
would have been applied to markers.

One-dimensional systematic jittering is not supported when this

option is in effect. Random jittering is supported when this option is
in effect and both the X and Y axes are interval axes. See JITTER=
on page 827.

Note If a numeric column is used, then its values are converted to strings
using the format associated with the column or using BEST6 if no
format is defined.

Tips You can use the MARKERCHARACTERPOSITION= option to

change the justification of the marker character.

You can use the OUTLINEDMARKERCHARACTERS= option to

enhance the appearance of the marker characters.

You can use the LABELSTRIP= option to strip the leading and
trailing blanks from numeric value strings in order to center each
string on its data point.

If the MARKERCHARACTER= variable contains one or more long

character strings, large axis offsets might result. In that case, use
the OFFSETMIN= and OFFSETMAX= suboptions in the
XAXISOPTS= and YAXISOPTS= axis options to override the
automatic axis offsets.

MARKERCHARACTERATTRS=style-element | style-element (text-options) |

(text-options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.
834 Chapter 6 / Plot Statements

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the MARKERCHARACTER= option

must also be used.

When the GROUP= option is also specified, each distinct group

value might be represented by a different color (depending on the
ODS style setting or the setting on the INDEX= option). The marker
character that is associated with the group is assigned the group
color.

This option’s COLOR= suboption can be used to specify a single

color for all marker characters in a grouped plot, without affecting
items that have a group color, such as error bars and marker
symbols.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |

TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters. The following figure shows
the effect of each of the values on the position of marker character M.

Note: The red dots represent the center in each case for reference.

Default CENTER

Interaction This option is ignored if the MARKERCHARACTER= option is not

specified.

Tip You can use the LABELSTRIP= option to strip the leading and
trailing blanks from value strings in order to properly justify each
string on its data point.
 SCATTERPLOT Statement 835

MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression

in SAS 9.4M1 and earlier releases, specifies the column or range attribute map
variable that is used to determine the marker colors.

Note: Starting with SAS 9.4M2, the MARKERCOLORGRADIENT= option is

deprecated and replaced with the COLORRESPONSE= option. The syntax and
functionality are the same. The MARKERCOLORGRADIENT= option is still
honored, but the COLORRESPONSE= option is preferred.

Note Starting with SAS 9.4M2, if you use the MARKERCOLORGRADIENT=

option, then be aware that the TIP=, TIPFORMAT=, and TIPLABEL=
options recognize the MARKERCOLORGRADIENT role and not the
COLORRESPONSE role.

See COLORRESPONSE= on page 812

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element

For grouped data, the Color attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interactions This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

When the COLORRESPONSE= or MARKERCOLORGRADIENT=

option is in effect, this option’s COLOR= specification is ignored.

Note When style-element is specified, only the style element’s Color

attribute is used.

See “General Syntax for Attribute Options” on page 1599

“Fill Color Options” on page 1600

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

836 Chapter 6 / Plot Statements

Interaction This option is ignored when a data skin is applied by the current
style or by the DATASKIN= option. In the latter case, the outline is
set by the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1599

“Line Options” on page 1602

MARKERSIZEMAX=dimension
for SAS 9.4M1 and for earlier releases, specifies a drawing size for the largest
marker when the marker size represents response values.

Note: Starting with SAS 9.4M2, the SIZEMAX= option replaces the
MARKERSIZEMAX= option. The syntax and functionality are the same. The
MARKERSIZEMAX= option is still honored, but the SIZEMAX= option is
preferred.

See SIZEMAX= on page 839

“dimension” on page xiii

MARKERSIZEMIN=dimension
for SAS 9.4M1 and for earlier releases, specifies a drawing size for the smallest
marker when the marker size represents response values.

Note: Starting with SAS 9.4M2, the SIZEMIN= option replaces the
MARKERSIZEMIN= option. The syntax and functionality are the same. The
MARKERSIZEMIN= option is still honored, but the SIZEMIN= option is
preferred.

See SIZEMIN= on page 839

“dimension” on page xiii

MARKERSIZERESPONSE=numeric-column | expression
for SAS 9.4M1 and for earlier releases, specifies a column that is used to map
the drawing size of the markers.

Note: Starting with SAS 9.4M2, the SIZERESPONSE= option replaces the
MARKERSIZERESPONSE= option. The syntax and functionality are the same.
The MARKERSIZERESPONSE= option is still honored, but the SIZERESPONSE=
option is preferred.

Interaction One-dimensional systematic jittering is not supported when this

option is in effect. Random jittering is supported when this option is
 SCATTERPLOT Statement 837

in effect and both the X and Y axes are interval axes. See JITTER=
on page 827.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEDMARKERCHARACTERS=TRUE | FALSE
specifies whether the characters that are used as marker symbols are outlined
in order to enhance their appearance in the graph.

Note: This option is valid starting in SAS 9.4M1.

Note: This feature is deprecated starting with SAS 9.4M2. The

OUTLINEDMARKERCHARACTERS= option is still honored, but the TEXTPLOT
statement is now the preferred method of creating a scatter plot using text
markers.

The following figure shows the marker characters M and F displayed when
OUTLINEDMARKERCHARACTERS=FALSE (default) and when
OUTLINEDMARKERCHARACTERS=TRUE.

Default FALSE

Restriction Outline marker characters are not supported in vector graphics

output. When this option is set and vector graphics output is
requested, the graph is converted into an image instead. A note
indicating the conversion is written to the SAS log. To restore the
vector graphics output in that case, remove the
OUTLINEDMARKERCHARACTERS= option from the
SCATTERPLOT statement.

Interaction The MARKERCHARACTER= option must be specified for this option

to have any effect.
838 Chapter 6 / Plot Statements

See “boolean ” on page xii for other Boolean values that you can use.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, MARKERCHARACTER,
COLORRESPONSE or MARKERCOLORGRADIENT,
XERRORLOWER, XERRORUPPER, YERRORLOWER,
YERRORUPPER, GROUP, and INDEX.
 SCATTERPLOT Statement 839

SIZEMAX=dimension
starting with SAS 9.4M2, specifies a drawing size for the largest marker when
the marker size represents response values.

Note: Starting with SAS 9.4M2, the SIZEMAX= option replaces the
MARKERSIZEMAX= option. The syntax and functionality are the same. The
MARKERSIZEMAX= option is still honored, but the SIZEMAX= option is
preferred.

Default Three times the GraphDataDefault:MarkerSize setting (typically

7px) or 21px.

Restriction The SIZEMAX= value must be greater than the SIZEMIN= or

MARKERSIZEMIN= value.

Interactions For this option to take effect, the SIZERESPONSE= or

MARKERSIZERESPONSE= option must also be used.

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use
the TEXTPLOT statement.

Tip If you specify this size as a percent, then the specified value is
interpreted as a percent of the graph’s height. You can control the
height by using the DESIGNHEIGHT= option of the BEGINGRAPH
statement, or by using the HEIGHT= option of the ODS GRAPHICS
statement. For a standard 640px by 480px output size, a
percentage value of 4.5% sets a maximum size of about 21px, which
is approximately the same marker size that would result from this
option’s typical default setting.

See “dimension” on page xiii

SIZEMIN=dimension
starting with SAS 9.4M2, specifies a drawing size for the smallest marker when
the marker size represents response values.

Note: Starting with SAS 9.4M2, the SIZEMIN= option replaces the
MARKERSIZEMIN= option. The syntax and functionality are the same. The
MARKERSIZEMIN= option is still honored, but the SIZEMIN= option is
preferred.

Default The GraphDataDefault:MarkerSize setting, which is typically 7px.

Restriction The SIZEMIN= value must be less than the SIZEMAX= or

MARKERSIZEMAX= value.

Interactions For this option to take effect, you must also specify the
SIZERESPONSE= or MARKERSIZERESPONSE= option.
840 Chapter 6 / Plot Statements

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use
the TEXTPLOT statement.

Tip If you specify this size as a percent, then the specified value is
interpreted as a percent of the graph’s height. You can control the
height by using the DESIGNHEIGHT= option of the BEGINGRAPH
statement, or by using the HEIGHT= option of the ODS GRAPHICS
statement. For a standard 640px by 480px output size, a
percentage value of 1.5% sets a minimum size of about 7px, which is
approximately the same marker size that would result from this
option’s typical default setting.

See “dimension” on page xiii

SIZERESPONSE=numeric-column | expression
starting with SAS 9.4M2, specifies a column that is used to map the drawing size
of the markers.

Note: Starting with SAS 9.4M2, the SIZERESPONSE= option replaces the
MARKERSIZERESPONSE= option. The syntax and functionality are the same.
The MARKERSIZERESPONSE= option is still honored, but the SIZERESPONSE=
option is preferred.

By default, the minimum and maximum values of this column establish a range
over which the marker sizes vary in linear proportion. The actual drawing size of
the smallest marker and the largest marker is set automatically.

Default The GraphDataDefault:MarkerSize setting, which is typically 7px.

Interactions This option overrides the SIZE= setting in the MARKERATTRS=

option.

This option is ignored when the MARKERCHARACTER= option is

used. To manage the size of marker characters in a scatter plot, use
the TEXTPLOT statement.

Tip You can adjust the smallest and largest marker size with the
SIZEMIN= and SIZEMAX= options, or with the MARKERSIZEMIN=
and MARKERSIZEMAX= options.

SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the scatter
plot is rendered.

Note: This feature applies to SAS 9.4M3 and to later releases.

AUTO
sets subpixel rendering OFF, unless SUBPIXEL=ON is specified in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. In that case,
subpixel rendering is ON.
 SCATTERPLOT Statement 841

OFF
disables subpixel rendering for image output only for this scatter plot.

Default AUTO

Requirement Anti-aliasing must be enabled for this option to have any effect.

Notes This option affects subpixel rendering only for this plot. Subpixel
rendering for other plots in the graph is not affected.

For vector-graphics output, this option is ignored, and subpixel

rendering is always enabled.

When subpixel rendering is used for the graph but is turned OFF
for this scatter plot, some elements in the scatter plot such as the
plot markers might be offset a half pixel. This can make the plot
markers appear blurry in the image output.

Tips Anti-aliasing is enabled by default. If anti-aliasing has been

disabled, use the ANTIALIAS=ON option in the ODS GRAPHICS
statement to re-enable it.

Anti-aliasing is disabled automatically for this plot when the

resources required to anti-alias it exceed a preset threshold. When
anti-aliasing is disabled for this or any other plot in the graph,
subpixel rendering is disabled for the entire graph. A note is written
to the SAS log that provides information about how to use the
ANTIALIASMAX= option in an ODS GRAPHICS statement to re-
enable anti-aliasing.

See “Using Subpixel Rendering” in SAS Graph Template Language: User’s

Guide

“ODS GRAPHICS Statement” on page 1561 for information about

the ANTIALIAS=, ANTIALIASMAX=, and SUBPIXEL= options.

TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over the
scatter points. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the scatter
plot can be specified along with roles that do.

(role-list)
an ordered, space-separated list of unique SCATTERPLOT and user-defined
roles.

SCATTERPLOT roles include X , Y , DATALABEL , MARKERCHARACTER ,

COLORRESPONSE or MARKERCOLORGRADIENT , XERRORLOWER ,
XERRORUPPER , YERRORLOWER , YERRORUPPER , GROUP , and INDEX .

User-defined roles are defined with the ROLENAME= option.

Requirement If you use the COLORRESPONSE= option, you must specify the
COLORRESPONSE role for the color values. Likewise, if you
use the MARKERCOLORGRADIENT= option, you must specify
the MARKERCOLORGRADIENT role for the color values.
842 Chapter 6 / Plot Statements

Although they are functionally the same, you cannot mix the
COLORRESPONSE= and MARKERCOLORGRADIENT= options,
and their corresponding roles.

Example The following example displays data tips for the columns
assigned to the roles X, XERRORUPPER, and XERRORLOWER,
as well as the column Obs, which is not assigned to any pre-
defined SCATTERPLOT role. The Obs column must first be
assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X XERRORUPPER XERRORLOWER)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X , Y , DATALABEL ,
MARKERCHARACTER , COLORRESPONSE or
MARKERCOLORGRADIENT , XERRORLOWER , XERRORUPPER ,
YERRORLOWER , YERRORUPPER , FREQ , and GROUP .

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)
 SCATTERPLOT Statement 843

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a point is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each marker that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable markers, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pairs. In that case,
the same action is taken when the points for those X and Y pairs
are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

USEDISCRETESIZE=TRUE | FALSE
specifies that the marker size should be based on fraction of the midpoint
spacing that is set by the DISCRETEMARKERSIZE= option.
844 Chapter 6 / Plot Statements

Default FALSE

See DISCRETEMARKERSIZE=

“boolean ” on page xii for other Boolean values that you can use.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interactions This option is ignored if the X= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error bars are
drawn from the markers to the endpoints.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars. The error bars are
drawn from the markers to the endpoints.

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
 SCATTERPLOT Statement 845

Default Y

Interactions This option is ignored if the Y= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars. The error bars are
drawn from the markers to the endpoints.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars. The error bars are
drawn from the markers to the endpoints.

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

Examples

Example 1: Grouped Scatter Plot

The following graph was generated by the “Example Program” on page 846:
 846 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";
layout overlay;
scatterplot x=height y=weight /
group=sex name="scatter" datalabel=name;
discretelegend "scatter";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=scatterplot;

run;

Example 2: Discrete Scatter Plot

The following graph was generated by the “Example Program” on page 847:
 SCATTERPLOT Statement 847

Example Program
/* Summarize SASHELP.CARS for mean MPG_HIGHWAY */
proc summary data=sashelp.cars nway;
class type;
var mpg_highway;
output out=mileage mean=mean;
run;

/* Sort by mileage */
proc sort data=mileage;
by mean;
run;

/* Define the graph template */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Average Highway MPG By Type";
layout overlay /
xaxisopts=(griddisplay=on gridattrs=(color=lightgray))
yaxisopts=(griddisplay=on gridattrs=(color=lightgray)
linearopts=(minorgrid=true minortickcount=9
minorgridattrs=(color=lightgray pattern=dot)));
scatterplot x=type y=mean /
group=type groupdisplay=cluster
markerattrs=graphDataDefault;
endlayout;
endgraph;
 848 Chapter 6 / Plot Statements

end;
run;

/* Render the graph */

proc sgrender data=mileage template=scatterplot;
run;

Details
This example creates a scatter plot of average highway mileage by vehicle type.
The SUMMARY procedure is used to compute the mean highway mileage for each
vehicle type. The SORT procedure is then used to sort the data by mileage in
ascending order. Normally, when you plot discrete values using the SCATTERPLOT
statement, the values at each end of the category axis are offset to accommodate
the width of the largest marker. When a small symbol is used, this can result in the
minimum and maximum symbols being placed very close to the edge of the axis.
You can use the OFFSETMAX= and OFFSETMIN= axis options to specify offsets
for each end of the category axis. However, you have to determine an appropriate
offset value.

In this example, the GROUP= and GROUPDISPLAY= options are used to create
offsets automatically. The GROUP= option creates a group for each category value.
The GROUPDISPLAY=CLUSTER option displays the group values in a cluster and
automatically offsets the end values by one-half of the available midpoint spacing,
which is similar to the offsets that are used in bar charts. To restore the default
colors and symbols that are used for ungrouped data, the MARKERATTRS= option
is used to set the marker attributes to the graphDataDefault style element.

Example 3: Dot Plot

The following graph was generated by the “Example Program” on page 847:
 SCATTERPLOT Statement 849

Example Program
proc summary data=sashelp.class nway;
var height;
class age;
output out=classSum mean=mean std=std;
run;

proc template;
define statgraph dotplot;
begingraph;
entrytitle "Mean Height by Age";
layout overlay /
xaxisopts=(label="Height (Mean)")
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot)
reverse=true);
scatterplot x=mean y=age / name="scatter"
markerattrs=(symbol=circlefilled)
xerrorupper=eval(mean + std)
xerrorlower=eval(mean - std)
legendlabel="Height (Mean), 1 Standard Deviation";
discretelegend "scatter";
endlayout;
endgraph;
end;
run;

proc sgrender data=classSum template=dotplot;

where age < 16;
run;
 850 Chapter 6 / Plot Statements

Details
This example uses the SCATTERPLOT statement to create a dot plot from data
computed by the SUMMARY procedure. The SUMMARY procedure computes the
mean and standard deviation for variable Height by variable Age in data set
Sashelp.Class, and stores the results in data set Work.classSum. In the graph
template code, the SCATTERPLOT statement plots variables Mean and Age in data
set Work.classSum. The MARKERATTRS= option specified filled circles as the plot
dots. The XERRORUPPER= and XERRORLOWER= options specify the statistic
limits. The EVAL function computes the limits from variables Mean and Std.

SCATTERPLOTMATRIX Statement
Creates a matrix of all pairwise scatter plots of the specified columns.

Restriction: The SCATTERPLOTMATRIX statement cannot appear within a LAYOUT OVERLAY,

LAYOUT OVERLAY3D, or LAYOUT OVERLAYEQUATED block. It is typically placed
in a LAYOUT GRIDDED block.

Syntax
SCATTERPLOTMATRIX numeric-column-list </options>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with SAS 9.4M2, specifies the column or range attribute map
variable to use to determine the marker colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the plot markers.
DATATRANSPARENCY=number
specifies the degree of the transparency of the markers and data labels.
DIAGONAL=LABEL | (graph-list)
specifies whether the diagonal cells of the matrix are labeled with the
labels (names) of the required arguments, or with a plot.
INDEX=positive-integer-column | expression
 SCATTERPLOTMATRIX Statement 851

specifies indices for mapping marker attributes (color and symbol) to one
of the GraphData1–GraphDataN style elements.
LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are
displayed in the plot.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the data markers.
MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of
marker symbols.
MARKERCHARACTERATTRS=style-element | style-element (text-options) |
(text-options)
specifies the color and font attributes of the marker character specified
on the MARKERCHARACTER= option.
MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |
TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters.
MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression
in SAS 9.4M1 and earlier releases, specifies the column or range attribute
map variable that is used to determine the marker colors.
MATRIXTYPE=FULL | UPPERTRIANGLE | LOWERTRIANGLE
specifies whether to display the full matrix, or just the upper or lower
triangle of the matrix.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
START=TOPLEFT | BOTTOMLEFT
specifies whether to start populating the matrix from the top left or
bottom left corner.
SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the
scatter plots are rendered.
WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.
WALLDISPLAY=STANDARD | ALL | NONE | (display-options)
specifies whether the plot’s wall and wall outline are displayed.

Confidence options
ELLIPSE=(<ellipse-suboptions>)
specifies that a confidence ellipse be included in each cell containing a
scatter plot.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
852 Chapter 6 / Plot Statements

specifies the information to display when the cursor is positioned over

the scatter points.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Inset options
INSET=(info-options)
specifies what information is displayed in an inset.
INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information.

Label options
DATALABEL=column
specifies a column for marker labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of scatter markers, error bars, and data labels for
each unique group value of the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
CORROPTS=(correlation-options)
 SCATTERPLOTMATRIX Statement 853

specifies options for computing measures of association between pairs of

columns.
FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each
observation that is read.
ROWVARS=(numeric-column-list)
specifies a secondary list of numeric columns to be paired with the
required column list that is specified by the SCATTERPLOTMATRIX
statement.

Required Argument
numeric-column-list
specifies a list of numeric columns to plot.

Requirements There must be at least two columns to produce a useful matrix.

All of the columns must be numeric.

Note The default width is 640px, and the default height is 480px. The
graph size is not automatically adjusted to accommodate a large
number of columns.

Tips To change the graph size for the current template, use the
DESIGNHEIGHT= and DESIGNWIDTH= options in the
BEGINGRAPH statement.

To change the graph size for all templates in the current SAS
session, use the HEIGHT= and WIDTH= options in the ODS
GRAPHICS statement. Size settings in the ODS GRAPHICS
statement override size settings in the BEGINGRAPH statement.

You can also limit the number of columns in the matrix (perhaps
to seven in each dimension, for example) so that the resulting
graphs are not too small to be useful.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
MARKERCOLORGRADIENT= option.

style-element
specifies the name of a style element.

The style element should contain these style attributes:

STARTCOLOR specifies a color for the smallest data value of the

COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
NEUTRALCOLOR specifies a color for the midpoint of the range of the
COLORRESPONSE= or
MARKERCOLORGRADIENT= column.
854 Chapter 6 / Plot Statements

ENDCOLOR specifies a color for the highest data value of the

COLORRESPONSE= or
MARKERCOLORGRADIENT= column.

(color-list)
specifies a space-separated list of colors to use in the color ramp.

You can use style attribute references such as GraphData3:Color, color

names, or RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color.
The list can contain a mix of style attribute references, color names, and
color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interactions For this option to take effect, the COLORRESPONSE= or

MARKERCOLORGRADIENT= option must also be specified.

The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with SAS 9.4M2, specifies the column or range attribute map variable to
use to determine the marker colors.

Note: Starting with SAS 9.4M2, the COLORRESPONSE= option replaces the
MARKERCOLORGRADIENT= option. The syntax and functionality are the same.
The MARKERCOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGEALTCOLOR= or

RANGEALTCOLORMODEL= option in the RANGE statement
determines the marker colors.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
 SCATTERPLOTMATRIX Statement 855

You can use this option to add a second response variable to an analysis. For
example, in an analysis of weight by height, you can specify an age column by
using the COLORRESPONSE= or MARKERCOLORGRADIENT= option so that
the change in the gradient color of the markers reflects the change in age.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

Suboption COLOR= in the MARKERATTRS= option or in the

MARKERCHARACTERATTRS= option overrides this option for the
marker colors.

Note You can use MARKERCOLORGRADIENT= as an alternative to

COLORRESPONSE=. However, if you use
MARKERCOLORGRADIENT=, be aware that the TIP=,
TIPFORMAT=, and TIPLABEL= options recognize
MARKERCOLORGRADIENT as the color role, not
COLORRESPONSE.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

If the MARKERCHARACTER= option is also specified, then the

gradients that would be applied to the markers are applied to the
text strings.

CORROPTS=(correlation-options)
specifies options for computing measures of association between pairs of
columns. The following correlation-options are available:
856 Chapter 6 / Plot Statements

EXCLNPWGT=TRUE | FALSE
specifies whether observations with non-positive weight values are
excluded (TRUE) from the analysis.

Default FALSE (observations with negative weights are treated like those
with zero weights and counted in the total number of observations).

See “boolean ” on page xii for other Boolean values that you can use.

NOMISS=TRUE | FALSE
specifies whether observations with missing values are excluded (TRUE)
from the analysis.

Default FALSE (correlation statistics are computed using all of the

nonmissing pairs of columns).

Note Using NOMISS=TRUE is computationally more efficient.

See “boolean ” on page xii for other Boolean values that you can use.

WEIGHT=numeric-column
specifies a weighting column to use in the calculation of Pearson weighted
product-moment correlation.

The observations with missing weights are excluded from the analysis. If you
use this WEIGHT correlation option, then consider which value of the
VARDEF=correlation option is appropriate.

Default For observations with non-positive weights, the weights are set to
zero and the observations are included in the analysis.

Tip You can include EXCLNPWGT among the correlation options to

exclude observations with negative or zero weights from the
analysis.

VARDEF=DF | N | WDF | WEIGHT

specifies the variance divisor in the calculation of variances and covariances.

DF
Degrees of Freedom (N – 1)

N
number of observations

WDF
sum of weights minus 1 (WEIGHT – 1)

WEIGHT
sum of weights

Default DF

Interaction This option has effect only when the INSET= option is also used.

See the CORR procedure information in Base SAS Procedures Guide:

Statistical Procedures for the statistical and computational details of
these options.
 SCATTERPLOTMATRIX Statement 857

DATALABEL=column
specifies a column for marker labels. The label positions are adjusted to prevent
the labels from overlapping.

Interactions If a numeric column is specified and the column has no format, then
a BEST6 format is applied.

This option is ignored if the MARKERCHARACTER= option is used.

Note The position of the labels are adjusted to prevent the labels from
overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also
be specified.

This option is ignored if the MARKERCHARACTER= option is

specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to
avoid overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the markers.

AUTO
uses a data-label collision avoidance algorithm to position the labels.

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

858 Chapter 6 / Plot Statements

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.
 SCATTERPLOTMATRIX Statement 859

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.
860 Chapter 6 / Plot Statements

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELSPLIT=FALSE, or if

DATALABELSPLIT=TRUE and DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the plot markers. The following figure shows
large CIRCLEFILLED markers with each of the skins applied.
 SCATTERPLOTMATRIX Statement 861

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

When a data skin is applied, all marker outlines are set by the skin,
and the outline settings from the ODS style or from the
MARKERATTRS= option are ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the markers and data labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

Note This option does not apply to the HISTOGRAM, DENSITY, and
NORMAL plots in the scatter-plot matrix.

DIAGONAL=LABEL | (graph-list)
specifies whether the diagonal cells of the matrix are labeled with the labels
(names) of the required arguments, or with a plot. The plot for each diagonal
cell consists of an overlay combination of a histogram, normal, or kernel curves.

LABEL
labels the diagonal cells of the matrix with the names of the required
arguments.
862 Chapter 6 / Plot Statements

(graph-list)
specifies the plots that are displayed in the diagonal cells of the matrix.

The graph-list is a space-separated list of one or more of the following:

HISTOGRAM
specifies a histogram.

NORMAL
specifies a normal density curve.

KERNEL
specifies a kernel density estimate.

Default LABEL. Column labels (or names) are displayed in the diagonal
cells.

Requirement When specifying multiple graphs in the graph-list, you must

separate the values with a space. For example, the following
specification requests both a histogram and a normal density curve
in each diagonal cell:

DIAGONAL=(HISTOGRAM NORMAL)

Interactions The computation for HISTOGRAM, NORMAL, and KERNEL is

always computed on all the data for the current column (including
the FREQ= column, if used). The GROUP= option is not considered
in any of these computations.

This option is ignored if the ROWVARS= option is used.

Notes When this option is specified, the labels are drawn around the
outside of the matrix, and the matrix axes are dropped.

When NORMAL and KERNEL are specified, both curves have the
same visual attributes.

ELLIPSE=(<ellipse-suboptions>)
specifies that a confidence ellipse be included in each cell containing a scatter
plot. The ellipse is always drawn behind the scatter points. Ellipse-suboptions
can include one or more of the following:

ALPHA=positive-number
specifies the confidence level to compute for each ellipse.

Default 0.05

Range 0 < number < 1

Tip ALPHA=0.05 represents a 95% confidence level.

CLIP=TRUE | FALSE
specifies whether the X and Y values for the ellipse are considered when
determining the data ranges for the axes.
 SCATTERPLOTMATRIX Statement 863

TRUE
the X and Y values for the ellipses are ignored when the axis ranges are
determined.

Clipping occurs if the X and Y values for an ellipse exceed the axis range.

FALSE
the X and Y values for the ellipses contribute to the data range for each
axis.

If necessary, each axis is extended in order to display the entire ellipse.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

TYPE=MEAN | PREDICTED
specifies the type of ellipse.

MEAN
specifies a confidence ellipse of the mean

PREDICTED
specifies a prediction ellipse of the data

Default MEAN

See “ELLIPSE Statement” on page 504 for statistical details about how
the ellipse is calculated.

Default TYPE=MEAN ALPHA=0.05 You can set defaults by specifying the

option without arguments: ELLIPSE=( ).

Interactions The ellipse might be clipped by the data range for the scatter
points.

The ellipse is always computed on all the data for the current pair
of X and Y columns (including the FREQ= column, if used). The
GROUP= option is not considered when computed the ellipse.

Tip The display properties of each ellipse are controlled by the style
elements. The GraphDataDefault element controls the outline and
fill properties, and the GraphEllipse element controls whether the
outline, fill, or both are shown.

FREQ=numeric-column | expression
specifies a numeric column that provides frequencies for each observation that
is read.

Default All observations have a frequency count of 1.

Restriction If the value of numeric-column is missing or is less than 1 for an

observation, then that observation is not used in the analysis. If the
value is not an integer, then only the integer portion is used.

Note If n is the value of the numeric column for a given observation, then
that observation is used n times for the purposes of any statistical
864 Chapter 6 / Plot Statements

computation. The marker for that observation is drawn n times,

which increases the marker weight slightly with frequency.

GROUP=column | discrete-attr-var | expression

creates a distinct set of scatter markers, error bars, and data labels for each
unique group value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the graph by a

different combination of color and marker symbol. Markers vary
according to the ContrastColor and MarkerSymbol attributes of the
GraphData1–GraphDataN and GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of markers and
colors.

The marker size is set by the MARKERATTRS= option.

The MARKERCHARACTER= , COLORRESPONSE= , and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden. For example, each distinct group value is represented
by a different marker symbol, but the
MARKERATTRS=(SYMBOL=marker) option could be used to assign
the same symbol to all of the plot’s marker symbols, letting marker
color indicate group values. Likewise,
MARKERATTRS=(COLOR=color) could be used to assign the same
color to all markers, letting marker symbol indicate group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.
 SCATTERPLOTMATRIX Statement 865

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker attributes (color and symbol) to one of the
GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interactions For this option to take effect, the GROUP= option must also be
specified.

The MARKERCHARACTER=, COLORRESPONSE=, and

MARKERCOLORGRADIENT= options override the group settings
for the marker symbol and marker color.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.
866 Chapter 6 / Plot Statements

INSET=(info-options)
specifies what information is displayed in an inset. Insets appear in all cells of
the matrix except the diagonal and are displayed as a small table of name-value
pairs. The following info-options are available:

NOBS
displays the total number of observations where both the X and Y columns
have nonmissing values.

If the FREQ= option is used, then this number is adjusted accordingly. The
value of NOBS can be further adjusted by the use of the NOMISS=,
WEIGHT=, and EXCLNPWGT= suboptions of the CORROPTS= option.

PEARSON
displays the Pearson product-moment correlation.

The computation of the correlation is affected by the FREQ= and

CORROPTS= options. The computation is not done on a per group value
when GROUP= is used.

PEARSONPVAL
displays the probability value for the Pearson product-moment correlation.

Tip The location and appearance of the inset is controlled by the

INSETOPTS= option.

See PROC CORR in the documentation for Base SAS for statistical and
computational details of these options.

Example Here is an example of a typical inset:

N 150
r 0.96287
p(r) <.0001
In this example, NOBS is represented by N, PEARSON is represented
by r, and PEARSONPVAL is represented by p(r).

INSETOPTS=(appearance-options)
specifies location and appearance options for the inset information. The
appearance options can be any one or more of the settings that follow. The
options must be enclosed in parentheses, and each option is specified as a
name=value pair.

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the inset is automatically aligned within the layout.

NONE
does not automatically align the inset.

The inset’s position is therefore set by the HALIGN= and

VALIGN=appearance-options.

AUTO
attempts to center the inset in the area that is farthest from any
surrounding markers.

Data cells might have different inset placements.

 SCATTERPLOTMATRIX Statement 867

(location-list)
restricts the inset’s possible locations to those locations in the specified
location-list, and uses the location-list position that least collides with
the data cell’s other graphics features.

The location-list is space-separated and can contain any of these

locations: TOPLEFT TOP TOPRIGHT LEFT CENTER RIGHT
BOTTOMLEFT BOTTOM BOTTOMRIGHT. Example:
AUTOALIGN=(TOPRIGHT TOPLEFT)

Default NONE

Interaction When AUTOALIGN=AUTO or (location-list), the enclosing layout

statement’s HALIGN= and VALIGN=appearance-options are
ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the inset background

style-reference
specifies a style reference in the form style-element : style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default GraphWalls:Color style reference

BORDER=TRUE | FALSE
specifies whether a border is displayed around the inset.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

HALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment of the inset.

Default LEFT

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

OPAQUE=TRUE | FALSE
specifies whether the inset background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not
used.

See “boolean ” on page xii for other Boolean values that you can use.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the entire inset.

Default The GraphDataText style element.

868 Chapter 6 / Plot Statements

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TITLE="string"
specifies a title for the inset.

The title is added at the top of the inset and spans the full inset width.

Note Space is not reserved for the title when this option is not set

Tip Text properties for the title string can be set with TITLEATTRS=.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the inset’s title string.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment of the inset.

Default TOP

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

LABELSTRIP=TRUE | FALSE
specifies whether leading and trailing blanks are stripped from marker
characters or data labels that have a fixed position before they are displayed in
the plot. The MARKERCHARACTER= option specifies the column that provides
the marker strings that are to be used in place of marker symbols.

Default FALSE

Interactions This option effects marker strings only when the

MARKERCHARACTER= option is specified.

This option effects data labels only when DATALABEL= is specified

and DATALABELPOSITION= is not AUTO.

Tip Stripping the blanks from the numeric value strings helps center
each string relative to its data point. Stripping is useful when you
want to overlay the data values near or inside the markers for a
plot.

See “boolean ” on page xii for other Boolean values that you can use.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.
 SCATTERPLOTMATRIX Statement 869

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If the COLORRESPONSE= or MARKERCOLORGRADIENT= option

is specified, then this option’s COLOR= suboption is ignored.

If the MARKERCHARACTER= option is specified, then this option’s

SYMBOL= and WEIGHT= suboptions are ignored.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that
case, all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data. When the SYMBOL= suboption is specified in
that case, all markers have the same symbol, and the symbol color
alone distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MARKERCHARACTER=column | expression
specifies a column that defines strings that are to be used instead of marker
symbols.

Each string is centered horizontally and vertically at the data point. The data
point positions are not adjusted to prevent text overlap.
870 Chapter 6 / Plot Statements

Interactions This option overrides the DATALABEL= option.

If the GROUP= option is also used, then color is displayed for a

DISCRETE legend, but the character is not displayed in the legend.
For text strings, the same colors are applied to the text strings as
would have been applied to markers.

Note If a numeric column is used, then its values are converted to strings
using the format associated with the column or using BEST6 if no
format is defined.

Tips You can use the MARKERCHARACTERPOSITION= option to

change the justification of the marker character.

You can use the LABELSTRIP= option to strip the leading and
trailing blanks from numeric value strings in order to center each
string on its data point.

If the MARKERCHARACTER= variable contains one or more long

character strings, large axis offsets might result. In that case, use
the OFFSETMIN= and OFFSETMAX= suboptions in the
XAXISOPTS= and YAXISOPTS= axis options to override the
automatic axis offsets.

MARKERCHARACTERATTRS=style-element | style-element (text-options) |

(text-options)
specifies the color and font attributes of the marker character specified on the
MARKERCHARACTER= option.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the MARKERCHARACTER= option

must also be used.

When the GROUP= option is also specified, each distinct group

value might be represented by a different color (depending on the
ODS style setting or the setting on the INDEX= option). The marker
character that is associated with the group is assigned the group
color.

This option’s COLOR= suboption can be used to specify a single

color for all marker characters in a grouped plot, without affecting
items that have a group color, such as error bars and marker
symbols.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

 SCATTERPLOTMATRIX Statement 871

MARKERCHARACTERPOSITION=CENTER | TOP | BOTTOM | LEFT | RIGHT |

TOPLEFT | TOPRIGHT | BOTTOMLEFT | BOTTOMRIGHT
specifies the justification of the marker characters. The following figure shows
the effect of each of the values on the position of marker character M.

Note: The red dots represent the center in each case for reference.

Default CENTER

Interaction This option is ignored if the MARKERCHARACTER= option is not

specified.

Tip You can use the LABELSTRIP= option to strip the leading and
trailing blanks from value strings in order to properly justify each
string on its data point.

MARKERCOLORGRADIENT=numeric-column | range-attr-var | expression

in SAS 9.4M1 and earlier releases, specifies the column or range attribute map
variable that is used to determine the marker colors.

Note: Starting with SAS 9.4M2, the MARKERCOLORGRADIENT= option is

deprecated and replaced with the COLORRESPONSE= option. The syntax and
functionality are the same. The MARKERCOLORGRADIENT= option is still
honored, but the COLORRESPONSE= option is preferred.

Note Starting with SAS 9.4M2, if you use the MARKERCOLORGRADIENT=

option, then be aware that the TIP=, TIPFORMAT=, and TIPLABEL=
options recognize the MARKERCOLORGRADIENT role and not the
COLORRESPONSE role.

See COLORRESPONSE= on page 854

MATRIXTYPE=FULL | UPPERTRIANGLE | LOWERTRIANGLE

specifies whether to display the full matrix, or just the upper or lower triangle of
the matrix. By default, the full matrix is displayed. The cells in the grid are filled
beginning with the cell that is specified by the START= option. When you
display only the upper or lower triangle, you can use the START= option to
control the orientation of the triangle. The following figure shows the effect of
872 Chapter 6 / Plot Statements

the MATRIXTYPE= option when START=TOPLEFT and

DIAGONAL=(HISTOGRAM).

The next figure shows the effect of the START=BOTTOMLEFT option on the
previous graph.

Default FULL

Interactions This option is ignored if the ROWVARS= option is used.

The START= option specifies the corner where the matrix fill
begins.

The DIAGONAL= option specifies the content of the diagonal cells.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.
 SCATTERPLOTMATRIX Statement 873

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, MARKERCHARACTER,
COLORRESPONSE or MARKERCOLORGRADIENT, GROUP, and
INDEX.

ROWVARS=(numeric-column-list)
specifies a secondary list of numeric columns to be paired with the required
column list that is specified by the SCATTERPLOTMATRIX statement. The
labels for the columns appear vertically on the left side of the matrix.

Requirement All of the columns must be numeric.

Interaction When this option is specified, the DIAGONAL= and MATRIXTYPE=

options are ignored.

START=TOPLEFT | BOTTOMLEFT
specifies whether to start populating the matrix from the top left or bottom left
corner.

Default TOPLEFT

SUBPIXEL=AUTO | OFF
specifies whether subpixel rendering is used for image output when the scatter
plots are rendered.

Note: This feature applies to SAS 9.4M3 and to later releases.

874 Chapter 6 / Plot Statements

AUTO
sets subpixel rendering OFF, unless SUBPIXEL=ON is specified in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. In that case,
subpixel rendering is ON.

OFF
disables subpixel rendering for image output only for the scatter plots in this
scatter-plot matrix.

Default AUTO

Restriction This option applies only to the scatter plots in the matrix. It does
not affect the confidence ellipses.

Requirement Anti-aliasing must be enabled for this option to have any effect.

Notes This option affects subpixel rendering for the scatter plots only in
this scatter-plot matrix. Subpixel rendering for other plots in the
graph is not affected.

For vector-graphics output, this option is ignored, and subpixel

rendering is always enabled.

When subpixel rendering is used for the graph but is turned OFF
for this scatter-plot matrix, some elements in the scatter plots
such as the plot markers might be offset a half pixel, which can
make them appear blurry in the image output.

Tips Anti-aliasing is enabled by default. If anti-aliasing has been

disabled, use the ANTIALIAS=ON option in the ODS GRAPHICS
statement to re-enable it.

Anti-aliasing is disabled automatically for this plot when the

resources required to anti-alias it exceed a preset threshold. When
anti-aliasing is disabled for this or any other plot in the graph,
subpixel rendering is disabled for the entire graph. A note is written
to the SAS log that provides information about how to use the
ANTIALIASMAX= option in an ODS GRAPHICS statement to re-
enable anti-aliasing.

To disable subpixel rendering for the scatter plots and the ellipses,
specify SUBPIXEL=OFF in the template’s BEGINGRAPH statement
or in an ODS GRAPHICS statement.

See “Using Subpixel Rendering” in SAS Graph Template Language: User’s

Guide

“ODS GRAPHICS Statement” on page 1561 for information about

the ANTIALIAS=, ANTIALIASMAX=, and SUBPIXEL= options.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
scatter points. If this option is used, then it replaces all of the information that is
displayed by default. Roles for columns that do not contribute to the scatter
plot can be specified along with roles that do.
 SCATTERPLOTMATRIX Statement 875

(role-list)
an ordered, space-separated list of unique SCATTERPLOTMATRIX and user-
defined roles.

SCATTERPLOT roles include: X, Y, GROUP, DATALABEL,

MARKERCHARACTER, and COLORRESPONSE or
MARKERCOLORGRADIENT.

User-defined roles are defined with the ROLENAME= option.

Requirement If you use the COLORRESPONSE= option, you must specify the
COLORRESPONSE role for the color values. Likewise, if you
use the MARKERCOLORGRADIENT= option, you must specify
the MARKERCOLORGRADIENT role for the color values.
Although they are functionally the same, you cannot mix the
COLORRESPONSE= and MARKERCOLORGRADIENT= options,
and their corresponding roles.

Example The following example displays data tips for the columns
assigned to the roles TIP1, TIP2, TIP3, and TIP4:

ROLENAME=(TIP1=ID TIP2=AGE TIP3=HEIGHT TIP4=WEIGHT)

TIP=(TIP1 TIP2 TIP3 TIP4)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: current X, current Y,
MARKERCHARACTER, COLORGROUP or
MARKERCOLORGRADIENT, and GROUP.

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)
876 Chapter 6 / Plot Statements

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

WALLCOLOR=style-reference | color
specifies the fill color of the plot wall area.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style- attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphWalls:Color style reference.

Interaction This option is ignored if WALLDISPLAY=NONE or

WALLDISPLAY=(OUTLINE).

WALLDISPLAY=STANDARD | ALL | NONE | (display-options)

specifies whether the plot’s wall and wall outline are displayed.

STANDARD
displays a filled wall.

The setting of the FRAMEBORDER=ON | OFF attribute of the GraphWalls

style element determines whether the wall outline is displayed.

ALL
displays a filled, outlined wall.

NONE
displays no wall, no wall outline.
 SCATTERPLOTMATRIX Statement 877

(display-options)
a space-separated list of one or more display options enclosed in
parentheses.

The list can include one or both of the following options:

OUTLINE
displays the wall outline.

FILL
displays a filled wall area.

Default STANDARD

Tips Use the WALLCOLOR= option to control the fill color of the wall.

The appearance attributes of the wall fill and outline are set by the
GraphWalls style element.

Details
By default, the SCATTERPLOTMATRIX statement produces a symmetric scatter
plot matrix. For n columns, it produces an n columns by n rows matrix of scatter
plots. By default, the columns of the matrix are in the same left-to-right order as
the order of the numeric-column-list. The rows of the matrix are in the same
bottom-to-top order as the numeric-column-list. You can reverse the direction of
the diagonal by setting START=TOPLEFT.

To produce a rectangular matrix of scatter plots, use the ROWVARS= option.

Specifying n columns in the SCATTERPLOTMATRIX statement and m columns on
the ROWVARS= option produces an n-columns by m-rows matrix of scatter plots.
For example, the following statement specifies 2 columns on
SCATTERPLOTMATRIX and 3 columns on the ROWVARS= option to produce the 2-
columns by 3-rows matrix:

SCATTERPLOTMATRIX Height Weight

/ ROWVARS=(Age Height Weight);

The SCATTERPLOTMATRIX statement cannot appear within an overlay-type

layout. It generates its own matrix of plots and is typically placed in a LAYOUT
GRIDDED block.

If there are missing values in a column or a row, then all of the points that can be
plotted are plotted in each scatter plot.

Example: SCATTERPLOTMATRIX Statement

The following graph was generated by the “Example Program” on page 878:
 878 Chapter 6 / Plot Statements

Example Program
proc template;
define statgraph scatterplotmatrix;
begingraph;
entrytitle "Scatter Plot Matrix";
layout gridded;
scatterplotmatrix
sepallength sepalwidth petallength petalwidth /
group=species name="matrix";
discretelegend "matrix";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=scatterplotmatrix;

run;

SERIESPLOT Statement
Displays a series of line segments that connect observations of input data.

Note: The SERIESPLOT statement plots data in data order. For this reason, the input data
set should be sorted by the X variable. Otherwise, unexpected results might occur.
 SERIESPLOT Statement 879

If the data is not sorted, you can specify CONNECTORDER=XAXIS in the

SERIESPLOT statement. See CONNECTORDER= on page 887.
Tip: Starting with SAS 9.4M3, subpixel rendering is enabled by default. To disable
subpixel rendering, specify SUBPIXEL=OFF in the BEGINGRAPH statement or in an
ODS GRAPHICS statement. For information about the BEGINGRAPH statement
SUBPIXEL= option, see SUBPIXEL= on page 39. For information about the ODS
GRAPHICS statement SUBPIXEL= option, see “ODS GRAPHICS Statement” on
page 1561.

Syntax
SERIESPLOT X=column | expression
Y=column | expression </options>;

Summary of Optional Arguments

Appearance options
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.
ARROWHEADSCALE=positive-number | numeric-column | expression
specifies an arrowhead scale factor based on the thickness of the arrow
line.
ARROWHEADSHAPE= OPEN | FILLED | BARBED | column
specifies a shape for arrowheads.
BREAK=TRUE | FALSE
breaks the plot line at missing values of the Y variable.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction
of the midpoint spacing. On an interval axis, specifies the width of the
group clusters as a fraction of the minimum interval between adjacent
data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line
and marker colors.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the series lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, line label,
and data labels, when displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which series-line features to display.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
880 Chapter 6 / Plot Statements

specifies the attributes of the error bars that are associated with the data
points.
ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the error-bar caps.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and
line pattern) to one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the series line.
LINECOLORGROUP=column | discrete-attr-var | expression
specifies a column that determines the line colors for a grouped plot
independently of the GROUP= column.
LINEJOIN= MITER | ROUND | BEVEL
specifies how to shape the vertices of the series line.
LINEPATTERNGROUP=column | discrete-attr-var | expression
specifies a column that determines the line patterns for a grouped plot
independently of the GROUP= column.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line
thickness.
LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to
map a line thickness to each group value.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the data markers.
MARKERCOLORGROUP=column | discrete-attr-var | expression
specifies a column that determines the marker colors for a grouped plot
independently of the GROUP= column.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.
MARKERSYMBOLGROUP=column | discrete-attr-var | expression
specifies a column that determines the marker symbols for a grouped
plot independently of the GROUP= column.
SMOOTHCONNECT=TRUE | FALSE
 SERIESPLOT Statement 881

specifies that a smoothed line passes through all vertices.

SPLINEPOINTS=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the
INTERVAL= axis option.
SPLINETYPE=NONE | QUADRATICBEZIER
specifies the type of spline interpolation that is used to draw the series
line.
XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error
bars are drawn from the markers to the endpoints.
XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars.
YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars.
YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Connect options
CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the series line.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over
the series line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string" | column | expression
882 Chapter 6 / Plot Statements

specifies a label for the series line.

CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the series-line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the series-line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the series-line label relative to the series line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the series-line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the series-line label can be
split if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the series-line label
text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the series-line
label block.
DATALABEL=column | expression
specifies a column that supplies values for the data point labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the vertices of the
series line and the markers, when displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all series lines and markers from discrete X
values, or discrete Y values, or both.
GROUP=column | discrete-attr-var | expression
 SERIESPLOT Statement 883

creates a separate series plot for each unique group value in the specified
column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a point or a segment of the
curve is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column for the X values.

Y=column | expression
specifies the column for the Y values.

Optional Arguments
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.

Note: This feature applies to SAS 9.4M3 and to later releases.

NONE
no arrowheads are displayed.

START
an arrowhead is displayed at the starting point of each line.

END
an arrowhead is displayed at the ending point of each line.

BOTH
an arrowhead is displayed at each end of each line.

column
specifies a column that provides an arrowhead position for each group value.

Default NONE
884 Chapter 6 / Plot Statements

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead position values are assumed to be constant for each
group value. If the column has multiple values for a single group
value, then only one of the arrowhead position values is used for
that group.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

ARROWHEADSCALE=positive-number | numeric-column | expression

specifies an arrowhead scale factor based on the thickness of the arrow line.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1.0

Restriction The arrowhead scale values are assumed to be constant for each
line. If a column or expression provides multiple values for a line,
only one of the values is used.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

ARROWHEADSHAPE= OPEN | FILLED | BARBED | column

specifies a shape for arrowheads.

Note: This feature applies to SAS 9.4M3 and to later releases.

The following figure shows each of the arrowhead shapes.

Default OPEN

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead shape values are assumed to be constant for each group
value. If the column has multiple values for a single group value,
only one of the arrowhead shape values is used for that group.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.
 SERIESPLOT Statement 885

BREAK=TRUE | FALSE
breaks the plot line at missing values of the Y variable.

Default FALSE

Interaction This option is ignored when the Y axis is discrete and option
INCLUDEMISSINGDISCRETE=TRUE is specified in the
BEGINGRAPH statement. To break the line in that case, remove
INCLUDEMISSINGDISCRETE=TRUE from the BEGINGRAPH
statement.

Note When this option is set to FALSE, missing values are skipped and a
continuous line is drawn.

See “boolean ” on page xii for other Boolean values that you can use.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO
uses the discrete axis for clustering groups when only one axis is discrete.
Uses the X axis for clustering if both axes are discrete or interval.

X
Y
uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters
as a fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in effect,
the size of the markers in each cluster might be reduced to no less
than 5 pixels in order to display the cluster within the smallest
effective midpoint space. If you need larger markers in that case,
use the MARKERATTRS= option to specify a larger marker size.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
886 Chapter 6 / Plot Statements

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line and
marker colors.

Note: This feature applies to SAS 9.4M3 and to later releases.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
 SERIESPLOT Statement 887

Requirement The COLORRESPONSE values should remain constant for each

group value in a grouped plot and for the entire plot in an
ungrouped plot. If the COLORRESPONSE column has multiple
values for a single GROUP value or for a non-grouped plot,
unexpected results might occur.

Interactions When this option is specified without the GROUP= option, only a
single line is generated for the plot, and the line color is derived
from the COLORRESPONSE= value.

When the GROUP= option, LINECOLORGROUP= option, and

MARKERCOLORGROUP= options are used with the
COLORRESPONSE= option, the COLORRESPONSE= option
controls the color attributes.

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the LINEATTRS=

option and varies the line color according to the color gradient or
the attribute map.

Suboption COLOR= in the MARKERATTRS= option or in the

FILLEDOUTLINEDMARKERS= option overrides this option for the
marker colors.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorAltRamp style

element defines the line color gradient.

CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the series line.

XVALUES
connects data points in the order read from the X column.

Note In SAS 9.4M6 and earlier releases, unexpected results might occur
when XVALUES is specified and data stored in a CAS in-memory table
is used. In that case, use XAXIS instead. Starting with SAS Viya 3.5
and SAS Studio 5.2, XVALUES is ignored when data stored in a CAS in-
memory table is used.

XAXIS
connects data points as they occur min-to-max along the X axis.

Defaults XVALUES if the data is a SAS data set

XAXIS if the data is a CAS in-memory table

Tip For certain types of series lines (for example, time series) when the
input data might not be sorted by the X column, set this option to
XAXIS to assure the expected connect order.
888 Chapter 6 / Plot Statements

CURVELABEL="string" | column | expression

specifies a label for the series line.

Default No series-line label is displayed

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 900

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the series-line labels. When the
GROUP= option is used, each distinct group value might be represented by a
different color. The series-line label that is associated with the group is assigned
the group color. This option can be used to specify a single color for all series-
line labels in a plot, without affecting items that have the group color, such as
lines and marker symbols.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the series-line label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE
 SERIESPLOT Statement 889

Restriction OUTSIDE cannot be used when the SERIESPLOT is used in

multicell layouts such as LATTICE, DATAPANEL, or DATALATTICE,
where axes might be external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the series-
line labels appear. For more information, see “Location and Position
of Curve Labels” on page 213.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the series-line label relative to the series line.

AUTO
automatically locates the series-line label near the end series line along
unused axes whenever possible (typically Y2 or X2) in order to avoid
collision with tick values.

Restriction This option is used only when CURVELABELLOCATION=

OUTSIDE.

MAX
forces the series-line label to appear near maximum series values (typically,
to the right)

MIN
forces the series-line label to appear near minimum series values (typically,
to the left)

START
forces the series-line label to appear near the beginning of the series line.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the series line has a spiral
shape.

END
forces the series-line label to appear near the end of the series line.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the series line has a spiral
shape.

Defaults AUTO when CURVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION=INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.
890 Chapter 6 / Plot Statements

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the series-
line label appears. For more information, see “Location and Position
of Curve Labels” on page 213.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the series-line label might fall outside of the graph area.
In that case, the series-line label might not be displayed or might be
positioned incorrectly.

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the series-line label at the specified split characters.
When a series-line label is split, the label is split on each occurrence of the
specified split characters.

Default FALSE. The series-line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the series-line label can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the series-line label. In that case, all of the specified split
characters together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the series-

line label is split unconditionally at each occurrence of any of the specified split
characters. If the series-line label does not contain any of the specified
characters, then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

 SERIESPLOT Statement 891

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the series-line label or are
dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the series-line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the series-line label text.

TRUE
drops the split characters from the series-line label text.

FALSE
includes the split characters in the series-line label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a series-line label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the series line for reference.
892 Chapter 6 / Plot Statements

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the series-line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the series-line label
block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the series line for reference.
 SERIESPLOT Statement 893

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATALABEL=column | expression
specifies a column that supplies values for the data point labels.

Default No data labels are displayed

Note The label positions are adjusted to prevent the labels from overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also
be specified.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

Note When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to
avoid overlapping labels and markers.

Tip When the GROUP= option is used, each distinct group value might
be represented by a different color. The data label that is
associated with the group is assigned the group color. This option
can be used to specify a single color for all data labels in a plot,
without affecting items that have the group color, such as error bars
and marker symbols.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the vertices of the series line
and the markers, when displayed.

AUTO
uses a data-label collision avoidance algorithm to position the labels.
894 Chapter 6 / Plot Statements

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
 SERIESPLOT Statement 895

datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.
896 Chapter 6 / Plot Statements

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

 SERIESPLOT Statement 897

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the series lines. The following figure shows a
grouped series plot with each of the skins applied.

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data
skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The data skin appearance is based on the LINEATTRS= color.

DATATRANSPARENCY=number
specifies the degree of the transparency of the line, markers, line label, and data
labels, when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

DISCRETEOFFSET=number
specifies an amount to offset all series lines and markers from discrete X values,
or discrete Y values, or both. This option is useful when graphing multiple
response variables side by side on a common axis

Default 0 (no offset, all series lines and markers are centered on the discrete
X values, or discrete Y values, or both)
898 Chapter 6 / Plot Statements

Range -0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right on discrete X values
and up on discrete Y values. If the layout’s axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=(display-options) | STANDARD | ALL

specifies which series-line features to display.

(display-options)
a space-separated list of one or more options enclosed in parentheses.

Currently, only the MARKERS option is supported, which displays a series

line with markers.

STANDARD
displays a series line without markers.

ALL
displays a series line with markers.

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the error-bar caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1

Interaction This option is ignored when ERRORBARCAPSHAPE=NONE is in

effect.

Tips Use the ERRORBARCAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the ERRORBARATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the error-bar caps:

errorbarcapscale=2
 SERIESPLOT Statement 899

Reduce the size of the error-bar caps by 50%:

errorbarcapscale=0.5

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data
points.

Note: This feature applies to SAS 9.4M5 and to later releases.

Defaults For ungrouped data, the GraphError style element ContrastColor,

LineStyle, and LineThickness attributes.

For grouped data, a GraphData1–GraphDataN style element

ContrastColor attribute, and the GraphDataDefault style elemement
LineStyle and LineThickness attributes.

Interaction For this option to take effect, error bars must be displayed by the
XERRORLOWER= , XERRORUPPER= , YERRORLOWER= , or
YERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default The GraphError:CapStyle style reference. If attribute CapStyle is not

defined in the active style, then SERIF is the default value.

Interaction Starting with SAS 9.4M5, when ERRORBARCAPSCALE= is in effect,

the display width for SERIF caps is the cap's default width multiplied
by the value specified in ERRORBARCAPSCALE=.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.

TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.
900 Chapter 6 / Plot Statements

When this option is TRUE, the fill color and outline color for filled markers
are determined in the following ways:
n If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and
the marker outlined color is derived from the GraphData1–GraphDataN
style elements ContrastColor attribute.
n If the GROUP= option is not specified, then the marker fill is drawn by
using the MARKERFILLATTRS= specification, and the outline is drawn by
using the MARKEROUTLINEATTRS= specification.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set
this option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 900

MARKERFILLATTRS= on page 911

MARKEROUTLINEATTRS= on page 911

“boolean ” on page xii for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a separate series plot for each unique group value in the specified
column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol.
Lines and markers vary according to the ContrastColor, LineStyle,
and MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements. Line thickness (for grouped and
ungrouped data) is controlled by the LINEATTRS= option.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned
line cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000.
If the limit is exceeded, the plot is not drawn and a note is written
to the SAS log stating that the limit has been exceeded. In that
 SERIESPLOT Statement 901

case, increase LINEPATTERNOBSMAX= to the value suggested in

the note or change the line pattern to SOLID.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and the COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden. For example, each distinct group value is often
represented by a different line pattern, but you can use the
LINEATTRS=(PATTERN=pattern) option to assign the same line
pattern to all of the plot’s line patterns, letting line color indicate
group values. Likewise, you could use LINEATTRS=(COLOR= color)
to assign the same color to all lines, letting line pattern indicate
group values.

Starting with SAS 9.4M2, you can use the LINECOLORGROUP=,

LINEPATTERNGROUP=, MARKERCOLORGROUP=, and
MARKERSYMBOL= options to assign line colors, line patterns,
marker colors, and marker symbols based on a different group
column.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.

OVERLAY
draws markers for a given group value at the exact coordinate.

Depending on the data, markers at a given coordinate might overlap.

CLUSTER
draws markers for a given group value adjacent to each other.
902 Chapter 6 / Plot Statements

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

Default OVERLAY

Interaction For this option to take effect, the GROUP= option must also be
specified.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
 SERIESPLOT Statement 903

formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

Tips Use the CLUSTERWIDTH= option to control the distance between

the group markers in a cluster.

Use the INDEX= option to alter the default sequence of visual

attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
904 Chapter 6 / Plot Statements

then a modulo operation remaps that index value to a number

less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the series line.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphDataDefault:LineThickness style reference.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned
line cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000.
If the limit is exceeded, the plot is not drawn and a note is written
to the SAS log stating that the limit has been exceeded. In that
case, increase LINEPATTERNOBSMAX= to the value suggested in
the note or change the line pattern to SOLID.

Interactions This option’s COLOR= suboption overrides the default behavior for
grouped data when the LINECOLORGROUP= option and the
COLORRESPONSE= option are not specified. When the COLOR=
suboption is specified without the LINECOLORGROUP= and
COLORRESPONSE= options, all lines have the same color.

The LINECOLORGROUP= option and the COLORRESPONSE=

option override this option’s COLOR= suboption.

This option’s PATTERN= suboption overrides the default behavior

for grouped data when the LINEPATTERNGROUP= option is not
specified. When the PATTERN= suboption is specified without the
LINEPATTERNGROUP= option, all lines have the same pattern.

The LINEPATTERNGROUP= option overrides this option’s

PATTERN= suboption.
 SERIESPLOT Statement 905

Tip If you specify a thickness of 2 pixels or greater for the series line,
use the LINEJOIN= option to shape the line vertices so that the line
does not extend beyond the marker symbols.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Line Options” on page 1602 for available line-options.

LINECOLORGROUP=column | discrete-attr-var | expression

specifies a column that determines the line colors for a grouped plot
independently of the GROUP= column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with SAS 9.4M3.

When this option is specified with the GROUP= option, the series line colors are
selected from the GraphData1–GraphDataN style elements based on the
LINECOLORGROUP= column instead of on the GROUP= column.

Default The line colors are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that
define a series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the COLOR= suboption of the LINEATTRS=

option.

The COLORRESPONSE= option overrides this option.

Note If you specify a column in a SAS data set, colors are mapped to
group values in data order. If you specify a column in a CAS in-
memory table, colors are mapped to group values in ascending
order of the group column character values or unformatted
numeric values.

Tip Use the LINEATTRS= option to set the line thickness.

See GROUP= on page 900

906 Chapter 6 / Plot Statements

LINEATTRS=

LINEJOIN= MITER | ROUND | BEVEL

specifies how to shape the vertices of the series line.

Note: This feature applies to SAS 9.4M4 and to later releases.

MITER
joins the line vertices without adjustment.

ROUND
rounds the apex of each line vertex, effectively smoothing the apex edges.

BEVEL
bevels the apex of each line vertex, effectively trimming the apex to remove
any part of the line that extends beyond the apex point.

The following figure shows the effect each value of LINEJOIN= on the vertex of
two segments of a 10-point line. The red dot depicts the data point at the
vertex.

Default MITER

Tip Use this option when the series line thickness is 2 pixels or greater to
prevent the line from extending beyond the marker symbols at points
where large slope changes occur.

LINEPATTERNGROUP=column | discrete-attr-var | expression

specifies a column that determines the line patterns for a grouped plot
independently of the GROUP= column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with SAS 9.4M3.
 SERIESPLOT Statement 907

When this option is specified with the GROUP= option, the series line patterns
are selected from the GraphData1–GraphDataN style elements based on the
LINECOLORGROUP= column instead of on the GROUP= column.

Default The line patterns are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that
define a series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the PATTERN= suboption of the

LINEATTRS= option.

Note If you specify a column in a SAS data set, the group-value line
patterns are assigned in data order. If you specify a column in a
CAS in-memory table, group-value line patterns are assigned in
ascending order of the group column character values or
unformatted numeric values.

Tip Use the LINEATTRS= option to set the line thickness.

See GROUP= on page 900

LINEATTRS=

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness. By default, this option determines the thickness of
the line that represents the maximum response column value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default Ten times the thickness that is specified by the GraphDataDefault

style element LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached.
The line thickness for response values that exceed the
LINETHICKNESSMAXRESPONSE= value are set to the value that
is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.
908 Chapter 6 / Plot Statements

See “dimension” on page xiii

“Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this
option.

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value
is set to the value specified in the LINETHICKNESSMAX= option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum line

thickness.

See “dimension” on page xiii

“Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this
option.
 SERIESPLOT Statement 909

LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing,
this option is ignored. In these cases, the default line thickness is
used for all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

Notes The line thicknesses are not reflected in a discrete legend.

The LINETHICKNESSRESPONSE= values are assumed to be

constant for each group value in a grouped plot and for the entire
plot in an ungrouped plot. If the LINETHICKNESSRESPONSE
column has multiple values for a single GROUP value or ungrouped
plot, unpredictable results might occur.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this
option.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s COLOR=

suboption is ignored. In that case, to specify the marker fill color,
use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data when the MARKERCOLORGROUP= option is not
specified. When the COLOR= suboption is specified without the
MARKERCOLORGROUP option, all markers have the same color,
and the marker symbol alone distinguishes the markers.

The MARKERCOLORGROUP= option and theCOLORRESPONSE=

option override this option’s COLOR= suboption.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data when the MARKERSYMBOLGROUP= option is not
specified. When the SYMBOL= suboption is specified without the
910 Chapter 6 / Plot Statements

MARKERSYMBOLGROUP= option, all markers have the same

symbol, and the symbol color alone distinguishes the markers.

The MARKERSYMBOLGROUP= option overrides this option’s

SYMBOL= suboption.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display

of the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current
ODS style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MARKERCOLORGROUP=column | discrete-attr-var | expression

specifies a column that determines the marker colors for a grouped plot
independently of the GROUP= column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with SAS 9.4M3.

When this option is specified with the GROUP= option, the series line markers
are selected from the GraphData1–GraphDataN style elements based on the
MARKERCOLORGROUP= column instead of on the GROUP= column.

Default The line marker colors are selected based on the GROUP= column.

Requirement The column value must be the same for all of the points that
define a series line.
 SERIESPLOT Statement 911

Interactions The DISPLAY= option must enable the display of the series line
markers for this option to have any effect.

The GROUP= option must be specified for this option to have any
effect.

This option overrides the COLOR= suboption of the

MARKERATTRS= option.

The COLORRESPONSE= option overrides this option.

Note If you specify a column in a SAS data set, the group-value marker
colors are assigned in data order. If you specify a column in a CAS
in-memory table, group-value marker colors are assigned in
ascending order of the group column character values or
unformatted numeric values.

Tip Use the MARKERATTRS= option to set the marker symbol size.

See GROUP= on page 900

MARKERATTRS= on page 909

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element

For grouped data, the Color attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

Note When style-element is specified, only the style element’s Color

attribute is used.

See “General Syntax for Attribute Options” on page 1599

“Fill Color Options” on page 1600

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.
912 Chapter 6 / Plot Statements

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current
style or by the DATASKIN= option. In the latter case, the outline is
set by the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1599

“Line Options” on page 1602

MARKERSYMBOLGROUP=column | discrete-attr-var | expression

specifies a column that determines the marker symbols for a grouped plot
independently of the GROUP= column.

Note: This feature applies to SAS 9.4M2 and to later releases.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Note A discrete attribute map variable is valid for this option starting
with SAS 9.4M3.

When this option is specified with the GROUP= option, the series line marker
symbols are selected from the GraphData1–GraphDataN style elements based
on the MARKERSYMBOLGROUP= column instead of on the GROUP= column.

Default The line marker symbols are selected based on the GROUP=
column.

Requirement The column value must be the same for all of the points that
define a series line.

Interactions The GROUP= option must be specified for this option to have any
effect.

This option overrides the SYMBOL= suboption of the

MARKERATTRS= option.

The DISPLAY= option must enable the display of the series line
markers for this option to have any effect.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a
column in a CAS in-memory table, the visual attributes for each
 SERIESPLOT Statement 913

group value are assigned in ascending order of the group column

character values or unformatted numeric values.

Tip Use the MARKERATTRS= option to set the marker symbol size.

See GROUP= on page 900

MARKERATTRS= on page 909

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)
914 Chapter 6 / Plot Statements

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, CURVELABEL, DATALABEL, GROUP,
and INDEX.

SMOOTHCONNECT=TRUE | FALSE
specifies that a smoothed line passes through all vertices.

Default FALSE

Interaction Starting with SAS 9.4M3, this option is ignored when

SPLINETYPE=QUADRATICBEZIER is in effect.

See “boolean ” on page xii for other Boolean values that you can use.

SPLINEPOINTS=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the
INTERVAL= axis option.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 20

Interaction When this option is set to a non-default value, markers and data
labels, when displayed, are positioned at their original data points.

SPLINETYPE=NONE | QUADRATICBEZIER
specifies the type of spline interpolation that is used to draw the series line.

Note: This feature applies to SAS 9.4M3 and to later releases.

The following figure shows an example of SPLINETYPE= values

QUADRATICBEZIER and NONE.

Figure 6.2 Spline Types QUADRATICBEZIER and NONE

Default NONE

Interaction The SMOOTHCONNECT= option is ignored when this option is set

to a value other than NONE.
 SERIESPLOT Statement 915

Note Markers and data labels, when displayed, are positioned at their
original data points as shown in Figure 6.5 on page 914.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the
series line. If this option is used, then the information specified replaces all the
information that is displayed by default. You can specify roles for columns that
do not contribute to the series plot along with roles that do.

(role-list)
an ordered, space-separated list of unique SERIESPLOT roles and user-
defined roles.

SERIESPLOT roles include X, Y, CURVELABEL, DATALABEL, GROUP, and

COLORRESPONSE.

Define user-defined roles with the ROLENAME= option.

Notes CURVELABEL is considered a role only when it is assigned a

column of values. It is not considered a role and does not display
data tips when it is assigned a string.

Starting with SAS 9.4M3, the COLORRESPONSE role is valid.

Example The following example displays data tips for the columns assigned
to the roles X and Y as well as the column Obs, which is not
assigned to any pre-defined SERIESPLOT role. The Obs column
must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X, Y, DATALABEL, COLORRESPONSE,
and GROUP.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.
916 Chapter 6 / Plot Statements

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a point or a segment of the curve is
selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each series line segment that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable points or segments of the curve, you must

include an ODS GRAPHICS ON statement that specifies the
IMAGEMAP option, and you must write the output to the ODS
HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.
 SERIESPLOT Statement 917

Note When you select a portion of a segment that is not an endpoint,

the nearest segment endpoint is used.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding segment is selected.

The URL value can be the same for any X and Y pair, meaning that
the same action is taken when the segment for those X and Y pairs
is selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

XERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the X error bars. The error bars are
drawn from the markers to the endpoints.

Note: This option is valid in SAS 9.4M5 and later releases.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

XERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the X error bars. The error bars are
drawn from the markers to the endpoints.

Note: This option is valid in SAS 9.4M5 and later releases.

918 Chapter 6 / Plot Statements

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YERRORLOWER=numeric-column | expression
specifies values for the lower endpoints on the Y error bars. The error bars are
drawn from the markers to the endpoints.

Note: This option is valid in SAS 9.4M5 and later releases.

Default The lower segment of the error bars is not drawn.

Note The values are actual values, not relative values.

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

YERRORUPPER=numeric-column | expression
specifies values for the upper endpoints on the Y error bars. The error bars are
drawn from the markers to the endpoints.

Note: This option is valid in SAS 9.4M5 and later releases.

Default The upper segment of the error bars is not drawn.

Note The values are actual values, not relative values.

 SERIESPLOT Statement 919

Tips The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

If markers are displayed in the plot, then the markers overlay the error
bars. Large filled markers can obscure short error bars. To enable the
error bars to show through the markers in that case, you can use the
MARKERATTRS= option to specify a degree of transparency for the
filled markers.

Details
The SERIESPLOT statement is typically used to show time-dependent data.

Examples

Example 1: Overlay Series Plot

The following graph was generated by the “Example Program” on page 919:

Example Program
/* Mean temperature data for August in 2014. */
data augusttemps;
input day temp dewpoint heat_index;
label day="Day of Month" Temp="Temperature"
Dewpoint="Dew Point" Heat_Index="Heat Index";
920 Chapter 6 / Plot Statements

datalines;
1 68.74 67.96 71.14
2 66.49 66.49 68.23
3 70.05 69.41 72.70
4 73.70 70.19 76.82
5 76.09 70.89 79.89
6 78.57 70.16 81.93
7 77.10 71.02 80.98
8 74.14 70.33 77.20
9 67.16 66.66 69.01
10 70.39 68.82 72.75
11 73.75 71.53 77.40
12 74.11 72.81 78.89
13 73.36 67.32 75.66
14 70.79 61.63 71.61
15 73.40 66.76 75.61
16 76.79 70.37 80.37
17 77.64 72.23 82.17
18 76.73 74.12 81.65
19 75.44 72.92 79.85
20 78.03 73.85 83.52
21 79.32 74.57 85.02
22 78.58 75.37 84.45
23 76.80 75.15 81.79
24 73.46 67.93 75.74
25 69.71 64.37 71.18
26 70.63 65.01 72.13
27 71.98 64.41 73.45
28 76.43 70.46 81.25
29 77.17 74.14 81.68
30 78.12 73.98 83.58
31 79.77 74.46 85.64
;
run;

proc template;
define statgraph seriesoverlay;
begingraph;
entrytitle "Mean Daily Outdoor Temperatures in August";
layout overlay / cycleattrs=true
xaxisopts=(type=linear
griddisplay=on gridattrs=(color=lightgray pattern=solid)
linearopts=(minorticks=true minortickcount=9
minorgrid=true minorgridattrs=(color=lightgray
pattern=dot)))
yaxisopts=(type=linear
label="Temperature ((*ESC*){unicode '00B0'x}F)"
griddisplay=on gridattrs=(color=lightgray pattern=solid)
linearopts=(minorticks=true minortickcount=4
minorgrid=true minorgridattrs=(color=lightgray
pattern=dot)));
seriesplot x = day y = temp / name="temp"
smoothconnect=true lineattrs=(thickness=3px);
seriesplot x = day y = heat_index / name = "heat_index"
smoothconnect=true
lineattrs=(thickness=2px pattern=shortdash);
 SERIESPLOT Statement 921

seriesplot x = day y = dewpoint / name = "dewpoint"

smoothconnect=true
lineattrs=(thickness=2px pattern=dot);
discretelegend "temp" "dewpoint" "heat_index";
endlayout;
endgraph;
end;
run;

proc sgrender data=augusttemps template=seriesoverlay;

run;

Example 2: Grouped Series Plot

The following graph was generated by the “Example Program” on page 921:

Example Program
proc template;
define statgraph seriesplot;
begingraph;
entrytitle "Tech Stock Trends";
layout overlay /
xaxisopts=(label="Date (mm/yy)"
griddisplay=on gridattrs=(color=lightgray pattern=dot)
timeopts=(tickvalueformat=mmyys5. minorticks=true
minortickinterval=month))
yaxisopts=(griddisplay=on gridattrs=(color=lightgray
pattern=dot)
 922 Chapter 6 / Plot Statements

linearopts=(minorticks=true minortickcount=5));
seriesplot x=date y=close / group=stock name="stocks";
discretelegend "stocks";
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.stocks template=seriesplot;

where date > "31dec1999"d;
run;

Example 3: Series Plot with Line-Thickness Response and

Arrowheads
Starting with SAS 9.4M3, you can specify a response variable to control the
thickness of the lines in your grouped series plot. You can also add arrowheads to
one or both ends of your grouped or ungrouped series lines. This example creates a
grouped series plot that shows the monthly closing price for IBM, Intel, and
Microsoft stock in 2001. The mean volume is computed for each stock and is used
to control the thickness of the series line for each stock. The minimum line
thickness is set to 2px, and the maximum line width is set to 7px. A barbed
arrowhead, scaled to 1.25, is positioned at the end of each series line. In order to
position the arrowheads properly, the data must be sorted in ascending order by
date. The following figure shows the output.

Example Program
/* Extract the 2001 data from Sashelp.Stocks
and convert Volume to millions. */
data stocks;
 SERIESPLOT Statement 923

set sashelp.stocks(where=
(date between "1jan02"d and "31dec02"d));
volume = volume / 1000000;
format date MONNAME3. volume 6.2;
;

/* Compute the average volume for each stock */

proc means data=stocks noprint;
by stock notsorted;
var volume;
output out=meanvolume(keep=stock meanvolume) mean=meanvolume;
run;

/* Merge the average volume data with the stock data */

data stocks;
merge stocks meanvolume;
by stock;
run;

/* Sort the data by Date */

proc sort data=stocks;
by date;
run;

/* Define the graph template */

proc template;
define statgraph seriesplot;
begingraph / drawspace=wallpercent;
entrytitle "Stock Trends in 2001";
entryfootnote "Line thickness reflects the average volume.";
layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot)
type=discrete label="Month")
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
seriesplot x=date y=close / group=stock name="stocks"
arrowheadposition=end arrowheadshape=barbed
arrowheadscale=1.25
linethicknessresponse=meanvolume
linethicknessmin=2px linethicknessmax=7px
curvelabel=meanvolume;
drawtext "Average Volume (Millions)" / x=94 y=90
width=80 widthunit=pixel;
discretelegend "stocks";
endlayout;
endgraph;
end;

/* Render the graph */

proc sgrender data=stocks template=seriesplot;
run;
 924 Chapter 6 / Plot Statements

Example 4: Series Plot with Quadratic Bézier Spline Curves

This example shows you how to create a series plot with quadratic Bézier spline
curves. The SERIESPLOT statement is used to generate a grouped plot of stock
closing prices in the year 2002. The SPLINETYPE=QUADRATICBEZIER option is
used in the SERIESPLOT statement to specify quadratic Bézier spline curves. Here
is the output for this example.

Example Program
proc template;
define statgraph seriesplot;
begingraph;
entrytitle "Stock Performance in 2002";
layout overlay /
xaxisopts=(type=discrete label="Month"
discreteopts=(tickvaluefitpolicy=thin));
seriesplot x=date y=close / name="series" group=stock
splinetype=quadraticbezier;
discretelegend "series" / across=3 displayclipped=true;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=seriesplot;

where date between "1jan02"d and "31dec02"d;
format date MONNAME3.;
run;
 STEPPLOT Statement 925

STEPPLOT Statement
Displays a series of horizontal and vertical line segments that connect observations of input data.

Note: The STEPPLOT statement plots data in data order. For this reason, the input data
set should be sorted by the X variable. Otherwise, unexpected results might occur.
If the data is not sorted, you can specify CONNECTORDER=XAXIS in the
STEPPLOT statement. See CONNECTORDER= on page 931.

Syntax
STEPPLOT X=column | expression
Y=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.
ARROWHEADSCALE=positive-number | numeric-column | expression
specifies an arrowhead scale factor based on the thickness of the arrow
line.
ARROWHEADSHAPE= OPEN | FILLED | BARBED | column
specifies a shape for arrowheads.
BREAK=TRUE | FALSE
determines whether the plot line should show breaks at occurrences of
missing values of the Y column.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction
of the midpoint spacing. On an interval axis, specifies the width of the
group clusters as a fraction of the minimum interval between adjacent
data values.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the step plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the step lines, markers, error
bars, step-line labels, and data labels, when displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which step-line features to display.
ERRORBARATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the error bars that are associated with the data
points.
926 Chapter 6 / Plot Statements

ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the
default width of the error-bar caps.
ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.
ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and
line pattern) to one of the GraphData1–GraphDataN style elements.
JUSTIFY=(LEFT | CENTER | RIGHT)
specifies the location of the data point relative to the step.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the step line connecting the data points.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line
thickness.
LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to
map a line thickness to each group value.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the attributes of the data markers.
MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.
MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
 STEPPLOT Statement 927

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Connect options
CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the step line.
JOIN=TRUE | FALSE
specifies whether the steps are connected.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over
the step line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
CURVELABEL="string" | column | expression
specifies a label for the step line.
CURVELABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the step-line labels.
CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the step-line label relative to the plot area.
CURVELABELPOSITION=AUTO | MAX | MIN | START | END
specifies the position of the step-line labels relative to the step line.
CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the step-line label at the specified split
characters.
CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the step-line label can be split
if needed.
CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the step-line label
text.
CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the step-line label
block.
DATALABEL=column | expression
specifies a column that supplies values for the data point labels.
DATALABELATTRS=style-element | style-element (text-options) | (text-
options)
specifies the color and font attributes of the data labels.
928 Chapter 6 / Plot Statements

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the data points and
markers, when displayed.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies an amount to offset all step lines and markers from discrete X
values when graphing multiple response variables side by side on a
common axis.
GROUP=column | discrete-attr-var | expression
creates a distinct set of lines, markers, and data labels for each unique
group value of the specified column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped step lines and markers are overlaid or
clustered around the category midpoints.
GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING
specifies the ordering of the groups within a category.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a step line segment is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=column | expression
specifies the column of the X values.

Y=numeric-column | expression
specifies the numeric column of the Y values.
 STEPPLOT Statement 929

Optional Arguments
ARROWHEADPOSITION= NONE | START | END | BOTH | column
specifies a position for arrowheads.

Note: This feature applies to SAS 9.4M3 and to later releases.

NONE
no arrowheads are displayed.

START
an arrowhead is displayed at the starting point of each line.

END
an arrowhead is displayed at the ending point of each line.

BOTH
an arrowhead is displayed at each end of each line.

column
specifies a column that provides an arrowhead position for each group value.

Default NONE

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead position values are assumed to be constant for each
group value. If the column has multiple values for a single group
value, then only one of the arrowhead position values is used for
that group.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

ARROWHEADSCALE=positive-number | numeric-column | expression

specifies an arrowhead scale factor based on the thickness of the arrow line.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1.0

Restriction The arrowhead scale values are assumed to be constant for each
line. If a column or expression provides multiple values for a line,
only one of the values is used.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

ARROWHEADSHAPE= OPEN | FILLED | BARBED | column

specifies a shape for arrowheads.
930 Chapter 6 / Plot Statements

Note: This feature applies to SAS 9.4M3 and to later releases.

The following figure shows each of the arrowhead shapes.

Default OPEN

Restriction When you specify a column and the GROUP= option is in effect, the
arrowhead shape values are assumed to be constant for each group
value. If the column has multiple values for a single group value,
only one of the arrowhead shape values is used for that group.

Interaction This option is ignored when ARROWHEADPOSITION=NONE is in

effect.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

BREAK=TRUE | FALSE
determines whether the plot line should show breaks at occurrences of missing
values of the Y column.

Default FALSE

Note When this option is set to FALSE, missing values are skipped and the
line continues through the missing value and to the next point.

See “boolean ” on page xii for other Boolean values that you can use.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO
uses the discrete axis for clustering groups when only one axis is discrete.
Uses the X axis for clustering if both axes are discrete or interval.

X
Y
uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters
as a fraction of the minimum interval between adjacent data values.
 STEPPLOT Statement 931

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interactions For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

When markers are displayed for interval data and

GROUPDISPLAY=CLUSTER and CLUSTERWIDTH= are in effect,
the size of the markers in each cluster might be reduced to no less
than 5 pixels in order to display the cluster within the smallest
effective midpoint space. If you need larger markers in that case,
use the MARKERATTRS= option to specify a larger marker size.

CONNECTORDER=XVALUES | XAXIS
specifies how to connect the data points to form the step line.

XVALUES
connects data points in the data order of the X column.

Note In SAS 9.4M6 and earlier releases, unexpected results might occur
when XVALUES is specified and data stored in a CAS in-memory table
is used. In that case, use XAXIS instead. Starting with SAS Viya 3.5
and SAS Studio 5.2, XVALUES is ignored when data stored in a CAS in-
memory table is used.

XAXIS
connects data points sorted by their X values.

Tip When the input data for the step lines is not sorted by the X column,
you can use XAXIS to assure the expected connect order.

Defaults XVALUES if the data is a SAS data set

XAXIS if the data is a CAS in-memory table

CURVELABEL="string" | column | expression

specifies a label for the step line.

Restrictions When the GROUP= option is specified, "string" and expression are
not valid. Use column in that case.

When the GROUP= option is not specified, column is not valid. Use
"string" or expression in that case.

The line label for missing values is ignored.

Tip The font and color attributes for the label are specified by the
CURVELABELATTRS= option.

See GROUP= on page 944

932 Chapter 6 / Plot Statements

CURVELABELATTRS=style-element | style-element (text-options) | (text-

options)
specifies the color and font attributes of the step-line labels. When the
GROUP= option is used, each distinct group value might be represented by a
different color. The series label that is associated with the group is assigned the
group color. This option can be used to specify a single color for all series labels
in a plot, without affecting items that have the group color, such as lines and
marker symbols.

Defaults For non-grouped data, the GraphValueText style element.

For grouped data, text color is derived from the

GraphData1:ContrastColor–GraphDataN:ContrastColor style
references. The font is derived from the GraphValueText style
element.

Interactions For this option to take effect, the CURVELABEL= option must also
be used.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

CURVELABELLOCATION=INSIDE | OUTSIDE
specifies the location of the step-line label relative to the plot area.

INSIDE
locates the labels inside the plot area

OUTSIDE
locates the labels outside the plot area

Default INSIDE

Restriction OUTSIDE cannot be used when the STEPPLOT is used in multicell

layouts such as LATTICE, DATAPANEL, or DATALATTICE, where
axes are external to the grid.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELPOSITION= option to determine where the step-line
labels appear. For more information, see “Location and Position of
Curve Labels” on page 213.

CURVELABELPOSITION=AUTO | MAX | MIN | START | END

specifies the position of the step-line labels relative to the step line.
 STEPPLOT Statement 933

AUTO
automatically positions the step-line label near the step boundary along
unused axes whenever possible (typically Y2 and X2) in order to avoid
collision with tick values.

Restriction This option is used only when

CURVELABELPOSITION=OUTSIDE.

MAX
forces the step-line label to appear near maximum step values (typically,
upper right)

MIN
forces the step-line label to appear near minimum step values (typically,
lower left)

START
forces the step-line label to appear near the beginning of the steps.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the step line spirals
around.

END
forces the step-line label to appear near the end of the steps.

Restriction This option is used only when CURVELABELLOCATION=INSIDE.

Tip This option is particularly useful when the step line spirals
around.

Defaults AUTO when CUVELABELLOCATION=OUTSIDE.

END when CURVELABELLOCATION=INSIDE.

Restriction The AUTO setting is ignored if CURVELABELLOCATION= INSIDE is

specified. The START and END settings are ignored if
CURVELABELLOCATION=OUTSIDE is specified.

Interactions For this option to take effect, the CURVELABEL= option must also
be specified.

This option is used in conjunction with the

CURVELABELLOCATION= option to determine where the step-line
label appears. For more information, see “Location and Position of
Curve Labels” on page 213.

Note When you specify TICKVALUELIST=, VIEWMAX=, or VIEWMIN= in

an axis statement, the data points that are used to determine the
position of the step-line label might fall outside of the graph area.
In that case, the step-line label might not be displayed or might be
positioned incorrectly.
934 Chapter 6 / Plot Statements

Tip When END is in effect, no attempt is made to avoid curve label

collisions. If your curve labels collide, try using MAX instead.

CURVELABELSPLIT=TRUE | FALSE
specifies whether to split the step-line label at the specified split characters.
When a step-line label is split, the label is split on each occurrence of the
specified split characters.

Default FALSE. The step-line label is not split.

Requirement The CURVELABEL= option must also be specified.

Interactions The CURVELABELSPLITCHAR= option specifies one or more

characters on which the splits occur.

This option has no effect when CURVELABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITCHAR="character-list"
specifies one or more characters on which the step-line label can be split if
needed. When multiple split characters are specified, each character in the list
is treated as a separate split character unless the specified characters appear
consecutively in the step-line label. In that case, all of the specified split
characters together are treated as a single split character.

When CURVELABEL= is specified and CURVELABELSPLIT=TRUE, the step-line

label is split unconditionally at each occurrence of any of the specified split
characters. If the step-line label does not contain any of the specified
characters, then the label is not split.

"character-list"
one or more characters with no delimiter between each character and
enclosed in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the split characters a, b, and c, use the
following option:
curvelabelsplitchar="abc"

The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if CURVELABELPOSITION=AUTO.

The CURVELABELSPLITCHARDROP= option specifies whether

the split characters are included in the step-line label or are
dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.
 STEPPLOT Statement 935

The split characters are case sensitive.

Tip Use the CURVELABELSPLITJUSTIFY= option to specify the

justification of the strings in the step-line label block.

CURVELABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the step-line label text.

TRUE
drops the split characters from the step-line label text.

FALSE
includes the split characters in the step-line label text.

When CURVELABELSPLIT=TRUE and

CURVELABELSPLITCHARDROP=FALSE, each split character remains as the
last character in the current line. The characters that follow the split
character, up to and including the next split character, are then wrapped to
the next line.

The following figure shows an example of a step-line label with the following
specifications:
n CURVELABELPOSITION=MAX

n CURVELABEL="Product*Group*A"

n CURVELABELSPLIT=TRUE

n CURVELABELSPLITCHARDROP=TRUE | FALSE

n CURVELABELSPLITCHAR="*"

Note: The horizontal line to the left of the label represents the maximum end of
the step line for reference.

When CURVELABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When CURVELABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the step-line label.

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction The CURVELABELSPLITCHAR= option specifies the split

characters.
936 Chapter 6 / Plot Statements

See “boolean ” on page xii for other Boolean values that you can use.

CURVELABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the step-line label block.

AUTO
justifies the labels based on the CURVELABELPOSITION= option.

CURVELABELPOSITION= Value Justification

MAX or END LEFT

MIN or START RIGHT

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which CURVELABELPOSITION=MAX.

Note: The horizontal line to the left of each label represents the maximum end
of the step line for reference.

In this case, because CURVELABELPOSITION=MAX, AUTO left-justifies the

lines of text.

Default AUTO

Requirement The CURVELABEL= option and the CURVELABELSPLIT=TRUE

option must also be specified.

Interaction This option has no effect if CURVELABELPOSITION=AUTO.

DATALABEL=column | expression
specifies a column that supplies values for the data point labels.

Note The label positions are adjusted to prevent the labels from overlapping.

DATALABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the data labels. When the GROUP=
option is used, each distinct group value might be represented by a different
color. The data label that is associated with the group is assigned the group
color. This option can be used to specify a single color for all data labels in a
plot, without affecting items that have the group color, such as error bars and
marker symbols.
 STEPPLOT Statement 937

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interactions For this option to take effect, the DATALABEL= option must also
be specified.

This option’s COLOR= setting overrides the colors indicated by the

GROUP= option.

The LINEATTRS= on page 948 COLOR= suboption overrides the

default data label color for grouped plots.

Notes When the DATALABELPOSITION=AUTO option is in effect, in

some cases, the data label font size might be reduced in order to
avoid overlapping labels and markers.

If an attribute map is used with a grouped plot and the attribute

map specifies line color, the line color specified in the attribute map
overrides the default data label color.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the data points and markers,
when displayed.

AUTO
uses a data-label collision avoidance algorithm to position the labels.

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at the specified split characters.
938 Chapter 6 / Plot Statements

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
 STEPPLOT Statement 939

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

940 Chapter 6 / Plot Statements

DATALABELPOSITION= Value Justification

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, the bottom center of the text box
is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the step plot lines. The following figure
shows a step plot with each of the skins applied.
 STEPPLOT Statement 941

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data skin
is not applied to that plot. In that case, use the DATASKINMAX=
option in your ODS GRAPHICS statement to increase the maximum
limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the step lines, markers, error bars,
step-line labels, and data labels, when displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

DISCRETEOFFSET=number
specifies an amount to offset all step lines and markers from discrete X values
when graphing multiple response variables side by side on a common axis.

Default 0 (no offset, all step lines and markers are centered on the discrete
X values)

Range -0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. A positive offset is to the right. If the layout’s axis
options set REVERSE=TRUE, then the offset direction is also
reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=(display-options) | STANDARD | ALL

specifies which step-line features to display.

(display-options)
a space-separated list of one or more options enclosed in parentheses.

Currently, only the MARKERS option is supported, which displays a step line
with markers.

STANDARD
displays a step line without markers.
942 Chapter 6 / Plot Statements

ALL
displays a step line with markers.

Default STANDARD

Tip Use the MARKERATTRS= and LINEATTRS= options to control the

appearance of the line and markers.

ERRORBARATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the error bars that are associated with the data
points.

Defaults For ungrouped data, the GraphError style element ContrastColor,

LineStyle, and LineThickness attributes.

For grouped data, a GraphData1–GraphDataN style element

ContrastColor attribute, and the GraphDataDefault style elemement
LineStyle and LineThickness attributes.

Interaction For this option to take effect, error bars must be displayed by the
ERRORLOWER= , ERRORUPPER= options.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ERRORBARCAPSCALE=positive-number
specifies a positive number to use as a multiplier for determining the default
width of the error-bar caps.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default 1

Interaction This option is ignored when ERRORBARCAPSHAPE=NONE is in

effect.

Tips Use the ERRORBARCAPSHAPE= option to specify the cap shape.

The thickness (height) of the SERIF, LINE, and BRACKET is

determined by the thickness specified in the ERRORBARATTRS=
option. The cap line pattern is always solid.

Examples Double the size of the error-bar caps:

errorbarcapscale=2

Reduce the size of the error-bar caps by 50%:

errorbarcapscale=0.5

ERRORBARCAPSHAPE=SERIF | NONE
specifies whether the error bars have a serif cap.
 STEPPLOT Statement 943

Interaction Starting with SAS 9.4M5, when ERRORBARCAPSCALE= is in effect,

the display width for SERIF caps is the cap's default width multiplied
by the value specified in ERRORBARCAPSCALE=.

Tip The appearance of the error bars is controlled by the

ERRORBARATTRS= option.

ERRORLOWER=numeric-column | expression
specifies the values of the lower endpoints on the Y error bars.

Default The lower segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction When this option is specified, the DATALABEL= on page 308 and
DATALABELTYPE= on page 314 options are ignored, and data
labels are not displayed.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

ERRORUPPER=numeric-column | expression
specifies the values of the upper endpoints on the Y error bars.

Default The upper segment of the error bars is not drawn.

Requirement The error bar values must be absolute data values, not data values
relative to the value of the bar.

Interaction When this option is specified, the DATALABEL= on page 308 and
DATALABELTYPE= on page 314 options are ignored, and data
labels are not displayed.

Tip You can use the ERRORBARATTRS= option to control the

appearance of the error bars.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.

TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.

When this option is TRUE, the fill color and outline color for filled markers
are determined in the following ways:
n If the GROUP= option is specified, then by default, the fill color is derived
from the GraphData1–GraphDataN style elements Color attribute, and
944 Chapter 6 / Plot Statements

the marker outlined color is derived from the GraphData1–GraphDataN

style elements ContrastColor attribute.
n If the GROUP= option is not specified, then the marker fill is drawn by
using the MARKERFILLATTRS= specification, and the outline is drawn by
using the MARKEROUTLINEATTRS= specification.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

Tip To specify the marker fill and outline colors for a non-grouped plot, set
this option to TRUE, and then use the MARKERFILLATTRS= and
MARKEROUTLINEATTRS= options to specify the colors.

See GROUP= on page 944

MARKERFILLATTRS= on page 951

MARKEROUTLINEATTRS= on page 952

“boolean ” on page xii for other Boolean values that you can use.

GROUP=column | discrete-attr-var | expression

creates a distinct set of lines, markers, and data labels for each unique group
value of the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Default Each distinct group value might be represented in the plot by a

different combination of color, line pattern, and marker symbol.
Lines and markers vary according to the ContrastColor, LineStyle,
and MarkerSymbol attributes of the GraphData1–GraphDataN and
GraphMissing style elements. Line thickness (for grouped and
ungrouped data) is controlled by the LINEATTRS= option.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned
line cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000.
If the limit is exceeded, the plot is not drawn and a note is written
to the SAS log stating that the limit has been exceeded. In that
case, increase LINEPATTERNOBSMAX= to the value suggested in
the note or change the line pattern to SOLID.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of marker
symbols, colors, and line patterns.
 STEPPLOT Statement 945

The marker size is set by the MARKERATTRS= option.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips The representations that are used to identify the groups can be
overridden. For example, each distinct group value is often
represented by a different line pattern, but the
LINEATTRS=(PATTERN=pattern) option could be used to assign
the same line pattern to all of the plot’s line patterns, letting line
color indicate group values. Likewise, LINEATTRS=(COLOR=color)
could be used to assign the same color to all lines, letting line
pattern indicate group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

GROUPDISPLAY=OVERLAY | CLUSTER
specifies whether grouped step lines and markers are overlaid or clustered
around the category midpoints.

OVERLAY
centers the step lines and markers for matching category values on the
midpoints.

The step lines in each set of group values are superimposed on each other.

CLUSTER
clusters the step lines and markers for matching category values around the
midpoints on a discrete axis or around the intervals on an interval axis.

Each cluster of group values is centered at the midpoint for the category.

Default OVERLAY

Restriction For this option to take effect, the GROUP= option must also be
specified.

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

GROUPORDER=DATA | REVERSEDATA | ASCENDING | DESCENDING

specifies the ordering of the groups within a category.
946 Chapter 6 / Plot Statements

DATA
orders the groups within a category in the group-column data order.

REVERSEDATA
orders the groups within a category in the reverse group-column data order.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip This option is useful when you want to reverse the category axis.

ASCENDING
orders the groups within a category in ascending order.

DESCENDING
orders the groups within a category in descending order.

Defaults DATA if the data is a SAS data set.

ASCENDING if the data is a CAS in-memory table. Sorting is based

on nonformatted values for numeric group values or on formatted
values for character group values.

Interactions This option is ignored if the GROUP= option is not also specified.

By default, the groups in the legend are shown in the order that is
specified in GROUPORDER.

Notes When plotting a SAS data set, attributes such as color, symbol, and
pattern are assigned to each group in DATA order by default. When
plotting a CAS in-memory table, attributes are assigned in
ascending order of the group column character values or
unformatted numeric values by default.

The ASCENDING and DESCENDING settings linguistically sort the

group values within each category (or X value) for display position
purposes only. For numeric data, the order is based on the
unformatted values. For character data, the order is based on the
formatted values. The data order of the observations and the visual
attributes that are assigned to the group values remain unchanged.

In SAS 9.4M6 and earlier releases, when data stored in a CAS in-
memory table is used and DATA or REVERSEDATA is specified,
unpredictable results might occur. To generate consistent graphs in
that case, specify ASCENDING or DESCENDING. Starting with SAS
Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-memory
table is used and DATA or REVERSEDATA is specified, it is ignored,
and ASCENDING is used instead.

Tip Use the INDEX= option to alter the default sequence of visual
attributes that is assigned to the groups.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.
 STEPPLOT Statement 947

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color, marker symbol, and line
pattern) to one of the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

JOIN=TRUE | FALSE
specifies whether the steps are connected.

Default TRUE

See “boolean ” on page xii for other Boolean values that you can use.

JUSTIFY=(LEFT | CENTER | RIGHT)

specifies the location of the data point relative to the step.
948 Chapter 6 / Plot Statements

Default LEFT

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Y-column label. If a label is not defined, then the Y-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the step line connecting the data points.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor and LineStyle attributes of the

GraphData1–GraphDataN style elements, and the
GraphDataDefault:LineThickness style reference.

Restriction Starting with SAS 9.4M5, the number of vertices for a patterned line
cannot exceed the maximum specified by the ODS GRAPHICS
statement option LINEPATTERNOBSMAX=. The default is 10,000. If
the limit is exceeded, the plot is not drawn and a note is written to
the SAS log stating that the limit has been exceeded. In that case,
increase LINEPATTERNOBSMAX= to the value suggested in the
note or change the line pattern to SOLID.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness. By default, this option determines the thickness of
the line that represents the maximum response column value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default Ten times the thickness that is specified by the GraphDataDefault

style element LineThickness attribute.
 STEPPLOT Statement 949

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached.
The line thickness for response values that exceed the
LINETHICKNESSMAXRESPONSE= value are set to the value that
is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.

See “dimension” on page xiii

“Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value
is set to the value specified in the LINETHICKNESSMAX= option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.
950 Chapter 6 / Plot Statements

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum line

thickness.

See “dimension” on page xiii

“Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing,
this option is ignored. In these cases, the default line thickness is
used for all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

Notes The line thicknesses are not reflected in a discrete legend.

The LINETHICKNESSRESPONSE= values are assumed to be

constant for each group value in a grouped plot and for the entire
plot in an ungrouped plot. If the LINETHICKNESSRESPONSE
column has multiple values for a single GROUP value or ungrouped
plot, unpredictable results might occur.

See “Example 3: Series Plot with Line-Thickness Response and

Arrowheads” on page 922 for an example of how to use this option.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the attributes of the data markers.

Defaults For non-grouped data, GraphDataDefault style element.

For grouped data, the MarkerSymbol and ContrastColor attributes

of the GraphData1–GraphDataN style elements, and the
GraphDataDefault:MarkerSize style reference.
 STEPPLOT Statement 951

Interactions If FILLEDOUTLINEDMARKERS=TRUE, then this option’s COLOR=

suboption is ignored. In that case, to specify the marker fill color,
use the MARKERFILLATTRS= option instead.

This option’s COLOR= suboption overrides the default behavior for

grouped data. When the COLOR= suboption is specified in that
case, all markers have the same color, and the marker symbol alone
distinguishes the markers.

This option’s SYMBOL= suboption overrides the default behavior

for grouped data. When the SYMBOL= suboption is specified in
that case, all markers have the same symbol, and the symbol color
alone distinguishes the markers.

The TRANSPARENCY= fill option overrides this option’s

DATATRANSPARENCY= suboption.

This option is ignored if the DISPLAY= option disables the display

of the markers.

If the DATASKIN= option is in effect, then the data skin determines

the marker outlines. Any outline-related settings from the current
ODS style or from the marker attribute options are ignored.

Note When style-element is specified, only the style element’s

MARKERSYMBOL, CONTRASTCOLOR, and MARKERSIZE
attributes are used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

MARKERFILLATTRS=style-element | (fill-options)
specifies the appearance of the filled markers.

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element

For grouped data, the Color attribute of a GraphData1–GraphDataN

style element

Restriction The TRANSPARENCY= fill option is ignored. Use the

MARKERATTRS= option to set the marker transparency.

Interaction This option is in effect only when

FILLEDOUTLINEDMARKERS=TRUE and the DISPLAY= option
enables fill display.

Note When style-element is specified, only the style element’s Color

attribute is used.

See “General Syntax for Attribute Options” on page 1599

“Fill Color Options” on page 1600

952 Chapter 6 / Plot Statements

MARKEROUTLINEATTRS=style-element | (line-options)
specifies the appearance of the marker outlines.

Defaults For non-grouped data, the GraphOutlines style element.

For grouped data, the LineThickness attritube of the GraphOutlines

style element and the ContrastColor attribute of a GraphData1–
GraphDataN style element.

Restriction The line style of the marker outline is always solid.

Interaction This option is ignored when a data skin is applied by the current
style or by the DATASKIN= option. In the latter case, the outline is
set by the data skin.

Note When style-element is specified, only the style element’s

CONTRASTCOLOR and LINETHICKNESS attributes are used.

See “General Syntax for Attribute Options” on page 1599

“Line Options” on page 1602

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046
 STEPPLOT Statement 953

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, CURVELABEL, DATALABEL,
ERRORLOWER, ERRORUPPER, GROUP, and INDEX.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the step
line. If this option is used, then it replaces all of the information that is displayed
by default. Roles for columns that do not contribute to the step plot can be
specified along with roles that do.

(role-list)
an ordered, space-separated list of unique STEPPLOT and user-defined
roles.

STEPPLOT roles include X , Y , CURVELABEL , DATALABEL ,

ERRORLOWER , ERRORUPPER , and GROUP .

User-defined roles are defined with the ROLENAME= option.

Note CURVELABEL is considered a role only when it is assigned a

column of values. It is not considered a role and does not display
data tips when assigned a string.

Example The following example displays data tips for the columns assigned
to the roles X and Y as well as the column Obs, which is not
assigned to any pre-defined STEPPLOT role. The Obs column must
first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X , Y , DATALABEL , ERRORLOWER ,
ERRORUPPER , and GROUP .
954 Chapter 6 / Plot Statements

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a step line segment is selected.
 STEPPLOT Statement 955

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each step line segment that is to have an active link.

Example http://www.sas.com

Requirement To generate selectable segments, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Note When selecting a portion of a segment that is not an endpoint, the

nearest segment endpoint is used.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding segment is selected.

The URL value can be the same for any X and Y pairs, meaning that
the same action is taken when the segment for those X and Y pairs
are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.
 956 Chapter 6 / Plot Statements

Details
For character columns, the X axis is always discrete. For numeric columns, both the
X and Y axes are linear by default. You can change the axis type for numeric axes
with the XAXISOPTS= and YAXISOPTS= options of the containing OVERLAY
layout.

By default, the STEPPLOT statement uses the X values in data order. You can use
the CONNECTORDER= option to change the order.

Example: STEPPLOT Statement

The following graph was generated by the “Example Program” on page 957:

Input Data
Here is the DATA step code that generates the input data for this example.
data Study;
input Group : $10. Time Status @@;
label Time="Time (Days)";
datalines;
Low-Risk 2569 0 Low-Risk 2506 0 Low-Risk 2409 0
Low-Risk 2218 0 Low-Risk 1857 0 Low-Risk 1829 0
Low-Risk 1562 0 Low-Risk 1470 0 Low-Risk 1363 0
Low-Risk 1030 0 Low-Risk 860 0 Low-Risk 1258 0
 STEPPLOT Statement 957

Low-Risk 2246 0 Low-Risk 1870 0 Low-Risk 1799 0

Low-Risk 1709 0 Low-Risk 1674 0 Low-Risk 1568 0
Low-Risk 1527 0 Low-Risk 1324 0 Low-Risk 957 0
Low-Risk 932 0 Low-Risk 847 0 Low-Risk 848 0
Low-Risk 1850 0 Low-Risk 1843 0 Low-Risk 1535 0
Low-Risk 1447 0 Low-Risk 1384 0 Low-Risk 414 1
Low-Risk 2204 1 Low-Risk 1063 1 Low-Risk 481 1
Low-Risk 105 1 Low-Risk 641 1 Low-Risk 390 1
Low-Risk 288 1 Low-Risk 421 1 Low-Risk 79 1
Low-Risk 748 1 Low-Risk 486 1 Low-Risk 48 1
Low-Risk 272 1 Low-Risk 1074 1 Low-Risk 381 1
Low-Risk 10 1 Low-Risk 53 1 Low-Risk 80 1
Low-Risk 35 1 Low-Risk 248 1 Low-Risk 704 1
Low-Risk 211 1 Low-Risk 219 1 Low-Risk 606 1
High-Risk 2640 0 High-Risk 2430 0 High-Risk 2252 0
High-Risk 2140 0 High-Risk 2133 0 High-Risk 1238 0
High-Risk 1631 0 High-Risk 2024 0 High-Risk 1345 0
High-Risk 1136 0 High-Risk 845 0 High-Risk 422 1
High-Risk 162 1 High-Risk 84 1 High-Risk 100 1
High-Risk 2 1 High-Risk 47 1 High-Risk 242 1
High-Risk 456 1 High-Risk 268 1 High-Risk 318 1
High-Risk 32 1 High-Risk 467 1 High-Risk 47 1
High-Risk 390 1 High-Risk 183 1 High-Risk 105 1
High-Risk 115 1 High-Risk 164 1 High-Risk 93 1
High-Risk 120 1 High-Risk 80 1 High-Risk 677 1
High-Risk 64 1 High-Risk 168 1 High-Risk 74 1
High-Risk 16 1 High-Risk 157 1 High-Risk 625 1
High-Risk 48 1 High-Risk 273 1 High-Risk 63 1
High-Risk 76 1 High-Risk 113 1 High-Risk 363 1
;
run;

Example Program
Here is the program code for this example.
proc template;
define statgraph stepplot;
begingraph;
entrytitle "Kaplan-Meier Survival Plot";
layout overlay /
yaxisopts=(linearopts=(viewmin=0 viewmax=1));
stepplot x=Time y=Survival /
group=Stratum name="step";
scatterplot x=Time y=Censored / name="scat"
legendlabel="Censored" markerattrs=(symbol=plus);
discretelegend "step" / location=inside
halign=right valign=top across=1;
discretelegend "scat" /
location=inside halign=center valign=top;
endlayout;
endgraph;
end;
run;
 958 Chapter 6 / Plot Statements

ods graphics;
ods exclude all;
ods output survivalplot=plotdata;
proc lifetest data=Study plots=(survival);
time Time * Status(0);
survival;
strata Group;
run;
ods select all;

proc sgrender data=plotdata template=stepplot;

run;

SURFACEPLOTPARM Statement
Creates a three-dimensional surface representing a response variable evaluated over a grid of X and
Y values.

Restriction: The SURFACEPLOTPARM statement does not support data tips.

Requirements: For surface plots, the input data should form an evenly spaced grid of horizontal
values (X and Y) and one or more vertical values (Z) for each combination.
If the plot data is a SAS data set, the input data must be sorted by Y and X in order
to obtain the correct lighting. If the plot data is a CAS in-memory table, the data is
sorted automatically before it is plotted.

Syntax
SURFACEPLOTPARM X=numeric-column | expression
Y=numeric-column | expression
Z=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
SURFACECOLORGRADIENT= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
starting with SAS 9.4M2, specifies the column or range attribute map
variable to use to determine the surface colors.
DATATRANSPARENCY=number
specifies the degree of the transparency of the surface.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the filled surface or the wire-frame mesh.
 SURFACEPLOTPARM Statement 959

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
SURFACECOLORGRADIENT=numeric-column | range-attr-var
specifies the column or range attribute map variable that is used to
determine the surface colors (in SAS 9.4M1 and earlier releases).
SURFACETYPE=FILLGRID | FILL | WIREFRAME
specifies how the surface is displayed.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Required Arguments
X=numeric-column | expression
specifies the X coordinates of the grid.

Y=numeric-column | expression
specifies the Y coordinates of the grid.

Z=numeric-column | expression
specifies the height of response values.

Note The input data should be sorted by both 1) the Y column and 2) the X
column. The sort direction for Y should be ascending. The sort direction
of X be either ascending or descending.

Optional Arguments
COLORMODEL=style-element | (color-list)
specifies a color ramp that is to be used with the COLORRESPONSE= or
SURFACECOLORGRADIENT= option.

style-element
specifies the name of a style element.

The style element should contain these style attributes:

STARTCOLOR specifies a color for the smallest data value of the

COLORRESPONSE= or
SURFACECOLORGRADIENT= column.
960 Chapter 6 / Plot Statements

NEUTRALCOLOR specifies a color for the midpoint of the range of the

COLORRESPONSE= or
SURFACECOLORGRADIENT= column.
ENDCOLOR specifies a color for the highest data value of the
COLORRESPONSE= or
SURFACECOLORGRADIENT= column.

(color-list)
specifies a space-separated list of colors to use in the color ramp.

You can use style attribute references such as GraphData3:Color, color

names, or RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color.
The list can contain a mix of style attribute references, color names, and
color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element.

Interaction For this option to have any effect, the COLORRESPONSE= or

SURFACECOLORGRADIENT= option must also be used.

Tip The REVERSECOLORMODEL= option can be used to reverse the

start and end colors of the ramp assigned to the color model.

COLORRESPONSE=numeric-column | range-attr-var | expression

starting with SAS 9.4M2, specifies the column or range attribute map variable to
use to determine the surface colors.

Note: Starting with SAS 9.4M2, the COLORRESPONSE= option replaces the
SURFACECOLORGRADIENT= option. The syntax and functionality are the
same. The SURFACECOLORGRADIENT= option is still honored, but the
COLORRESPONSE= option is preferred.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.
 SURFACEPLOTPARM Statement 961

You can use this option to add a second response variable to an analysis. For an
example, see “Example 1: Surface Plot” on page 963.

Interaction Suboption COLOR= in the FILLATTRS= option overrides this option

for the fill colors.

Tip To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

DATATRANSPARENCY=number
specifies the degree of the transparency of the surface.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the color of the filled surface or the wire-frame mesh.

Default The GraphDataDefault:Color style reference.

Interaction The COLORRESPONSE= or SURFACECOLORGRADIENT= option is

ignored if this option is specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The Z-column label. If a label is not defined, then the Z-column
name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.
962 Chapter 6 / Plot Statements

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an OVERLAY3D layout contribute to a common axis.

Default FALSE

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

SURFACECOLORGRADIENT=numeric-column | range-attr-var
specifies the column or range attribute map variable that is used to determine
the surface colors (in SAS 9.4M1 and earlier releases).

Note: Starting with SAS 9.4M2, the SURFACECOLORGRADIENT= option is

deprecated and replaced with the COLORRESPONSE= option. The syntax and
functionality are the same. The SURFACECOLORGRADIENT= option is still
honored, but the COLORRESPONSE= option is preferred.

See COLORRESPONSE= on page 960

SURFACETYPE=FILLGRID | FILL | WIREFRAME

specifies how the surface is displayed.

FILLGRID
specifies a filled surface with superimposed grid lines

FILL
specifies a filled surface without grid lines

WIREFRAME
specifies an unfilled surface with grid lines

Default FILLGRID

Details
The SURFACEPLOTPARM statement assumes that the Z response values have
been provided for a uniform X-Y grid. Missing Z values leave a “hole” in the surface.
The observations in the input data set should form an evenly spaced grid of
horizontal (X and Y) values and one vertical (Z) value for each of these
 SURFACEPLOTPARM Statement 963

combinations. The observations should be in sorted order of Y and X to obtain an

accurate plot.

The G3GRID procedure (requires a SAS/GRAPH license) can be used to interpolate

the necessary values to produce a data set with nonmissing Z values for every
combination of X and Y. The G3GRID procedure can also smooth data with spline
interpolations. For further details, see the documentation for PROC G3GRID in the
SAS/GRAPH: Reference.

The KDE procedure can produce an output data set of gridded X-Y values where the
Z value is computed to be a Kernel Density Estimate of the distribution of X and Y.
For more information, see The KDE Procedure in SAS/STAT User’s Guide.

Examples

Example 1: Surface Plot

The following graph was generated by the “Example Program” on page 963:

Example Program
Here is the code for this example. The COLORRESPONSE= option is valid starting
with SAS 9.4M2. For prior releases, use the SURFACECOLORGRADIENT= option
instead.
proc template;
define statgraph surfaceplotparm;
 964 Chapter 6 / Plot Statements

begingraph;
entrytitle "Surface Plot of Lake Bed";
layout overlay3d / cube=false;
surfaceplotparm x=length y=width z=depth /
reversecolormodel=true
colorresponse=depth
colormodel=twocoloraltramp;
endlayout;
endgraph;
end;

/* create gridded data for surface

* proc g3grid is a sas/graph procedure */
proc g3grid data=sashelp.lake out=gridded;
grid width*length = depth / naxis1=75 naxis2=75;
run;

proc sgrender data= gridded template=surfaceplotparm;

run;

Example 2: Spline Interpolation

The following graph was generated by the “Example Program” on page 965:

Input Data
Here is the SAS code that generates the data for this example.
data nums;
do i=1 to 30;
 SURFACEPLOTPARM Statement 965

X=10*ranuni(33)-5;
Y=10*ranuni(33)-5;
Z=sin(sqrt(x*x+y*y));
output;
end;
run;
proc g3grid data=nums out=gridded;
grid y*x=z / spline
axis1=-5 to 5 by 0.1
axis2=-5 to 5 by 0.1;
run;
proc sort data=gridded; by y x; run;

Example Program
proc template;
define statgraph g3grid_surface;
begingraph;
entrytitle "Spline Interpolation";
layout overlay3d;
surfaceplotparm x=x y=y z=z /
surfacetype=fill;
endlayout;
endgraph;
end;
run;

proc sgrender data=gridded template=g3grid_surface;

run;

Example 3: Kernel Density Estimate

The following graph was generated by the “Example Program” on page 966:
 966 Chapter 6 / Plot Statements

Input Data
Here is the SAS code that generates the data for this example.
data nums;
do i=1 to 30;
X=10*ranuni(33)-5;
Y=10*ranuni(33)-5;
Z=sin(sqrt(x*x+y*y));
output;
end;
run;
proc kde data=nums;
bivar x y / ngrid=100
gridl=-5 gridu=5
out=binned(rename=(value1=X value2=Y));
run;
proc sort data=binned; by y x;
label x="X" y="Y";
run;

Example Program
proc template;
define statgraph kde_surface;
begingraph;
entrytitle "Kernel Density Estimate";
 TEXTPLOT Statement 967

layout overlay3d;
surfaceplotparm x=x y=y z=density /
surfacetype=fill;
endlayout;
endgraph;
end;
run;

proc sgrender data=binned template=kde_surface;

run;

TEXTPLOT Statement
Displays text values at specific X and Y locations in the graph.

Note: This feature applies to SAS 9.4M2 and to later releases.

Tip: Use the TEXTPLOT statement, rather than the SCATTERPLOT statement with the
MARKERCHARACTER= option, when you want more control over the appearance
of the text. The TEXTPLOT statement enables you to rotate the text to any angle,
manage the text position, split the text into multiple lines, display a bounding box
around the text, add a back-light effect to the text, and so on.

Syntax
TEXTPLOT X=column | expression
Y=column | expression
TEXT=column | expression </options>;

Summary of Optional Arguments

Appearance options
BACKLIGHT=number | AUTO
specifies a back-light effect for the marker text.
CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction
of the midpoint spacing. On an interval axis, specifies the width of the
group clusters as a fraction of the minimum interval between adjacent
data values.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute map variable to use to determine
the text-marker text colors.
CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)
968 Chapter 6 / Plot Statements

specifies whether space requirements for this plot contribute to the

calculation of the axis offsets.
DATATRANSPARENCY=number
specifies the degree of the transparency of the text.
DISPLAY=(display-options) | STANDARD | ALL
specifies the text features to display.
FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled areas of the text.
INDEX=positive-integer-column | expression
specifies indices for mapping marker text color to one of the GraphData1–
GraphDataN style elements.
OUTLINEATTRS=style-element | style-element(line-options) | (line-options)
specifies the appearance of the text-marker outlines.
PAD=dimension | (pad-options)
specifies the amount of extra space to add inside the text-marker border.
REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by
either the ODS style that is in effect or by the COLORMODEL= option.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font properties of the text-marker text.

Axes options
CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when
GROUPDISPLAY=CLUSTER.
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over
the text values.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Experimental options
OUTFILE=fileref | "filename"
 TEXTPLOT Statement 969

specifies a file for storing information about the text bounding-box for
each text value in the column specified in the OUTID= option.
OUTID=column | expression
specifies a column that contains text values to write to the file specified
in the OUTFILE= option.

Label options
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
DISCRETEOFFSET=number
specifies the distance to offset all text values from discrete X values,
discrete Y values, or both.
GROUP=column | discrete-attr-var | expression
creates a separate text value for each unique group value in the specified
column.
GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

ODS options
URL=string-column
specifies an HTML page to display when a text value is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Text options
FORMAT=format
specifies a SAS format or a user-defined format for the TEXT= column.
POSITION=BOTTOM | BOTTOMLEFT | BOTTOMRIGHT | CENTER | LEFT |
RIGHT | TOP | TOPLEFT | TOPRIGHT | keyword-column
specifies the position of the text value with respect to the location of the
data point.
POSITIONOFFSETX=number | numeric-column
specifies the positive or negative X offset for an individual text item.
POSITIONOFFSETY=number | numeric-column
specifies the positive or negative Y offset for an individual text item.
ROTATE=number | numeric-column
specifies the angle of rotation, in degrees, for the text values.
SIZEMAX=dimension
specifies the maximum font size for a text marker when a response
variable is used to size the text-marker font.
SIZEMAXRESPONSE=numeric | scalar-numeric-expression
970 Chapter 6 / Plot Statements

specifies the response value that corresponds to the maximum font size
for text markers.
SIZEMIN=dimension
specifies the minimum font size for text markers when a response
variable is used to size the font for text values.
SIZERESPONSE=numeric-column | numeric-expression
specifies a response column that is used to determine the font size for
each text value.
SPLITCHAR="character-list"
specifies one or more characters on which the text-marker text can be
split.
SPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the text-marker
text.
SPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the lines of text in the text-marker text
blocks.
SPLITPOLICY=NONE | SPLIT | SPLITALWAYS
specifies a policy for avoiding collisions among the text values.
SPLITWIDTH=AUTO | width-in-characters
specifies the maximum width of each split line.
STRIP=TRUE | FALSE
specifies whether leading and trailing blanks should be stripped from the
text-marker text before it is displayed.
VCENTER=BBOX | BASELINE
specifies whether the text is vertically centered with respect to the text
bounding box or to the text baseline.

Required Arguments
X=column
specifies the column for the X values.

Y=column
specifies the column for the Y values.

TEXT=column
specifies the column for the text values that are to be used for the markers.

Optional Arguments
BACKLIGHT=number | AUTO
specifies a back-light effect for the marker text. The effect is applied only to the
marker text.

number
specifies the degree of the back-light effect.

Range 0–1, where 0 specifies no effect and 1 specifies maximum effect

AUTO
the system selects an appropriate level for the back-light effect.
 TEXTPLOT Statement 971

If the GROUP= or COLORRESPONSE= option is in effect, BACKLIGHT=0.75.

Otherwise, BACKLIGHT=0.5.

The following figure shows the effect on the text of an outlined, filled text
marker.

The back light is based on text color. For dark colors, a contrasting white back-
light effect is used. For lighter colors, a contrasting black back-light effect is
used. The following figure shows the two back-light types when BACKLIGHT=1.

Default 0 (no back-light effect)

Restriction Vector graphics output cannot be generated when the back-light

effect is applied. If you request vector graphics output and enable
the back-light effect, an image is generated instead.

Note If the background color is white, the white backlight effect for dark
text colors is not visible. Conversely, if the background color is
black, the black backlight effect for light text colors is not visible.

Tip The BACKLIGHT= option is most effective when the text color has a
low level of contrast with the background or when the background is
cluttered.

CLUSTERAXIS=AUTO | X | Y
specifies the axis to use for clustering groups when GROUPDISPLAY=CLUSTER.

AUTO
uses the discrete axis for clustering groups when only one axis is discrete.
Uses the X axis for clustering if both axes are discrete or interval.

X
Y
uses the X or Y axis for clustering groups.

Default AUTO

Interaction The GROUPDISPLAY= option must be set to CLUSTER for this

option to have any effect.
972 Chapter 6 / Plot Statements

CLUSTERWIDTH=number
on a discrete axis, specifies the width of the group clusters as a fraction of the
midpoint spacing. On an interval axis, specifies the width of the group clusters
as a fraction of the minimum interval between adjacent data values.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is
the maximum possible width

Interaction For this option to take effect, the GROUP= option must also be
specified, and the GROUPDISPLAY= option must be set to
CLUSTER.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

Tip To reverse the start and end colors of the ramp that is assigned to
the color model, use the REVERSECOLORMODEL= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute map variable to use to determine the
text-marker text colors.
 TEXTPLOT Statement 973

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

Tip For each range in the attribute map, the RANGECOLOR= or

RANGECOLORMODEL= option in the RANGE statement
determines the bounding-box fill colors. The RANGEALTCOLOR=
or RANGEALTCOLORMODEL= option determines the text colors
and the bounding-box outline colors.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

You can use this option to add a second response variable to an analysis.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the GROUP= option is ignored.

Prior to SAS 9.4M3, suboption COLOR= in the TEXTATTRS=,

FILLATTRS=, and OUTLINEATTRS= options overrides the color
attributes that are normally determined by this option.

Starting with SAS 9.4M3, interactions between this option and

suboption COLOR= in the TEXTATTRS=, FILLATTRS=, and
OUTLINEATTRS= options depend on the DISPLAY= option
settings. See “Response Colors in a Text Plot” on page 993.

Note The gradient in a continuous legend for this plot reflects the text
colors.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the OUTLINEATTRS= option to set the text bounding-box

color to a fixed color.

Use the FILLATTRS= option to set the text bounding-box fill color
to a fixed color or to modify the fill transparency.

Prior to SAS 9.4M3, when fill is displayed and the

COLORRESPONSE= option is in effect, the text color and the fill
color are derived from the color gradient, which makes the text
unreadable. In that case, use the BACKLIGHT= option to add a
backlight effect to the text, or use the TEXTATTRS= or
FILLATTRS= option to specify a different text or fill color.
974 Chapter 6 / Plot Statements

CONTRIBUTEOFFSETS=ALL | NONE | (axis-offset-list)

specifies whether space requirements for this plot contribute to the calculation
of the axis offsets. This plot’s layout container queries each of its plots for a
preferred offset and includes all of the offsets in the axis offset calculations. If
the DATALABEL= or MARKERCHARACTER= option is specified for this plot,
this plot might request a preferred offset that prevents the clipping of any data
labels or marker character strings that appear at the ends of the axes. The
requested offset is based on the longest string. If the label or marker character
lengths vary significantly, the result is wasted space when the shorter strings
appear near the ends of the axes. In that case, you can use the
CONTRIBUTEOFFSETS= option to modify or eliminate this plot’s contribution
to the offset calculations in order to reclaim that space.

ALL
the space requirements for this plot are contributed to the axis offset
calculations.

NONE
the space requirements for this plot are not contributed to the axis offset
calculations.

(axis-offset-list)
a space-delimited list of specific contributions that this plot makes to the
axis offset calculations.

The list is one or more of the following values, enclosed in parentheses:

XMAX
the space requirements for this plot are contributed to the X-axis offset
calculation for the maximum end.

XMIN
the space requirements for this plot are contributed to the X-axis offset
calculation for the minimum end.

YMAX
the space requirements for this plot are contributed to the Y-axis offset
calculation for the maximum end.

YMIN
the space requirements for this plot are contributed to the Y-axis offset
calculation for the minimum end.

Default ALL

Interaction Offsets that are set in the layout axis options are always honored,
regardless of the setting on this option.

Note This option does not affect offset requests from other plots.

DATATRANSPARENCY=number
specifies the degree of the transparency of the text.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

 TEXTPLOT Statement 975

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

DISCRETEOFFSET=number
specifies the distance to offset all text values from discrete X values, discrete Y
values, or both.

Default 0 (no offset, all text values are centered on the discrete X values, or
discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete tick marks. A positive offset is to the right on discrete X
values and is up on discrete Y values. If the layout’s axis options set
REVERSE=TRUE, then the offset direction is also reversed.

Restriction This option applies to discrete axes only. For nondiscrete axes, this
option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=(display-options) | STANDARD | ALL

specifies the text features to display.

(display-options)
a space-separated list of one or more display options, enclosed in
parentheses.

The list can include one or more of the following options:

FILL
displays filled bounding boxes around the text

Tip Use the FILLATTRS= option to modify the appearance of the

bounding-box fill.

OUTLINE
displays outlined bounding boxes around the text

Tip Use the OUTLINEATTRS= option to modify the appearance of the

bounding-box outline.

STANDARD
displays the text only

ALL
specifies FILL and OUTLINE

Default STANDARD

Restriction Vector graphics output cannot be generated when FILL or OUTLINE

is displayed. If you request vector graphics output and specify
976 Chapter 6 / Plot Statements

DISPLAY=FILL or DISPLAY=OUTLINE, an image is generated

instead.

Tip When fill is displayed and the COLORRESPONSE= option is in

effect, a low contrast between the fill color and the text color can
make some of the text difficult to read or unreadable. In that case,
use the TRANSPARENCY= suboption to adjust the fill transparency
or use the BACKLIGHT= option to add a backlight effect to the text.

FILLATTRS=style-element | (fill-options)
specifies the appearance of the filled areas of the text. When fill options are
specified, only the COLOR= and TRANSPARENCY= suboptions are honored.

Defaults For non-grouped data, the Color attribute of the GraphDataDefault

style element.

For grouped data, the Color attribute of the GraphData1–

GraphDataN style elements.

Interactions For this option to have any effect, the fill must be enabled by the
ODS style or by the DISPLAY= option.

When this option’s COLOR= suboption is specified with the

GROUP= or COLORRESPONSE= option, the bounding-box fill color
is set to the COLOR= specification for all of the text values.

Tip The DATATRANSPARENCY= option sets the transparency for the

text-marker text, fill, and outlines. You can combine this option
with DATATRANSPARENCY= to set one transparency for the text
and outlines but a different transparency for the fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FORMAT=format
specifies a SAS format or a user-defined format for the TEXT= column.

Default The format that is in effect for the column that is specified in the
TEXT= argument. If no format is in effect, BEST6 is used for numeric
columns.

Note Not all of the SAS formats are supported. See Appendix 4, “Format
Support in ODS Graphics,” on page 1621.

GROUP=column | discrete-attr-var | expression

creates a separate text value for each unique group value in the specified
column.

discrete-attr-var
specifies a discrete attribute map variable.
 TEXTPLOT Statement 977

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Tip For each value in the discrete attribute map, the VALUE
statement TEXTATTRS= option determines the text color, the
LINEATTRS= option determines the bounding-box outline color,
and the FILLATTRS= option determines the bounding-box fill
color and transparency.

Interactions The group values are mapped in the order in which they appear in
the data.

The INCLUDEMISSINGGROUP option controls whether missing

group values are considered a distinct group value.

The COLORRESPONSE= option overrides the group settings for the

text color of the text value. In that case, text color is set according
to the gradient.

The COLOR= suboption of the TEXTATTRS= option overrides the

group settings for the text color of the text value. In that case, the
text color for all of the text values is set to the COLOR=
specification.

Notes The legend entries for this plot reflect the text colors.

By default, for each text value, the bounding-box outline color is

set to the text color.

If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips Use the OUTLINEATTRS= option to set the text bounding-box

color to a fixed color.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

GROUPDISPLAY=OVERLAY | CLUSTER
specifies how marker groups are positioned for the coordinate pairs.

OVERLAY
draws text values for a given group value at the exact coordinate.

Depending on the data, markers at a given coordinate might overlap.

978 Chapter 6 / Plot Statements

CLUSTER
draws text values for a given group value adjacent to each other.

Restriction CLUSTER is supported only when at least one axis is discrete.

Default OVERLAY

Note When you plot a SAS data set, the items for each group value are
drawn in data order. When you plot a CAS in-memory table, they are
drawn in ascending order of the group column character values or
unformatted numeric values.

Tip Use the CLUSTERWIDTH= option to control the width of the clusters
when CLUSTER is in effect.

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.

INDEX=positive-integer-column | expression
specifies indices for mapping marker text color to one of the GraphData1–
GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.
 TEXTPLOT Statement 979

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The TEXT-column label. If a label is not defined, then the TEXT-
column name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

OUTFILE=fileref | "filename"
specifies a file for storing information about the text bounding-box for each text
value in the column specified in the OUTID= option. The information is written
in the comma-separated value (CSV) format.

CAUTION
OUTFILE= is an experimental option that is available in SAS 9.4M3. Do not
use this option in production jobs.

Interaction The OUTID= option must be specified for this option to have any
effect.

Note This option is a specialized feature for users who want to customize
the placement of the text in a text plot.

Tip Use the IMPORT procedure to import the CSV values into a SAS
data set.

See “Customizing Text Marker Placement (Experimental)” on page 994

OUTID=column | expression
specifies a column that contains text values to write to the file specified in the
OUTFILE= option.
980 Chapter 6 / Plot Statements

CAUTION
OUTID= is an experimental option that is available in SAS 9.4M3. Do not use
this option in production jobs.

Interaction The OUTFILE= option must be specified for this option to have any
effect.

Note This option is a specialized feature for users who want to customize
the placement of the text in a text plot.

See “Customizing Text Marker Placement (Experimental)” on page 994

OUTLINEATTRS=style-element | style-element(line-options) | (line-options)

specifies the appearance of the text-marker outlines.

Defaults For non-grouped data, the ContrastColor attribute of the

GraphOutlines style element.

For grouped data, text values use the ContrastColor attribute of

the GraphData1–GraphDataN style elements. If the
COLORRESPONSE= option is specified, the outline colors vary
according to the color gradient.

Restriction This option uses only the color specification in the style element or
line options. The line pattern and line thickness specifications are
ignored.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or by the DISPLAY= option.

When the COLOR= suboption is specified with the GROUP= option

or with the COLORRESPONSE= option, for all of the text values,
the bounding-box border color is set to the COLOR= specification.

Note This plot’s legend entries reflect the marker text color.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

COLOR= on page 1602

PAD=dimension | (pad-options)
specifies the amount of extra space to add inside the text-marker border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the text-marker border.

(pad-options)
a space-separated list of one or more pad options, enclosed in parentheses.

The list can include one or more of the following name-value-pair options:

LEFT=dimension TOP=dimension
RIGHT=dimension BOTTOM=dimension
 TEXTPLOT Statement 981

Default Padding is a fraction of the font height.

Note Sides that are not assigned padding are padded with the default
amount.

Tips This option is meaningful only when the DISPLAY= option displays
fills, outlines, or both.

Use pad-options to create non-uniform padding.

Note The default units for dimension are pixels.

See “dimension” on page xiii

POSITION=BOTTOM | BOTTOMLEFT | BOTTOMRIGHT | CENTER | LEFT | RIGHT

| TOP | TOPLEFT | TOPRIGHT | keyword-column
specifies the position of the text value with respect to the location of the data
point.

BOTTOM | BOTTOMLEFT | BOTTOMRIGHT | CENTER | LEFT | RIGHT | TOP |

TOPLEFT | TOPRIGHT
specifies a position option.

keyword-column
specifies a column that contains position options.

Interaction When keyword-column is specified and SPLITJUSTIFY= on page

986 is other than AUTO (default), the position values specified
in keyword-column are ignored and the text is positioned
BOTTOMRIGHT.

Note This feature applies to SAS 9.4M5 and to later releases.

The VCENTER= option specifies whether the position is relative to the text
bound box or the text baseline. See the VCENTER= option. By default, the
positions are relative to the text bounding box. The following figure shows the
effect of each of these values on the position of an outlined text value when
VCENTER=BBOX is in effect. The red dot indicates the marker data-point
location.
982 Chapter 6 / Plot Statements

When CENTER, LEFT, or RIGHT is specified, and VCENTER=BASELINE is in

effect, the positions are relative to the text baseline, as shown in the following
figure.

Default CENTER

POSITIONOFFSETX=number | numeric-column
specifies the positive or negative X offset for an individual text item.

Note: This feature applies to SAS 9.4M5 and to later releases.

The offset is relative to the item's data coordinate. The value can be greater
than 1. The offset unit is the font height of the current text item, which varies
according to the text's font size based on size response, and so on.

Notes A value of 1 implies an X offset that is the same distance as the current
text item's font height.

A reverse axis reverses the offset direction.

Tip This option is most useful when the text is placed around other objects
in the graph.

POSITIONOFFSETY=number | numeric-column
specifies the positive or negative Y offset for an individual text item.

Note: This feature applies to SAS 9.4M5 and to later releases.

The offset is relative to the item's data coordinate. The value can be greater
than 1. The offset unit is the font height of the current text item, which varies
according to the text's font size based on size response, and so on.

Notes A value of 1 implies an Y offset that is the same distance as the current
text item's font height.

A reverse axis reverses the offset direction.

Tip This option is most useful when the text is placed around other objects
in the graph.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE
 TEXTPLOT Statement 983

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

REVERSECOLORMODEL=TRUE | FALSE
specifies whether to reverse the gradient (color ramp) that is defined by either
the ODS style that is in effect or by the COLORMODEL= option.

Default FALSE

See COLORMODEL=

“boolean ” on page xii for other Boolean values that you can use.

ROTATE=number | numeric-column
specifies the angle of rotation, in degrees, for the text values. Positive angles
are measured in a counter-clockwise direction, and negative angles are
measured in a clockwise direction. You can use an angle that exceeds 360
degrees in absolute value.

Default 0

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, GROUP, and COLORRESPONSE.

SIZEMAX=dimension
specifies the maximum font size for a text marker when a response variable is
used to size the text-marker font. By default, the font size of the text values
that are associated with the maximum response column value is set to the value
specified by this option.
984 Chapter 6 / Plot Statements

Default Three times the size specified in the GraphDataText style element
for the maximum response column value marker.

Interactions The SIZERESPONSE= option must be specified for this option to

have any effect.

The SIZEMAXRESPONSE= option specifies the response value at

which the maximum font size for a text marker is reached. The font
size for all text values that exceed the SIZEMAXRESPONSE= value
is set to the value specified in this option.

Tips Use the SIZEMAXRESPONSE= option to specify the response value

at which the maximum font size for a text marker is reached.

Use the SIZEMIN= option to specify the minimum font size for text
markers.

See “dimension” on page xiii

SIZEMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum font size for text
markers.

Default The maximum value in the response column specified in the

SIZERESPONSE= option.

Interaction The SIZERESPONSE= option must be specified for this option to

have any effect.

Note The font size for all text values that exceed the maximum response
value is set to the value specified in the SIZEMAX= option.

SIZEMIN=dimension
specifies the minimum font size for text markers when a response variable is
used to size the font for text values.

Default The size specified in the GraphDataText style element for the
minimum response column value marker.

Interaction The SIZERESPONSE= option must be specified for this option to

have any effect.

Tip Use the SIZEMAX= option to specify the maximum text size.

See “dimension” on page xiii

SIZERESPONSE=numeric-column | numeric-expression
specifies a response column that is used to determine the font size for each text
value.

Default The size specified in the GraphDataText style element for all text
values.

Notes When the column value for an observation is 0, the font size for the
text value for that observation is set to the SIZEMIN= option value.
 TEXTPLOT Statement 985

When the column value for an observation is negative or missing, the

text value for that observation is not displayed in the text plot.
However, that observation still contributes to the axis ranges, legend,
and so on.

When all the column values are 0 or missing, this option is ignored. In
that case, the default font size is used for all of the text values.

Tip Use the SIZEMIN= and SIZEMAX= options to limit the minimum and
maximum font size for the text values.

SPLITCHAR="character-list"
specifies one or more characters on which the text-marker text can be split.

When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When SPLITPOLICY=SPLIT and a text value collision is detected, the text-

marker text is split on a specified split character only if a split is needed at that
point in order to make the text fit. In that case, a split might not occur on every
split character. When SPLITPOLICY=SPLITALWAYS, the text-marker text is
split unconditionally on every occurrence of a split character. If the text-marker
text does not contain any of the specified split characters, then the text is not
split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
splitchar="abc"

The SPLITPOLICY= option must specify SPLIT or SPLITALWAYS.

Interaction The SPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

SPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the text-marker text.
986 Chapter 6 / Plot Statements

TRUE
drops a split character from the text-marker text when a split occurs at that
character.

Split characters at which a split does not occur are left in place. The
SPLITPOLICY= option determines where the text is split. When
SPLITPOLICY=SPLIT, the text for each text value is split at a split character
only where a split is needed to fit the text in the available space. At each
split point, the split character is dropped, and the characters that follow the
split character, up to but not including the split character at the next split
point, are wrapped to the following line.

When SPLITPOLICY=SPLITALWAYS, the text-marker text is split at every

instance of a split character. All of the split characters are dropped. The
characters that follow each split character, up to but not including the next
split character, are wrapped to the next line.

FALSE
includes the split characters in the data label display.

The SPLITPOLICY= option determines how the split characters are

displayed. When SPLITPOLICY=SPLIT, each data label is split at a split
character only where a split is needed in order to make the label fit the
available space. A split might not occur at every split character in the label.
At each split point, the split character remains as the last character in the
current line. The characters that follow the split character, up to and
including the split character at the next split point, are then wrapped to the
following line. This process repeats until all of the text is displayed.

When SPLITPOLICY=SPLITALWAYS, the text for each marker is split at

every instance of a split character in the text regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line.

Default TRUE. A split character is dropped from the text-marker text when
a split occurs at that character.

Requirement The SPLITPOLICY= option must specify SPLIT or SPLITALWAYS.

See “boolean ” on page xii for other Boolean values that you can use.

SPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the lines of text in the text-marker text blocks.

AUTO
justifies the text based on the POSITION= option.

POSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

 TEXTPLOT Statement 987

CENTER
LEFT
RIGHT
justifies the text center, left, or right, as specified.

Default AUTO

SPLITPOLICY=NONE | SPLIT | SPLITALWAYS

specifies a policy for avoiding collisions among the text values.

NONE
does not split the text for text values that collide.

SPLIT
splits the text-marker text at a split character only if a split is needed at that
character in order to make the text fit the available space.

No split occurs at split characters that occur where a split is not needed. If
the text does not contain any of the specified split characters, then a split
does not occur. In that case, if the text does not fit the available space, then
it might collide with the adjoining text values.

See the SPLITCHAR= option for information about specifying the split
characters

SPLITALWAYS
splits the text-marker text at every occurrence of a split character.

If the text does not contain any of the specified split characters, then a split
does not occur.

See the SPLITCHAR= option for information about specifying the split
characters

Default NONE

SPLITWIDTH=AUTO | width-in-characters
specifies the maximum width of each split line.

AUTO
uses the width of the longest inter-split-character substring.

width-in-characters
specifies a fixed width, expressed as a character count.

Note When you specify a fixed width, the text-marker text is split
unconditionally every n characters, where n is the value of width-in-
characters.

Restriction This option has an effect only when SPLITPOLICY=SPLIT.

STRIP=TRUE | FALSE
specifies whether leading and trailing blanks should be stripped from the text-
marker text before it is displayed.
988 Chapter 6 / Plot Statements

Default FALSE

Tip Stripping the blanks from the numeric value strings helps center each
string relative to its data point.

See “boolean ” on page xii for other Boolean values that you can use.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font properties of the text-marker text.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the Font attribute of the GraphDataText style

element, and the ContrastColor attribute of a GraphDataN style
element.

Interactions When this option’s COLOR= suboption is used with the GROUP=
option, the COLOR= suboption specifies the color for all of the text
values.

This option’s COLOR= suboption overrides the COLORRESPONSE=

option. In that case, if a continuous legend is requested for the plot,
the legend is not drawn.

Note If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, and style), the
properties that are not specified are derived from the
GraphDataText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
for using a style-element.

“Text Options” on page 1606 for available text-options.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over the text
values. If you use this option, it replaces all of the information that is displayed
by default. Roles for columns that do not contribute to the scatter plot can be
specified along with roles that do.

(role-list)
specifies an ordered, space-separated list of unique TEXTPLOT roles and
user-defined roles. TEXTPLOT roles include X, Y, GROUP, and
COLORRESPONSE.

User-defined roles are defined with the ROLENAME= option.

Example The following example displays data tips for the columns assigned
to the roles X and Y, as well as the column Obs, which is not
assigned to any pre-defined TEXTPLOT role. The Obs column must
first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y)
 TEXTPLOT Statement 989

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to the following roles are automatically

included in the data tip information: X, Y, GROUP, and
COLORRESPONSE.

Restriction Data tips in text plots are not supported in the SVG output format.
To generate a text plot with data tips, use a bitmapped output
format such as PNG.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement with the IMAGEMAP option specified,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.
990 Chapter 6 / Plot Statements

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

URL=string-column
specifies an HTML page to display when a text value is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each text value that is an active link.

Example http://www.sas.com

Requirement To generate selectable text values, you must include an ODS

GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some X and Y pairs, meaning that
no action is taken when the corresponding point is selected.

The URL value can be the same for any X and Y pair. In that case,
the same action is taken when the points for that X and Y pair are
selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

VCENTER=BBOX | BASELINE
specifies whether the text is vertically centered with respect to the text
bounding box or to the text baseline.

BBOX
vertically centers the text with respect to its bounding box.

BASELINE
vertically centers the text with respect to the text baseline.

If the text is split into multiple lines, the text is centered on the baseline of
the last line of text.

Restriction This option is valid only when POSITION= is set to CENTER,

LEFT, or RIGHT. If POSITION= is set to any other value,
VCENTER=BBOX is used instead.
 TEXTPLOT Statement 991

Default BBOX

Tip Use the POSITION= option to specify the text position with respect to
the text bounding box or to the text baseline.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details

About Text Plots

A text plot is basically a scatter plot that uses text as markers. To generate a text
plot, you can use the SCATTERPLOT statement with the MARKERCHARACTER=
option or you can use the TEXTPLOT statement. However, the TEXTPLOT
statement provides more control over the marker text. With the TEXTPLOT
statement, the plot data provides the marker location as X and Y coordinates, and
the marker text. Each marker consists of the marker text enclosed in a bounding
box that is centered on its X and Y coordinates. By default, the bounding box is not
visible. You can specify options in the TEXTPLOT statement to add an outline,
background fill, or both, to the bounding box. The following figure shows a simple
example of a text marker with a filled and outlined bounding box.

Using additional options in the TEXTPLOT statement, you can rotate the text
around its anchor point to any angle, manage the text position, split the text into
multiple lines, add a back-light effect to the text, and so on. If you currently use the
SCATTERPLOT statement and the MARKERCHARACTER= option to generate text
plots, SAS recommends that you use the TEXTPLOT statement instead.
992 Chapter 6 / Plot Statements

Using Colors in a Text Plot

Text Marker Color Attributes

The TEXTPLOT statement enables you to control the following color attributes for
the text markers:
n text color

n bounding-box fill color

n bounding-box outline color

By default, only the marker text is displayed. Use the DISPLAY= option to specify
whether the bounding-box fill, bounding-box outline, or both are displayed.

The following table lists the options that you can use to control the color
attributes.

Option Color Attribute Default

TEXTATTRS=(COLOR=color) text color For non-grouped data, the

GraphDataText:color
style reference.
For grouped data, the
GraphDataN:ContrastCol
or style reference.

FILLATTRS=(COLOR=color) bounding-box fill For non-grouped data, the

color GraphDataDefault:color
style reference.
For non-grouped data, the
GraphDataN:color style
reference.

OUTLINEATTRS=(COLOR=color) bounding-box For non-grouped data, the

outline color GraphOutlines:ContrastC
olor style reference.
For non-grouped data, the
GraphDataN:ContrastCol
or style reference.

When a response variable is used, in some cases, an attribute option is overridden

by the response color. See “Response Colors in a Text Plot” on page 993.

Group Colors in a Text Plot

The TEXTPLOT statement supports the GROUP= option, which enables you to
create a separate text marker for each unique value in a specified column. When
the GROUP= option is in effect, by default, the text marker color attributes are
controlled by the GraphData1–GraphDataN style elements. You can use the
TEXTATTRS=, FILLATTRS=, and OUTLINEATTRS= options to set one or more color
 TEXTPLOT Statement 993

attributes to a fixed color. However, when you specify fixed colors, the contrast
between the fixed colors and the remaining variable colors might not be sufficient
for some group values. When both the GROUP= and COLORRESPONSE= options
are in effect, the COLORRESPONSE= option controls the color attributes.

Response Colors in a Text Plot

The TEXTPLOT statement supports the COLORRESPONSE= option, which enables
you to specify a numeric column or range attribute map variable that is used to
determine the text-marker colors. When a numeric column is specified, each unique
value is assigned a color from a color ramp that is specified in the COLORMODEL=
option. When a range attribute map variable is specified, each unique value is
assigned the attributes for that value that are defined in the attribute map. If the
attribute map does not define the color attributes for a value, that value is assigned
a color from the color ramp.

The TEXTPLOT statement COLORRESPONSE= option controls the color of the

marker text, fill, and outline, depending on which bounding-box attributes are
displayed. Prior to SAS 9.4M3, when the bounding-box fill and outline are
displayed, COLORRESPONSE= determines the color of the text and the bounding-
box fill. The bounding-box outline color in that case is determined by suboption
COLOR= in the OUTLINEATTRS= option. When only the outline is displayed, the
COLORRESPONSE= option determines the color of the text and the bounding-box
outline. You can use suboption COLOR= in the FILLATTRS=, OUTLINEATTRS=, and
TEXTATTRS= options to override the individual attributes.

Starting with SAS 9.4M3, when the bounding-box fill and outline are displayed, the
COLORRESPONSE= option determines the color of the text and of the bounding-
box outline. The bounding-box fill color in that case is determined by suboption
COLOR= in the FILLATRS= option. You can use suboption COLOR= in the
FILLATTRS=, OUTLINEATTRS=, and TEXTATTRS= options to override the
individual attributes, depending on which bounding-box attributes are displayed.
The following table lists the results when suboption COLOR= is specified in some
common color attribute option settings and the COLORRESPONSE= option is in
effect.

Attribute Option Result

TEXTATTRS= The result depends on the bounding-box display as follows:

n If the bounding-box fill, outline, or both are displayed, the
text color is set to its COLOR= specification. The bounding-
box outline color is determined by the GraphDataDefault
style element, and the bounding-box fill is assigned a color
from the color ramp.
n If no bounding-box attributes are displayed, suboption
COLOR= in TEXTATTRS= is ignored, and the text is assigned
a color from the color ramp.

TEXTATTRS= and The result depends on the bounding-box display as follows:

FILLATTRS=
994 Chapter 6 / Plot Statements

Attribute Option Result

n If the fill and outline are displayed, the text and fill colors are
set to their COLOR= specification. The bounding-box outline
is assigned a color from the color ramp.
n If only the fill is displayed, the text color is set to its COLOR=
specification. Suboption COLOR= in FILLATTRS= is ignored,
and the bounding-box fill is assigned a color from the color
ramp.

TEXTATTRS= and The result depends on the bounding-box display as follows:

OUTLINEATTRS= n If the fill and outline are displayed, the text and outline
colors are set to their COLOR= specification. The bounding-
box fill is assigned a color from the color ramp.
n If only the outline is displayed, the outline color is set to its
COLOR= specification. Suboption COLOR= in TEXTATTRS=
is ignored, and the text is assigned a color from the color
ramp.

TEXTATTRS=, When the bounding-box fill and outline are displayed, suboption
FILLATTRS=, and COLOR= in TEXTATTRS= is ignored, and the text is assigned a
OUTLINEATTRS= color from the color ramp. The fill and outline colors are set to
their COLOR= specification.

Customizing Text Marker Placement (Experimental)

Starting with SAS 9.4M3, you can use the TEXTPLOT statement OUTFILE= and
OUTID= options to write information about the bounding box size and position for
each text marker to a comma-separated values (CSV) file.

CAUTION
OUTFILE= and OUTID= are experimental options that are available in SAS
9.4M3. Do not use these options in production jobs.

These options are for users who want to customize the placement of the text in a
text plot. After you write the attributes to a CSV file, you can import the data into a
SAS data set, and then use the data to customize the placement of the text
markers.

The following table lists the information that is written to the CSV file.

Column Name Data Type Description

DataDescent Numeric The descent, in y-axis units, of this observation's

text string. For a discrete axis, one unit is the
distance between adjacent discrete category ticks.
 TEXTPLOT Statement 995

Column Name Data Type Description

DataHeight Numeric The height, in y-axis units, of this observation's

text string. For a discrete axis, one unit is the
distance between adjacent discrete category ticks.

DataWidth Numeric The width, in x-axis units, of this observation's

text string. For a discrete axis, one unit is the
distance between adjacent discrete category ticks.

Descent Numeric The descent, in pixels, of this observation's text

string.

Height Numeric The height, in pixels, of the bounding box for this
observation's text string.

Name Character The name that is specified in the NAME= option in

this TEXTPLOT statement. The name associates
this CSV file with a specific TEXTPLOT statement.
This column is useful when multiple CSV files
from different TEXTPLOT statements are merged
into a single SAS data set.

OutID Character The value for this observation from the column
specified in the OUTID= option in this TEXTPLOT
statement. The ID is used as a unique identifier for
this observation's text string.

RelativeDescent Numeric The relative descent of this observation's

bounding box as a proportion of the plot area
width.

RelativeHeight Numeric The relative height of this observation's text string

as a proportion of the plot area height.

RelativeWidth Numeric The relative width of this observation's bounding

box as a proportion of the plot area width.

Width Numeric The width, in pixels, of the bounding box for this
observation's text string.

Here is an example of using the OUTFILE= and OUTID= options in a TEXTPLOT

statement to write text bounding-box information to CSV file textboxdata.csv. It
also shows how to import the CSV file into SAS data set Work.TextBoxData.
filename csvout "textboxdata.csv";

/* Write the text bound-box data to file textboxdata.csv */

proc template;
define statgraph textplot;
begingraph;
996 Chapter 6 / Plot Statements

layout overlay;
textplot x=weight y=height text=name / name="textboxdata"
outfile=csvout outid=name;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=textplot;

run;

/* Import the CSV file into data set Work.TextBoxData */

proc import
out=work.textboxdata
datafile=csvout
dbms=csv replace;
getnames=yes;
run;

Here is a partial listing of data set Work.TextBoxData.

The SAS System

Obs NAME OUTID WIDTH HEIGHT DESCENT RELATIVEWIDTH

1 textboxdata Alfred 34.045307 19.028384 5.004367 0.060364

2 textboxdata Alice 31.040453 19.028384 5.004367 0.055036
3 textboxdata Barbara 43.059872 19.028384 5.004367 0.076347
…
Obs RELATIVEHEIGHT RELATIVEDESCENT DATAWIDTH DATAHEIGHT DATADESCENT

1 0.046411 0.012206 6.489479 1.062499 0.279432

2 0.046411 0.012206 5.916715 1.062499 0.279432
3 0.046411 0.012206 8.207772 1.062499 0.279432
…

After you import the CSV file into a SAS data set, you can then use the data to
position the text values as needed.

Example: TEXTPLOT Statement

This example creates a text plot of weight by age and sex. Column Name provides
the text for filled, outlined markers. The following figure shows the output.
 VECTORPLOT Statement 997

Here is the SAS code for this example.

proc template;
define statgraph textplot;
begingraph;
entrytitle "Weight by Age and Sex";
layout overlay / yaxisopts=(offsetmin=0.05 offsetmax=0.05);
textplot x=age y=weight text=name / name='textplot1'
display=all
textattrs=(weight=bold) fillattrs=(transparency=0.9)
group=sex groupdisplay=cluster clusterwidth=1;
discretelegend 'textplot1';
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=textplot;

run;

VECTORPLOT Statement
Creates a plot of vectors (directed line segments).

Syntax
VECTORPLOT X=numeric-column | expression
Y=numeric-column | expression
XORIGIN=numeric-constant | numeric-column | expression
998 Chapter 6 / Plot Statements

YORIGIN=numeric-constant | numeric-column | expression < /options >;

Summary of Optional Arguments

Appearance options
CLIP=TRUE | FALSE
specifies whether the origin is considered when determining the data
ranges for the axes.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the column or range attribute variable to use to map the line
colors to a continuous color gradient.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the vector plot lines.
DATATRANSPARENCY=number
specifies the degree of the transparency of the vector line and
arrowhead, and the vector labels.
INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to
one of the GraphData1–GraphDataN style elements.
LINEATTRS=style-element | style-element (line-options ) | (line-options )
specifies the attributes of the vector line and arrowhead.
LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line
thickness.
LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.
LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to
map a line thickness to each group value.

Axes options
PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.
 VECTORPLOT Statement 999

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over a
vector line.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
DATALABEL=column | expression
specifies the labels at the ends of the vectors.
DATALABELATTRS=style-element | style-element (text-options ) | (text-
options )
specifies the color and font attributes of the data labels.
DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |
RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end points and
arrow heads.
DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at specified split characters.
DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if
needed.
DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.
DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT
specifies the justification of the strings that are inside the data label
blocks.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Midpoint options
GROUP=column | discrete-attr-var | expression
creates a distinct set of vector lines and data label colors for each unique
group value in the specified column.
INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the
plot.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Vector options
1000 Chapter 6 / Plot Statements

ARROWDIRECTION=OUT | IN | BOTH
specifies the placement of the arrowhead(s) at the end of the vector.
ARROWHEADS=TRUE | FALSE
specifies whether arrowheads are displayed on the vectors.
ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED
specifies the shape of the arrowheads.
SCALE=number
specifies the scale factor of the vector length.

Required Arguments
X=numeric-column | expression
specifies the column for the X values of the vector endpoints.

Y=numeric-column | expression
specifies the column for the Y values of the vector endpoints.

XORIGIN=numeric-constant | numeric-column | expression

specifies the X data coordinate of the vector origin.

YORIGIN=numeric-constant | numeric-column | expression

specifies the Y data coordinate of the vector origin.

Optional Arguments
ARROWDIRECTION=OUT | IN | BOTH
specifies the placement of the arrowhead(s) at the end of the vector.

OUT
specifies a single arrowhead, pointing away from the origin, at the end of the
vector away from the origin.

IN
specifies a single arrowhead, pointing toward the origin, at the end of the
vector near the origin.

BOTH
specifies two arrowheads.

One arrowhead points away from the origin, at the end of the vector
opposite from the origin. The other arrowhead points toward the origin, at
the end of the vector near the origin.

Default OUT

Interaction If ARROWHEADS= FALSE, then this option is ignored.

Tip Use the ARROWHEADSHAPE= option to control arrowhead

appearance.

ARROWHEADS=TRUE | FALSE
specifies whether arrowheads are displayed on the vectors.

Default TRUE
 VECTORPLOT Statement 1001

Interaction When this option is set to FALSE, the ARROWDIRECTION= and

ARROWHEADSHAPE= options are ignored and all vectors are
displayed as undirected line segments.

See “boolean ” on page xii for other Boolean values that you can use.

ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED

specifies the shape of the arrowheads.

Default OPEN

Interaction This option is ignored if ARROWHEADS= FALSE.

Note No arrowhead is drawn for a zero-length vector. A zero-length

vector is represented as a dot at its starting point.

Tip Use the ARROWDIRECTION= option to control arrow direction.

CLIP=TRUE | FALSE
specifies whether the origin is considered when determining the data ranges for
the axes.

FALSE
includes the origin when establishing the axis scales.

Each axis might be extended to force the display of the origin.

TRUE
ignores the origin when establishing axis scales.

Each axis scale is determined by the other plots in the overlay. This might
result in the origin not being displayed if its data range is not within the data
ranges of tips of the vectors.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.
1002 Chapter 6 / Plot Statements

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorAltRamp style element.

Interaction For this option to take effect, the COLORRESPONSE= option must
also be specified.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the column or range attribute variable to use to map the line colors to
a continuous color gradient.

Note: This feature applies to SAS 9.4M3 and to later releases.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Requirement For a grouped plot, the COLORRESPONSE values should remain

constant for each group value. If the COLORRESPONSE column
has multiple values for a single GROUP value, unexpected results
might occur.

Interactions When the GROUP= option is specified with the

COLORRESPONSE= option, the color attributes are controlled by
the COLORRESPONSE= option.
 VECTORPLOT Statement 1003

Suboption COLOR= in the DATALABELATTRS= option overrides

this option for the data label color attribute.

This option overrides suboption COLOR= in the LINEATTRS=

option and varies the line color according to the color gradient or
the attribute map.

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

For a numeric column or expression, the ThreeColorAltRamp style

element defines the line color gradient.

DATALABEL=column | expression
specifies the labels at the ends of the vectors.

Default No data labels are displayed

Note The label positions are automatically adjusted to prevent the labels
from colliding with other labels and other arrows.

DATALABELATTRS=style-element | style-element (text-options ) | (text-options )

specifies the color and font attributes of the data labels.

Defaults For non-grouped data, the GraphDataText style element.

For grouped data, the GraphData1:ContrastColor–

GraphDataN:ContrastColor style references.

Interaction For this option to take effect, the DATALABEL= option must also be
specified.

Note When the DATALABELPOSITION=AUTO option is in effect, in some

cases, the data label font size might be reduced in order to avoid
overlapping labels and markers.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

DATALABELPOSITION=AUTO | TOPRIGHT | TOP | TOPLEFT | LEFT | CENTER |

RIGHT | BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies the location of the data labels relative to the end points and arrow
heads.

AUTO
uses a data-label collision avoidance algorithm to position the labels.

Note When AUTO is in effect, in some cases, the data label font size might
be reduced in order to avoid overlapping labels and markers.

See “LABELPLACEMENT=AUTO | GREEDY | SA” on page 36

1004 Chapter 6 / Plot Statements

TOPRIGHT
TOP
TOPLEFT
LEFT
CENTER
RIGHT
BOTTOMLEFT
BOTTOM
BOTTOMRIGHT
specifies a location for all of the data labels.

Default AUTO

DATALABELSPLIT=TRUE | FALSE
specifies whether to split the data labels at specified split characters.

Default FALSE. The data labels are not split.

Requirement The DATALABEL= option must also be specified.

Interactions The DATALABELSPLITCHAR= option specifies one or more

characters on which splits can occur.

This option has no effect when DATALABELPOSITION=AUTO.

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITCHAR="character-list"
specifies one or more characters on which the data labels can be split if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the data label. In that case, all of the specified split characters
together are treated as a single split character.

When DATALABEL= is specified and DATALABELSPLIT=TRUE, the data label is

split unconditionally at each occurrence of any of the specified split characters.
If the data label does not contain any of the specified characters, then the label
is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
datalabelsplitchar="abc"

The DATALABEL= option and the DATALABELSPLIT=TRUE

option must also be specified.
 VECTORPLOT Statement 1005

Interactions This option has no effect if DATALABELPOSITION=AUTO.

The DATALABELSPLITCHARDROP= option specifies whether the

split characters are included in the data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the DATALABELSPLITJUSTIFY= option to specify the

justification of the strings in the data label block.

DATALABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the data labels.

TRUE
drops the split characters from the data label.

FALSE
includes the split characters in the data label.

When DATALABELSPLIT=TRUE and DATALABELSPLITCHARDROP=FALSE,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

The following figure shows an example of a data label with the following
specifications:
n the data label text for this label is Product*Group*A

n DATALABELSPLIT=TRUE

n DATALABELSPLITCHARDROP=TRUE | FALSE

n DATALABELSPLITCHAR="*"

When DATALABELSPLITCHARDROP=TRUE, the asterisks are removed from

the label. When DATALABELSPLITCHARDROP=FALSE, each asterisk remains
as the last character in the line prior to the new line.

Default TRUE. The split characters are dropped from the data label.

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction The DATALABELSPLITCHAR= option specifies the split

characters.
1006 Chapter 6 / Plot Statements

See “boolean ” on page xii for other Boolean values that you can use.

DATALABELSPLITJUSTIFY=AUTO | CENTER | LEFT | RIGHT

specifies the justification of the strings that are inside the data label blocks.

AUTO
justifies the labels based on the DATALABELPOSITION= option.

DATALABELPOSITION= Value Justification

TOPLEFT, LEFT, or BOTTOMLEFT RIGHT

TOPRIGHT, RIGHT, or BOTTOMRIGHT LEFT

TOP, CENTER, or BOTTOM CENTER

CENTER
LEFT
RIGHT
justifies the labels center, left, or right, as specified.

The following figure shows an example in which DATALABELPOSITION=TOP.

Note: The gray vertical line at the bottom of each label represents the
horizontal center of the text box for reference.

In this case, because DATALABELPOSITION=TOP, AUTO centers the lines of

text. The text box is anchored the same way that the unsplit text is anchored.
For example, if DATALABELPOSITION=TOP, then the bottom center of the text
box is positioned at the top of the marker.

Default AUTO

Requirement The DATALABEL= option and the DATALABELSPLIT=TRUE option

must also be specified.

Interaction This option has no effect if DATALABELPOSITION=AUTO.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the vector plot lines. The following figure
shows a vector plot with each of the skins applied.
 VECTORPLOT Statement 1007

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If that value is not specified, then the
GraphSkins:DataSkin style element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned graphical
elements is limited to 200 per plot in an OVERLAY or PROTOTYPE
layout. When this limit is exceeded for a plot, the specified data skin
is not applied to that plot. In that case, use the DATASKINMAX=
option in your ODS GRAPHICS statement to increase the maximum
limit.

Interaction This option overrides the BEGINGRAPH statement DATASKIN=

option.

DATATRANSPARENCY=number
specifies the degree of the transparency of the vector line and arrowhead, and
the vector labels.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Restriction The specified transparency is not applied to missing values. Missing

values are always opaque.

GROUP=column | discrete-attr-var | expression

creates a distinct set of vector lines and data label colors for each unique group
value in the specified column.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.
1008 Chapter 6 / Plot Statements

Default Each distinct group value might be represented in the plot by a

different combination of color and line pattern. Lines vary according
to the ContrastColor and LineStyle attributes of the GraphData1–
GraphDataN and GraphMissing style elements.

Interactions The group values are mapped in the order of the data, unless the
INDEX= option is used to alter the default sequence of line patterns
and colors.

The INCLUDEMISSINGGROUP= option controls whether missing

group values are considered a distinct group value.

When both the GROUP= and the COLORRESPONSE= options are

specified, the color attributes are controlled by the
COLORRESPONSE= option.

Note If you specify a column in a SAS data set, the visual attributes for
each group value are assigned in data order. If you specify a column
in a CAS in-memory table, the visual attributes for each group value
are assigned in ascending order of the group column character
values or of unformatted numeric values.

Tips You can use the LINEATTRS= option to override the

representations that are used to identify the groups. For example,
You can use LINEATTRS=(PATTERN=SOLID) to assign the same
pattern to all of the lines, letting the line color distinguish group
values. Likewise, you can use LINEATTRS=(COLOR=BLACK) to
assign the same color to all of the lines, letting the line pattern
distinguish group values.

By default, ODS Graphics limits the number of groups to 1000.

When this limit is exceeded, the GROUP= option is ignored. Use the
GROUPMAX= option in the ODS GRAPHICS statement to change
the maximum number of groups that can be processed.

See “DISCRETEATTRVAR Statement” on page 1516

INCLUDEMISSINGGROUP=TRUE | FALSE
specifies whether missing values of the group variable are included in the plot.

Default TRUE

Interaction For this option to take effect, the GROUP= option must also be
specified.

Tip The attributes of the missing group value are determined by the
GraphMissing style element unless a discrete attribute map is in
effect, the INDEX= option is used, the MISSING= system option
changes the default missing character, or a user-defined format is
applied to the group value. In those cases, the attributes of the
missing group value are determined by a GraphData1–GraphDataN
style element instead of by the GraphMissing style element.

See “boolean ” on page xii for other Boolean values that you can use.
 VECTORPLOT Statement 1009

INDEX=positive-integer-column | expression
specifies indices for mapping line attributes (color and line pattern) to one of
the GraphData1–GraphDataN style elements.

Requirements The column or expression value must be an integer value of 1 or

greater. Otherwise, this option is ignored.

The positive-integer column must not contain missing values.

Otherwise, the entire column is invalidated and this option is
ignored.

All of the indexes for a specific group value must be the same.
Otherwise, the results are unpredictable.

Interaction For this option to take effect, the GROUP= option must also be
specified.

Notes The index values are 1-based indices. For the style attributes in
GraphData1–GraphDataN, if the index value is greater than N,
then a modulo operation remaps that index value to a number
less than N to determine which style to use.

If you do not use this option, then the group values are mapped in
the order of the data.

Tip You can use indexing to collapse the number of groups that are
represented in a graph. For more information, see “Remapping
Groups for Grouped Data” on page 211.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The string specified on the NAME= option.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the GROUP= option is specified, then this option is ignored.

LINEATTRS=style-element | style-element (line-options ) | (line-options )

specifies the attributes of the vector line and arrowhead.

Defaults For non-grouped data, the GraphDataDefault style element.

For grouped data, the ContrastColor, LineStyle, and LineThickness

attributes of the GraphData1–GraphDataN style elements.

Interactions The COLORRESPONSE= option overrides this option’s COLOR=

suboption.

The LINETHICKNESSRESPONSE= option overrides this option’s

THICKNESS= suboption.
1010 Chapter 6 / Plot Statements

Note The arrow head size is nonlinearly proportional to the line thickness
in order to maintain appropriately sized arrow heads for thicker
lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

“Line Options” on page 1602 for available line-options.

LINETHICKNESSMAX=dimension
specifies the maximum line thickness when a response variable is used to
determine the line thickness. By default, this option determines the thickness of
the line that represents the maximum response column value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default Ten times the thickness that is specified by the GraphDataDefault

style element LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The LINETHICKNESSMAXRESPONSE= option specifies the

response value at which this maximum line thickness is reached.
The line thickness for response values that exceed the
LINETHICKNESSMAXRESPONSE= value are set to the value that
is specified by this option.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMIN= option to specify the minimum line

thickness.

See “dimension” on page xiii

LINETHICKNESSMAXRESPONSE=numeric | scalar-numeric-expression
specifies the response value that corresponds to the maximum line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The maximum value in the response column that is specified in the
LINETHICKNESSRESPONSE= option.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

The thickness for all lines that exceed the maximum response value
is set to the value specified in the LINETHICKNESSMAX= option.
 VECTORPLOT Statement 1011

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

LINETHICKNESSMIN=dimension
specifies the minimum line thickness when a response variable is used to
determine the line thickness.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The thickness specified by the GraphDataDefault style element

LineThickness attribute.

Interactions The LINETHICKNESSRESPONSE= option must be specified for this

option to have any effect.

If the line thickness that is calculated from the

LINETHICKNESSMIN=, LINETHICKNESSMAX=, and
LINETHICKNESSMAXRESPONSE= option values is less than 0.5
for a line, that line is not drawn.

Tip Use the LINETHICKNESSMAX= option to specify the maximum line

thickness.

See “dimension” on page xiii

LINETHICKNESSRESPONSE=numeric-column | expression
specifies a response column or range attribute variable that is used to map a
line thickness to each group value.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default The GraphDataDefault style element LineThickness attribute.

Interactions When the column values are all zero, all negative, or all missing,
this option is ignored. In these cases, the default line thickness is
used for all of the lines.

This option overrides suboption THICKNESS= in the LINEATTRS=

option.

Note The line thicknesses are not reflected in a discrete legend.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.
1012 Chapter 6 / Plot Statements

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

PRIMARY=TRUE | FALSE
specifies that the data columns for this plot and the plot type be used for
determining default axis features. This option is needed only when two or more
plots within an overlay-type layout contribute to a common axis.

Default FALSE

Restriction This option is ignored if the plot is placed under a GRIDDED or

LATTICE layout block.

Note In an OVERLAY layout, only one plot in an overlay can be the

primary plot on a per-axis basis. When no plot is designated as the
primary plot, the first plot that can be a primary plot is considered
the primary plot. If multiple plots specify PRIMARY=TRUE for the
same axis, then the last such plot encountered is considered the
primary plot.

See “When Plots Share Data and a Common Axis” on page 1046

“boolean ” on page xii for other Boolean values that you can use.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles X, Y, DATALABEL, XORIGIN, YORIGIN,
GROUP, and INDEX.

SCALE=number
specifies the scale factor of the vector length.

Default 1.0

Restriction The number specified must be greater than 0.

TIP=(role-list ) | NONE
specifies the information to display when the cursor is positioned over a vector
line. If this option is used, then the information specified replaces all of the
 VECTORPLOT Statement 1013

information that is displayed by default. You can specify roles for columns that
do not contribute to the vector plot along with roles that do.

(role-list)
an ordered, space-separated list of unique VECTORPLOT roles and user-
defined roles.

VECTORPLOT roles include X, Y, DATALABEL, XORIGIN, YORIGIN, GROUP,

INDEX, and COLORRESPONSE.

Define user-defined roles with the ROLENAME= option.

Note Starting with SAS 9.4M3, the COLORRESPONSE role is valid.

Example The following example displays data tips for the columns assigned
to the roles X, Y, GROUP, and the column Obs, which is not
assigned to any pre-defined VECTORPLOT role. The Obs column
must first be assigned a role.

ROLENAME=(TIP1=OBS)
TIP=(TIP1 X Y GROUP)

NONE
suppresses data tips from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: X, Y, DATALABEL, XORIGIN, YORIGIN,
GROUP, and COLORRESPONSE.

Requirement To generate data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tip You can control the labels and formats for the TIP roles with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example ROLENAME=(TIP1=SALARY)
TIP=(TIP1)
TIPFORMAT=(TIP1=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.
1014 Chapter 6 / Plot Statements

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example ROLENAME=(TIP1=PCT)
TIP=(TIP1)
TIPLABEL=(TIP1="Percent")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles. (See the
ROLENAME= option.)

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
By default in a VECTORPLOT, each vector starts at 0, 0 in the data space and is
terminated with an open arrowhead. Zero-length vectors are represented by a dot
at the starting point. To specify alternative coordinates for the starting point, use
the XORIGIN= and YORIGIN= arguments.
 VECTORPLOT Statement 1015

Example: VECTORPLOT Statement

The following graph was generated by the “Example Program” on page 1015:

Example Program
data CarPref;
input Make $11. Model $13. (S1-S25) (1.);
datalines;
Cadillac Deville 8007990491240508971093809
Chevrolet Aveo 0051200423451043003515698
Chevrolet Cavalier 4053305814161643544747795
Chevrolet Malibu 6027400723121345545668658
Dodge Intrepid 7006000434101107333458708
Dodge Stratus 3005005635461302444675655
Dodge Neon 4005003614021602754476555
Ford Taurus 2024006715021443530648655
Ford Mustang 5007197705021101850657555
Ford Focus 0021000303030201500514078
Honda Accord 5956897609699952998975078
Honda Civic 4836709507488852567765075
Lincoln LS 7008990592230409962091909
Pontiac Firebird 0107895613201206958265907
Volkswagen Jetta 4858696508877795377895000
Volkswagen Beetle 4858509709695795487885000
Volvo S40 9989998909999987989919000
;
run;
* Compute Two Component Model;
 1016 Chapter 6 / Plot Statements

ods graphics;
ods exclude all;
ods output mdprefplot=plotdata;
proc prinqual data=CarPref n=2 replace mdpref method=mgv;
id model;
transform monotone(S1-S25);
run;
ods select all;

proc template;
define statgraph vectorplot;
begingraph;
entrytitle "Multidimensional Preference Analysis";
entrytitle "of Preference Ratings for Automobiles";
layout overlayequated / equatetype=fit cycleattrs=true;
referenceline y=0 / datatransparency=0.7;
referenceline x=0 / datatransparency=0.7;
vectorplot y=vec2 x=vec1 xorigin=0 yorigin=0 /
datalabel=label2var;
scatterplot y=prin2 x=prin1 /
datalabel=idlab1 primary=true
markerattrs=(symbol=circlefilled);
endlayout;
endgraph;
end;
run;

proc sgrender data=plotdata template=vectorplot;

run;

WATERFALLCHART Statement
Creates a waterfall chart that is computed from input data.

Restriction: The WATERFALLCHART statement does not support CAS in-memory tables. In
SAS 9.4M6 and earlier releases, unexpected results might occur when data stored
in a CAS in-memory table is used. Starting with SAS Viya 3.5 and SAS Studio 5.2,
the waterfall chart is dropped when data stored in a CAS in-memory table is used.
Interaction: A Waterfall chart accumulates response values in data order. Any change in the
order of the X-axis values from the data order can adversely affect the waterfall
chart. The X-axis value order can change when the Waterfall chart is overlaid with
other plots or when it is used in a Lattice with uniform axes. It can also change
when certain options are applied to the X axis.
Tip: Starting with SAS 9.4M3, you can use subpixel rendering with this statement. It is
enabled by default. To disable subpixel rendering, specify SUBPIXEL=OFF in the
BEGINGRAPH statement or in an ODS GRAPHICS statement. For information
about the BEGINGRAPH statement SUBPIXEL= option, see SUBPIXEL= on page 39.
For information about the ODS GRAPHICS statement SUBPIXEL= option, see “ODS
GRAPHICS Statement” on page 1561.
 WATERFALLCHART Statement 1017

Syntax
WATERFALLCHART CATEGORY=column | expression
RESPONSE=numeric-column | expression </options>;

Summary of Optional Arguments

Appearance options
BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.
BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.
COLORGROUP=column | discrete-attr-var | expression
specifies a column that is used to discretely color the transaction bars.
COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.
COLORRESPONSE=numeric-column | range-attr-var | expression
specifies the numeric column or range attribute map variable that is used
to determine the transaction-bar colors.
CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.
DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN
enhances the visual appearance of the filled bars.
DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, bar
labels, and trend lines, if displayed.
DISPLAY=(display-options) | STANDARD | ALL
specifies which bar features to display.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the filled transaction bars.
FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.
FINALBARATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the “final” bar, if displayed.
INITIALBARATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the appearance of the “initial” bar, if displayed.
INITIALBARVALUE=number
specifies a value for the initial bar.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the bar outlines.

Axes options
BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline.
FINALBARTICKVALUE="string"
1018 Chapter 6 / Plot Statements

specifies a tick value to use on the category axis when the “final” bar is
displayed
INITIALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “initial” bar is
displayed.
XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to
the secondary X2 (top) axis.
YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Data tip options

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the
data tips.
TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a
bar.
TIPFORMAT=(role-format-list)
specifies display formats for tip columns.
TIPLABEL=(role-label-list)
specifies display labels for tip columns.

Label options
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the
bar.
BARLABELATTRS=style-element | style-element (text-options) | (text-options)
specifies the text properties of the bar label text.
BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels
are displayed.
BARLABELFORMAT=format
specifies the text format used to display the bar label.
LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

ODS options
URL=string-column
specifies an HTML page to display when a bar is selected.

Plot reference options

NAME="string"
assigns a name to this plot statement for reference in other template
statements.

Statistics options
COLORSTAT=SUM | MEAN
 WATERFALLCHART Statement 1019

specifies the statistic to use for computing the response colors.

STAT=SUM | MEAN
specifies the statistic to be computed for the RESPONSE axis.

Trend line options

CONNECT=START | END
determines whether trend lines connect to the adjacent bar’s starting or
ending value.
CONNECTATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the trend lines that connect the bars.
CONNECTDECREASINGATTRS=style-element | style-element (line-options) |
(line-options)
specifies the appearance of trend lines that denote a decreasing value
between bars.
CONNECTINCREASINGATTRS=style-element | style-element (line-options) |
(line-options)
specifies the appearance of trend lines that denote an increasing value
between bars.

Required Arguments
CATEGORY=column | expression
specifies the column or expression for the category values. Duplicated category
values are summarized into a unique value. All values are treated as discrete.

RESPONSE=numeric-column | expression
specifies the numeric column or expression for the response values.

Optional Arguments
BARLABEL=TRUE | FALSE
specifies whether the bar statistic value is displayed at the end of the bar.

Default FALSE

Note By default, the bar-label format is derived from the format that is
assigned to the response column or from BEST6 if no format is
assigned. The derived format retains the precision of the response-
column format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Tip The font and color attributes for the label are specified by the
BARLABELATTRS= option. The text format is specified by the
BARLABELFORMAT= option.

See “boolean ” on page xii for other Boolean values that you can use.

BARLABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the text properties of the bar label text.

Default The GraphDataText style element.

1020 Chapter 6 / Plot Statements

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

BARLABELFITPOLICY=AUTO | NONE
specifies a policy for avoiding collisions among the bar labels when labels are
displayed.

AUTO
rotates the bar labels if the labels exceed the midpoint spacing.

If the labels collide horizontally due to thin bars, then AUTO drops all of the
labels. The following figure shows an example.

See the BARWIDTH= option for more information about the bar spacing.

NONE
does not rotate the bar labels.

Labels that are too long overlap.

Default AUTO

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

BARLABELFORMAT=format
specifies the text format used to display the bar label.

Default The bar-label format is derived from the format that is applied to
the response column or from BEST6 if no format is assigned. The
derived format retains the precision of the response-column
format and, if necessary, increases the format width to
accommodate the summarized value on the response axis.

Requirement For this option to take effect, BARLABEL=TRUE must be specified.

 WATERFALLCHART Statement 1021

Note When a bar-label format is specified with this option, the bar
labels are formatted as specified by format. The specified format is
not automatically expanded to accommodate values that are too
wide.

Tip If you want the bar-label format to expand automatically for

summarized values on the response axis, specify the format for the
response column rather than in this option.

BARWIDTH=number
specifies the width of a bar as a ratio of the maximum possible width.

Default 0.85

Range 0–1, where 0 is the minimum width, which is one pixel wide, and 1 is the
maximum possible width

Notes This option is needed only to change the default behavior.

By default, the bar width automatically adjusts based on the number of

bars to be displayed and the wall width.

Tip To remove any inter-bar gap, set BARWIDTH=1.

BASELINEATTRS=style-element | (line-options)
specifies the appearance of the baseline.

Default The ContrastColor, LineThickness, and LineStyle attributes of the

GraphAxisLines style element.

Notes The baseline is always drawn by default.

When style-element is specified, only the style element’s Color,

LineStyle, and LineThickness attributes are used.

Tip To suppress the baseline, set the line thickness to 0:

baselineattrs=(thickness=0)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

BASELINEINTERCEPT=number
specifies the response axis intercept for the baseline. The baseline is always
displayed in the chart, whether for a specified value or for the default value.
When the BASELINEINTERCEPT= option is used, the axis range is adjusted to
include the baseline, and the baseline is placed at the specified value on the
response axis.
1022 Chapter 6 / Plot Statements

Default 0

Interactions The value set by this option affects only the chart’s initial and final
bars. If no initial bar value is specified, then the first transaction bar
is drawn from 0, no matter what is set for the baseline value.

If necessary, the response axis data range is extended to include

the baseline intercept. When a logarithmic response axis is
requested and BASELINEINTERCEPT= specifies 0 or a negative
value, the response axis reverts to a linear axis. To restore the log
axis in that case, set BASELINEINTERCEPT= to a positive value.

Note Label positions are automatically adjusted to prevent the labels

from overlapping.

Tips Control the appearance of the baseline with the BASELINEATTRS=

option.

To suppress the baseline prior to SAS 9.4M5, use the

BASELINEATTRS= option to set the line thickness to 0. Starting
with SAS 9.4M5, specify DISPLAYBASELINE=OFF.

COLORGROUP=column | discrete-attr-var | expression

specifies a column that is used to discretely color the transaction bars.

discrete-attr-var
specifies a discrete attribute map variable.

The discrete attribute map variable is created by a DISCRETEATTRVAR

statement.
 WATERFALLCHART Statement 1023

Restriction A discrete attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set by a
dynamic variable.

Interactions This option is ignored if the COLORRESPONSE= option is also

specified.

If a column or expression is specified, then the unique column

values are found and the transaction bar colors are derived from
the GraphData1–GraphDataN style elements. The Color attribute is
used for the bar fill colors and the ContrastColor attribute is used
for the bar outline colors.

If a variable that is associated with an attribute map is specified,

then the color mapping defined by the associated
DISCRETEATTRMAP statement is used for the transaction bars.

Notes All of the COLORGROUP values for a specific category value must
be the same. Otherwise, the results are unpredictable.

All the transaction bars have only one fill and one outline color,
determined by the ODS style or set by the FILLATTRS= and
OUTLINEATTRS= options.

If you specify a column in a SAS data set, colors are mapped to

group values in data order. If you specify a column in a CAS in-
memory table, colors are mapped to group values in ascending
order of the group column character values or unformatted numeric
values.

Tip To manage the color of the “initial” bar, use the INITIALBARATTRS=
option. To manage the color of the “final” bar, use the
FINALBARATTRS= option.

See “DISCRETEATTRMAP Statement” on page 1505

“DISCRETEATTRVAR Statement” on page 1516

COLORMODEL=color-ramp-style-element | (color-list)
specifies a color ramp to use with the COLORRESPONSE= option.

color-ramp-style-element
specifies the name of a color-ramp style element.

The style element should contain these style attributes:

STARTCOLOR
specifies the color for the smallest data value of the COLORRESPONSE=
column.

NEUTRALCOLOR
specifies the color for the midpoint of the range of the
COLORRESPONSE= column.
1024 Chapter 6 / Plot Statements

ENDCOLOR
specifies the color for the highest data value of the COLORRESPONSE=
column.

(color-list)
specifies a space-separated list of colors to use in the color ramp. You can
use style attribute references such as GraphData3:Color, color names, or
RGB, CMYK, HLS, and HSV (HSB) color codes to specify a color. The list can
contain a mix of style attribute references, color names, and color codes.

Requirement The list of colors must be enclosed in parentheses.

See “color ” on page xii

Default The ThreeColorRamp style element

Interactions For this option to take effect, the COLORRESPONSE= option must
also be specified.

When FILLTYPE=GRADIENT and a color list is specified, the middle

color in the list is treated as the NEUTRAL color.

Tip To manage the color of the initial bar, use the INITIALBARATTRS=
option. To manage the color of the final bar, use the
FINALBARATTRS= option.

COLORRESPONSE=numeric-column | range-attr-var | expression

specifies the numeric column or range attribute map variable that is used to
determine the transaction-bar colors.

range-attr-var
specifies a range attribute map variable that is defined in a RANGEATTRVAR
statement.

“RANGEATTRVAR Statement”

Restriction A range attribute map variable specification must be a direct

reference to the attribute map variable. It cannot be set as a
dynamic variable.

See “RANGEATTRVAR Statement” on page 1531

When a numeric column or expression is specified, the range of column or

expression values are linearly mapped to the color ramp that is specified by the
COLORMODEL= option. When a range attribute map variable is specified, the
colors that are defined in the associated range attribute map are used instead.

Restriction This option affects only the fill colors. When only the bar outlines
are displayed, this option has no effect.

Interactions When the COLORGROUP= option is specified with this option, the
COLORGROUP= option is ignored.

This option overrides suboption COLOR= in the FILLATTRS=

option.
 WATERFALLCHART Statement 1025

Tips To display a legend with this option in effect, use a

CONTINUOUSLEGEND statement.

Use the COLORSTAT= option to specify the statistic to compute for

the COLORRESPONSE= column.

To produce discrete color mapping, the RANGEATTRMAP

statement can define an attribute map that maps a single color to
all values greater than 0, and another color to all values less than 0.

Use the FILLTYPE= option to indicate whether the color mapping is

used to produce solid or gradient fills. When FILLTYPE=GRADIENT
is in effect, the color at the end of the bar is based on the color
mapping, and the neutral color of the color ramp is used as the
starting color of each bar.

COLORSTAT=SUM | MEAN
specifies the statistic to use for computing the response colors.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default SUM

Interaction This option is ignored when the COLORRESPONSE= option is not

specified.

Tip This option might affect existing SAS programs. For programs
written before SAS 9.4M3, if STAT= and COLORRESPONSE= are
specified in a WATERFALLCHART statement, the chart colors and
color statistic might change from the previous SAS releases. To
restore the original colors and color statistic in that case, set
COLORSTAT= in the WATERFALLCHART statement to the same
statistic that is specified in STAT=.

CONNECT=START | END
determines whether trend lines connect to the adjacent bar’s starting or ending
value.

START
draws the trend lines horizontally and connects each to the adjacent bar’s
starting value.

Each connecting line extends from the right corner of one bar’s ending value
to the left corner of the adjacent bar’s starting value.

END
draws the trend lines diagonally and connects each to the adjacent bar’s
ending value.

Each connecting line extends from the right corner of one bar’s ending value
to the left corner of the adjacent bar’s ending value.

Default START
1026 Chapter 6 / Plot Statements

Restriction The last connect line is always drawn horizontally, extending from
the right corner of the last data bar’s ending value to the left corner
of the “final” bar’s starting value.

CONNECTATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the trend lines that connect the bars.

Default The GraphConnectLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CONNECTBREAK=TRUE | FALSE
specifies whether the connect line is broken for values that have no
observations.

Note: This feature applies to SAS 9.4M5 and to later releases.

Default FALSE

Requirement DISPLAY= must include CONNECT for this option to have any
effect.

Interaction This option is ignored when the GROUP= option is in effect.

CONNECTDECREASINGATTRS=style-element | style-element (line-options) |

(line-options)
specifies the appearance of trend lines that denote a decreasing value between
bars.

Default The appearance specified by the CONNECTATTRS= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

CONNECTINCREASINGATTRS=style-element | style-element (line-options) |

(line-options)
specifies the appearance of trend lines that denote an increasing value between
bars.

Default The appearance specified by the CONNECTATTRS= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

DATASKIN=NONE | CRISP | GLOSS | MATTE | PRESSED | SHEEN

enhances the visual appearance of the filled bars. The following figure shows
bars with each of the skins applied.
 WATERFALLCHART Statement 1027

Default The DATASKIN= option value that is specified in the BEGINGRAPH

statement. If not specified, then the GraphSkins:DataSkin style
element value is used.

Restriction Starting with SAS 9.4M1, the maximum number of skinned

graphical elements is limited to 200 per plot in an OVERLAY or
PROTOTYPE layout. When this limit is exceeded for a plot, the
specified data skin is not applied to that plot. In that case, use the
DATASKINMAX= option in your ODS GRAPHICS statement to
increase the maximum limit.

Requirement For this option to have any effect, the fill must be enabled by the
ODS style or the DISPLAY= option.

Interactions This option overrides the BEGINGRAPH statement DATASKIN=

option.

The skin appearance is based on the FILLATTRS= color.

When a data skin is applied, all bar outlines are set by the skin, and
the OUTLINEATTRS= option is ignored.

DATATRANSPARENCY=number
specifies the degree of the transparency of the bar fill, bar outline, bar labels,
and trend lines, if displayed.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
filled bar area. The INITIALBARATTRS= and FINALBARATTRS=
options can be used to specify transparency for the initial and final
bars. You can combine this option with FILLATTRS=,
INITIALBARATTRS=, and FINALBARATTRS= to set one transparency
for the bar outlines and trend lines but a different transparency for the
bar fills. Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
1028 Chapter 6 / Plot Statements

DISPLAY=(display-options) | STANDARD | ALL

specifies which bar features to display.

(display-options)
a space-separated list of one or more display options enclosed in
parentheses.

The list can include one or more of the following options:

CONNECT
displays line segments (trend lines) connecting adjacent bar. The
connection point is determined by the CONNECT= option.

FILL
displays filled bars.

FINALBAR
displays the final bar.

OUTLINE
displays outlined bars.

STANDARD
specifies CONNECT, FILL, FINALBAR, and OUTLINE

ALL
currently the same as STANDARD

Default STANDARD

Tips To control the appearance of the bars, use the COLORMODEL= ,

FILLATTRS= , and OUTLINEATTRS= options.

To control the appearance of the trend lines, use the

CONNECTATTRS= , CONNECTDECREASINGATTRS= , and
CONNECTINCREASINGATTRS= options.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the filled transaction bars.

Defaults If the COLORGROUP= option is not specified, then the

GraphDataDefault:Color style reference.

If the COLORGROUP= option is specified, then the

GraphData1:Color–GraphDataN:Color style references.

Interaction This option’s color specification is ignored if either the

COLORMODEL= or COLORRESPONSE= option is specified. The
transparency specification is honored in that case.

Tip The DATATRANSPARENCY= option sets the transparency for the

bar fill, bar outline, and trend lines. You can combine this option with
DATATRANSPARENCY= to set one transparency for the bar outlines
and trend lines but a different transparency for the bar fills.
Example:
datatransparency=0.2 fillattrs=(transparency=0.6)
 WATERFALLCHART Statement 1029

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FILLTYPE=SOLID | GRADIENT
specifies the bar fill type.

Note: This feature applies to SAS 9.4M2 and to later releases.

SOLID
fills each bar with the color that is assigned to that bar.

GRADIENT
fills each bar with a color gradient that starts with the neutral color from the
color model or with the middle color from the color model color list, if
specified, and ends with the color for that bar that is determined by the
COLORGROUP= or COLORRESPONSE= option.

Interaction The SHEEN data skin cannot be used when

FILLTYPE=GRADIENT is in effect. You can use one of the other
data skins.

Default SOLID

Interaction The DISPLAY= option must include FILL for this option to have any
effect.

Tip The colors that are used depend on whether the COLORGROUP=
option or the COLORRESPONSE= option is also specified.

See DATASKIN= on page 1026

FINALBARATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the “final” bar, if displayed.

Default the GraphFinal style element

Interaction This option is ignored if the DISPLAY= option does not display the
“final” bar.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

FINALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “final” bar is
displayed

Default "Final"

Interaction This option is ignored if the DISPLAY= option does not display the
“final” bar.
1030 Chapter 6 / Plot Statements

INITIALBARATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the “initial” bar, if displayed.

Default the GraphInitial style element

Interaction For this option to take effect, the INITIALBARVALUE= option must
also be specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

INITIALBARTICKVALUE="string"
specifies a tick value to use on the category axis when the “initial” bar is
displayed.

Default "Initial"

Interaction For this option to take effect, the INITIALBARVALUE= option must
also be specified.

INITIALBARVALUE=number
specifies a value for the initial bar. The initial bar’s value is used as the starting
response value for the first transaction bar.

Interactions If this option is not specified, then the initial bar is not included in
the chart and the first transaction bar is drawn from response value
0. This is true even if an intercept value is set by the
BASELINEINTERCEPT= option.

If necessary, the response axis data range is extended to include

the initial bar value. When a logarithmic response axis is requested
 WATERFALLCHART Statement 1031

and INITIALBARVALUE= specifies 0 or a negative value, the

response axis reverts to a linear axis. To restore the log axis in that
case, set INITIALBARVALUE= to a positive value.

Note The first transaction bar starts at response value 0.

LEGENDLABEL="string"
specifies a label to be used in a discrete legend for this plot.

Default The response-variable label. If a label is not defined, then the

response-variable name is used.

Restriction This option applies only to an associated DISCRETELEGEND

statement.

Interaction If the COLORGROUP= option is in effect, then this option is ignored.

NAME="string"
assigns a name to this plot statement for reference in other template
statements. The specified name is used primarily in legend statements to
coordinate the use of colors and line patterns between the plot and the legend.

Restrictions The string is case sensitive, cannot contain spaces, and must define
a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interaction The string is used as the default legend label if the LEGENDLABEL=
option is not used.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the bar outlines.

Default The ContrastColor and LineThickness attributes of the

GraphOutlines style element.

Interactions For this option to have any effect, outlines must be enabled by the
ODS style or the DISPLAY= option.

If the DATASKIN= option applies a data skin, then this option is

ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ROLENAME=(role-name-list)
specifies user-defined roles that can be used to display information in the data
tips. This option provides a way to add to the data columns that appear in data
tips that are specified by the TIP= option.

(role-name-list)
a space-separated list of role-name = column pairs.
1032 Chapter 6 / Plot Statements

Example The following example assigns the column Obs to the user-defined
role TIP:
ROLENAME=(TIP1=OBS)

Default No user-defined roles

Requirement The role names that you choose must be unique and different from
the predefined roles CATEGORY, RESPONSE, COLORGROUP, and
COLORRESPONSE.

STAT=SUM | MEAN
specifies the statistic to be computed for the RESPONSE axis.

Default SUM

Tip If you use this option with COLORRESPONSE= in SAS programs that
were written before SAS 9.4M3, the chart colors and color statistic
might change from the previous SAS releases. To restore the original
colors and color statistic in that case, set COLORSTAT= in the
WATERFALLCHART statement to the same statistic that is specified in
STAT=.

TIP=(role-list) | NONE
specifies the information to display when the cursor is positioned over a bar. If
this option is used, then it replaces all of the information that is displayed by
default. Roles for columns that do not contribute to the waterfall chart can be
specified along with roles that do.

(role-list)
an ordered, space-separated list of unique WATERFALLCHART and user-
defined roles.

WATERFALLCHART roles include CATEGORY , RESPONSE ,

COLORGROUP , and COLORRESPONSE .

Example The following example displays data tips only for the column that
is assigned to the RESPONSE role:

TIP=(RESPONSE)

NONE
suppresses data tips and URLs (if requested) from the plot.

Default The columns assigned to these roles are automatically included in

the data tip information: CATEGORY and RESPONSE .

Requirement To enable data tips in the output, you must include an ODS
GRAPHICS ON statement that specifies the IMAGEMAP option,
and you must write the output to the ODS HTML destination.

Interaction This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.
 WATERFALLCHART Statement 1033

Tip The labels and formats for the TIP roles can be controlled with the
TIPLABEL= and TIPFORMAT= options.

TIPFORMAT=(role-format-list)
specifies display formats for tip columns. This option provides a way to control
the formats of columns that appear in data tips.

(role-format-list)
a space-separated list of role-name = format pairs.

Example TIP=(RESPONSE)
TIPFORMAT=(RESPONSE=DOLLAR12.)

Default The column format of the column assigned to the role or BEST6 if
no format is assigned to a numeric column.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

TIPLABEL=(role-label-list)
specifies display labels for tip columns. This option provides a way to control
the labels of columns that appear in data tips.

role-label-list
a space-separated list of rolename ="string" pairs.

Example TIP=(RESPONSE)
TIPLABEL=(RESPONSE="Average Sales")

Default The column label or column name of the column assigned to the
role.

Restriction Only the roles that appear in the TIP= option are used.

Requirement A column must be assigned to each of the specified roles.

URL=string-column
specifies an HTML page to display when a bar is selected.

string-column
specifies a column that contains a valid HTML page reference (HREF) for
each bar that is to have an active link.

Example http://www.sas.com

Requirements To generate selectable bars, you must include an ODS GRAPHICS

ON statement that specifies the IMAGEMAP option, and you
must write the output to the ODS HTML destination.

For non-grouped data, the values of the column are expected to

be same for each unique RESPONSE value. If they are not, then
the results might be unpredictable.
1034 Chapter 6 / Plot Statements

For grouped data, the values of the column are expected to be the
same for each unique RESPONSE and group combination.

Interactions This option has no effect when TIP=NONE.

This option is ignored when the plot statement is in an OVERLAY

or PROTOTYPE layout and the INCLUDERANGES= option is
specified in the LINEAROPTS= or TIMEOPTS= option for either
axis.

Tips The URL value can be blank for some RESPONSE values, meaning
that no action is taken when the bars for those RESPONSE values
are selected.

The URL value can be the same for different RESPONSE values,
meaning that the same action is taken when the bars for those
RESPONSE values are selected.

By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET=
on page 1568.

XAXIS=X | X2
specifies whether data are mapped to the primary X (bottom) axis or to the
secondary X2 (top) axis.

Default X

Interaction The overall plot specification and the layout type determine the axis
display. For more information, see “How Axis Features Are
Determined” on page 1040.

YAXIS=Y | Y2
specifies whether data are mapped to the primary Y (left) axis or to the
secondary Y2 (right) axis.

Default Y

Interactions This option is ignored if the RESPONSE= argument is not specified.

The overall plot specification and the layout type determine the
axis display. For more information, see “How Axis Features Are
Determined” on page 1040.

Details
The input data for the WATERFALLCHART statement is raw, unsummarized input
data, and the statement calculates appropriate summarization statistics (sum or
mean). By default, the bars in the chart appear in the order in which the CATEGORY
 WATERFALLCHART Statement 1035

values are present in the input data. A waterfall chart is typically used to show
credit and debit transactions or successive changes to a given state.

In a waterfall chart, the bars that are calculated from the data are called
“transaction” bars. The transaction bars represent the values of the RESPONSE
variable across a series of intermediate values for the specified CATEGORY
variable. You can manage the color of the transaction bars using the
COLORGROUP , COLORMODEL , or COLORRESPONSE= option.

A waterfall chart can also display an “initial” bar and a “final” bar. The value of the
initial bar determines the starting response value for the first transaction bar. To
set the initial value, use the INITIALBARVALUE= option. If the initial bar is not
displayed, then the first transaction bar has a starting response value of 0. The
value of the final bar is set automatically to the ending value of the last transaction
bar.

Example: WATERFALLCHART Statement

The following graph was generated by the “Example Program” on page 1035:

Example Program
data transactions;
input ID $ Amount type $;
datalines;
Alpha 2000 credit
Beta -2500 debit
1036 Chapter 6 / Plot Statements

Gamma -2000 debit

Delta -500 debit
Epsilon 2250 credit
;
proc template;
define statgraph waterfallchart;
begingraph;
layout overlay;
waterfallchart category=id response=amount /
colorgroup=type
initialbarvalue=1000
name="waterfall";
discretelegend "waterfall";
endlayout;
endgraph;
end;
run;

proc sgrender data=transactions template=waterfallchart;

run;
 1037

PART 5

Plot Axes

Chapter 7
Axis Features in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039

Chapter 8
Axis Options in Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055
1038
 1039

7
Axis Features in Layouts

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039
How Axis Features Are Determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1040
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1040
Plot Data Are Mapped to a Designated Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1040
When Plots Share Data and a Common Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1046
Plot Axis Types Must Agree on Common Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049
Controlling Axis Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049
Setting the Axis Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1050
Adjusting the Axis View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1051
Adjusting Axis Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1052
Adjusting Axis Offsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1053

Overview
GTL plots are specified within layout blocks that enable you to control the graph
display features, including the display of the axes for the plots within the layout.
For example, the LAYOUT OVERLAY statement has XAXISOPTS= and
YAXISOPTS= options that enable you to specify axis features for the plots within
the layout.

The following sections explain how the axis features are determined in a layout.
The sections also discuss the issues that you must consider when setting the axis
type or adjusting the appearance of the axis display.
 1040 Chapter 7 / Axis Features in Layouts

How Axis Features Are Determined

Overview
GTL uses various criteria to determine the displayed axis features for a graph.
Generally, axis features are based on the following criteria:
n the layout type

n the order of plot statements in the layout and the options specified on those
statements
n the use of primary and secondary axes on the plots (when secondary axes are
supported)
n the plot type

n the column or columns of data that contribute to defining the axis range

n the data formats for the contributing data columns

Because the default axis features depend on a combination of the factors above, it
is useful to understand how the axis features are determined in the templates that
you build:
n how the data are mapped to the plot axes

n how the various layout types manage the axes.

Plot Data Are Mapped to a Designated Axis

Overview for Axis Mapping

Depending on the layout type and the plots that you specify within the layout, you
can manage up to four axes for two-dimensional plots:
n a primary X (bottom) axis

n a primary Y (left) axis

n a secondary X axis (X2 or “top” axis)

n a secondary Y axis (Y2 or “right” axis)

Within single-cell layouts (for example, OVERLAY layout), there can be just one
each of these four axes. However, within multi-cell layouts (for example, LATTICE
 How Axis Features Are Determined 1041

layouts), each cell can display the four axes. Thus, there can be multiple X, X2, Y,
and Y2 axes across the columns and rows in the layout grid. In a lattice-type layout,
you might have to use layout options to specify how the data ranges and axis
display should be managed. This section discusses the simpler case for axis
mapping, and “Axis Mapping in Lattice-type Layouts” on page 1042 discusses the
case for lattice-type layouts.

Note: GRIDDED layouts can be used to create a grid of cells, but the cells are
independent. Thus, axes in the grid cannot be managed across columns and rows,
so the simpler case applies to GRIDDED layouts.

Primary and Secondary Axes

By default, plot data are mapped to the primary axes. To enable you to override the
default, plot statements that support a secondary X2 axis provide an XAXIS=
option that can map data to the X or X2 axis. Plot statements that support a Y axis
provide a YAXIS= option that can map data to the Y or Y2 axis.

To determine the axis features within a layout, GTL must first determine what data
must be mapped to a particular axis. Thus, your use of primary and secondary axes
on plot specifications affects GTL’s determination of default axis features for the
layout.

For example, the plot statements in the following template specify Y-data
mappings to the Y2 and Y axes:
proc template;
define statgraph y2axis;
begingraph;
layout overlay;
histogram weight / scale=count yaxis=y2 ;
histogram weight / scale=percent yaxis=y ;
densityplot weight / normal();
endlayout;
endgraph;
end;
run;
proc sgrender data=sashelp.class template=y2axis;
run;
1042 Chapter 7 / Axis Features in Layouts

In this example, the first HISTOGRAM maps its Y-axis data to the Y2 axis, and the
second HISTOGRAM maps its Y-axis data to the Y axis. The DENSITY plot does not
explicitly map its Y-axis data, so the default Y axis is used. None of the plots
explicitly maps X-data, so the default X axis is used for all three plots. Thus, GTL
must manage any interactions that result from representing multiple plots on the X
and Y axes. For example, on the X axis, it must determine an appropriate data range
for representing the data values of all three plots.

When establishing axis features for each axis, GTL determines which plot
specifications map data to the axis. GTL also collects the data for all of the plots
that must be represented and maps that data to the designated axis. “When Plots
Share Data and a Common Axis” on page 1046 discusses the criteria GTL uses to
determine the axis features for the axes after this mapping has been done for each
axis.

Axis Mapping in Lattice-type Layouts

Lattice-type layouts (LAYOUT LATTICE, LAYOUT DATALATTICE, and LAYOUT
DATAPANEL) present a grid of graphs that automatically aligns plot areas and tick
display areas across grid cells. This alignment facilitates data comparisons among
graphs, and for those comparisons to be meaningful, the graph axes must be
coordinated across the columns and rows in the grid. All of the principles discussed
in “Overview for Axis Mapping” on page 1040 apply to the lattice-type layouts. In
addition, because there can be multiple X, X2, Y, and Y2 axes across grid cells, you
might have to use layout options to specify how the data ranges and axis display
should be managed.

For example, the following template uses a LAYOUT LATTICE to generate a grid
that displays a height analysis next to a weight analysis. By default in a LAYOUT
 How Axis Features Are Determined 1043

LATTICE statement, the options ROWDATARANGE= and ROW2DATARANGE= are

set to DATA. The DATA setting scales the Y-axis and Y2-axis data ranges
separately for each cell in the layout. To ensure that the Y-axis data range is the
same in both cells, the example specifies ROWDATARANGE=UNION. Similarly, to
ensure that the Y2-axis data range is the same in both cells, the example specifies
ROW2DATARANGE=UNION:
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union ;
layout overlay;
histogram height / scale=count yaxis=y2;
histogram height / scale=percent yaxis=y;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2;
histogram weight / scale=percent yaxis=y;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=y2axis;

run;

By default in a LAYOUT LATTICE statement, the options COLUMNDATARANGE=

and COLUMN2DATARANGE= are also set to DATA. But in this analysis, the height
is a separate measure from the weight, so the separate scales are appropriate for
the X-axes across cells. If the X-axes were displaying the same measure (for
1044 Chapter 7 / Axis Features in Layouts

example, comparing the height of female subjects to the height of male subjects),
then you could specify COLUMNDATARANGE=UNIONALL. This would set the
same scaling to the X-axis data ranges across the two layout columns. In this
example, you would not bother changing the default COLUMN2DATARANGE=
setting because the X2 axis is not needed.

Note: For DATALATTICE and DATAPANEL statements, UNIONALL is the default

value for the data ranges. Thus, you would not have to change the data ranges
unless you wanted to set UNION to scale data ranges per row or per column in the
layout.

In the example, scaling the data ranges across the row ensures proper axis scaling.
However, the graph display is cluttered by the duplicate display of ticks, axis
values, and axis labels on both the Y and Y2 axes. To simplify the display, you can
consolidate the axes. To do so, use a ROWAXES block to display a single Y axis for
both cells, and a ROW2AXES block to display a single Y2 axis for both cells. The
consolidated view removes the internal axes from the grid and displays only the
external axes:
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union;
rowaxes;
rowaxis / griddisplay=on;
endrowaxes;
row2axes;
rowaxis;
endrow2axes;
layout overlay;
histogram height / scale=count yaxis=y2;
histogram height / scale=percent yaxis=y;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2;
histogram weight / scale=percent yaxis=y;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=y2axis;

run;
 How Axis Features Are Determined 1045

When using ROWAXES or ROW2AXES blocks in a LATTICE layout, you nest within
the block one ROWAXIS statement for each row in the layout grid. The ROWAXIS
statements are applied sequentially to the rows, and each ROWAXIS statement
specifies the axis options for the Y or Y2 axes in its corresponding row. ROWAXIS
statements within the ROWAXES block apply to the Y axes, and ROWAXIS
statements within the ROW2AXES block apply to the Y2 axes. This example has
just a single row in the grid, so each block specifies only one ROWAXIS statement.
Notice that the ROWAXIS statement in the ROW2AXES block does not use any
options. Thus, it consolidates Y2 axes in the row into a single, external Y2 axis, but
it does not alter the default features of that axis. For columns in the grid, the
LATTICE layout provides COLUMNAXES and COLUMN2AXES blocks. These
blocks use COLUMNAXIS statements to externalize X and X2 axes and specify
their features.

When you use DATALATTICE and DATAPANEL layouts, the layout dynamically
generates a grid that contains as many cells as can be produced from the
combination of classification values. In those layouts the axes are always external,
and you can use the COLUMNAXISOPTS=, COLUMN2AXISOPTS=,
ROWAXISOPTS=, and ROW2AXISOPTS= options to specify the features for the
axes. The settings on each option apply across the entire grid. For example, if you
specify the ROWAXISOPTS= option in a DATALATTICE layout, then the specified
settings apply to the external Y axes in every row.
 1046 Chapter 7 / Axis Features in Layouts

When Plots Share Data and a Common Axis

Overview
If a layout block contains multiple plots that share data and a common axis, then
the plot settings often interact in ways that affect the axis features. Axis features
include the axis type, axis label, tick-mark layout, and so on. GTL resolves these
interactions in ways that vary according to the layout block and plot statements.

Note: Axis interactions might not occur if other settings in the template prevent
them. As discussed in “Plot Data Are Mapped to a Designated Axis” on page 1040,
if two plot statements are within an OVERLAY layout, then one of them might map
its data to the X axis and the other might map its data to the X2 (top) axis. Mapping
to separate axes can avoid the interactions that might occur if they both mapped
their data to the X axis.

Axis Features in Overlay-type Layouts

Overlay-type layouts (OVERLAY, OVERLAYEQUATED, and PROTOTYPE, for
example) build a composite from one or more GTL-statements.

Within overlay-type layouts, if you do not explicitly set axis features in your
template statements, then GTL automatically determines them. It sets the axis
features based on the layouts and plots in the layout block and the data that are
associated with the template at run time.

If only one plot statement within an overlay-type layout generates an axis, then
determining axis features is straight forward: the features are derived directly from
the plot type and the columns that are used for the plot data. For example, if a
LAYOUT OVERLAY block contains a single SCATTERPLOT and the X= argument
specifies a numeric column of children’s weights, then the default X-axis type is
LINEAR. The default X-axis label is the column label of the Weight column. If the
Weight column has no defined label, then the column name is used as a label.

When an overlay-type layout contains multiple plots that generate axes, GTL can
determine default axis features for the shared axes. Alternatively, you can use the
PRIMARY= option on one of the plot statements to specify which plot you want
GTL to use. The following code fragment explicitly specifies that the
SCATTERPLOT of children’s weights be used to determine axis features within the
layout:
layout overlay;
scatterplot x=weight ... / primary=true;
...
 How Axis Features Are Determined 1047

n If no plot in an overlay-type layout is designated as primary, then the first plot

that generates an axis is considered primary on a per-axis basis.
n If PRIMARY=TRUE for a plot within an overlay-type layout, then that plot’s data
columns, data type, and plot type determine the default axis features. An
explicitly specified primary plot determines the default axis features regardless
of where that plot statement occurs within the layout block.
n Only one plot can be primary on a per-axis basis. If multiple plots specify
PRIMARY=TRUE for the same axis, then the last one encountered is considered
primary.

The following SCATTERPLOT specifies a character column on the X= argument:

layout overlay;
scatterplot x=name ... / primary=true;
...

In this case, the default X-axis type is DISCRETE and the X-axis label is the label
that is assigned to column Name, or Name if no label is assigned to column Name.

Note: The SAS format on the primary plot’s column determines the axis format,
although the axis might not use that SAS format “as-is” from the column.

If a SCATTERPLOT’s X= argument specifies a column that has a SAS DATETIME

format, then the default X-axis type is TIME. The default X-axis label is the column
label or name of the DateTime column:
layout overlay;
scatterplot x=date ... / primary=true;
...

For some plot types, the default axis type does not directly correlate to the
specified column’s data type. For example, the following code fragment specifies a
BARCHART for the numeric column Age:
layout overlay;
barchart category=age ... / primary=true;
...

Because a BARCHART requires a discrete X axis, the default X-axis type in this
case is DISCRETE, in spite of the fact that column Age is numeric. The X-axis label
is the column label of Age, or the column name if no label exists.

Finally, consider a HISTOGRAM that is set as the primary plot in the layout and
that bins data values:
layout overlay;
histogram weight / binaxis=true primary=true;
...

In this case, the default X-axis type is LINEAR, but the histogram’s data bins are
used by default as the basis for the axis tick marks.
1048 Chapter 7 / Axis Features in Layouts

Axis Features in Data Panel and Data Lattice

Layouts
The criteria discussed in “Axis Features in Overlay-type Layouts” on page 1046
apply to determining the default axis features for the plots within DATAPANEL and
DATALATTICE layouts. Both of these layout types nest a LAYOUT PROTOTYPE
statement within their layout blocks. In both cases, the plot statements within the
LAYOUT PROTOTYPE block—an overlay-type layout—determine the axis features
for the plot display.

Axis Features in Lattice-type Layouts

The LAYOUT LATTICE statement can create a grid of graphs that automatically
aligns plot areas, data display areas, labels, and headers across the columns and
rows in the layout. The layout gives you the option of unifying the scale of the data
ranges that are displayed in the graphs.

If a LAYOUT LATTICE specification generates only one cell, then no competition

exists between cells for determining axis features in the display. In this case, the
axis features are derived directly from the plot type and the columns used for the
plot data.

Similarly, for multi-cell displays, if any or all of the options

COLUMNDATARANGE=, COLUMN2DATARANGE=, ROWDATARANGE=, or
ROW2DATARANGE= use the DATA setting to scale axis data ranges separately for
each cell in the layout, then the layout cells are data-independent. The data-
independent cells do not interact with each other for determining the axis features
in the display.

Axes are shared in the layout when one of the options COLUMNDATARANGE=,
COLUMN2DATARANGE=, ROWDATARANGE=, or ROW2DATARANGE= is used to
unite axis data ranges for layout cells. By default in those cases, the first cell that is
drawn (by default, the top left cell) determines the axis features in the display.
When UNIONALL is in effect, those same features are used in all of the grid’s
layout cells. When UNION is in effect, those same features are used on a per-row or
per-column basis. If you specify external axes for the columns or rows in the layout,
you can specify desired axis features on the appropriate COLUMNAXIS or
ROWAXIS statements used in the layout.

For an example LATTICE layout with external axes, see “Axis Mapping in Lattice-
type Layouts” on page 1042.

Axis Features in Gridded Layouts

In a GRIDDED layout the layout cells are independent of one another. Plot
statements within the layout cells do not share data and are not represented on a
 Controlling Axis Features 1049

common axis. Thus, no competition exists among layout cells for determining the
axis features.

Plot Axis Types Must Agree on Common Axes

GTL is extremely flexible and enables you to generate a wide variety of plot
displays. However, if you request incompatible plot displays within the same
layout, then the results are unpredictable.

“When Plots Share Data and a Common Axis” on page 1046 discusses the criteria
GTL uses to determine the default axis features. After the axis type has been
determined, GTL expects that all plots that share that axis will have the assigned
axis type. The expectation applies whether you specify axis features in your
template or let GTL determine default features.

For example, a BOXPLOT cannot be overlaid by a LINEPARM: the two types of plot
cannot share axes because the plot types are incompatible within the same set of
axes. Thus, if you were to use both a BOXPLOT statement and a LINEPARM
statement within a LAYOUT OVERLAY block, then only one of them can be
displayed. GTL therefore displays the primary plot (the first specified plot by
default, or the plot designated as primary by setting PRIMARY=TRUE). The other
plot is not displayed.

Similarly, a BARCHART requires a discrete X axis, whereas a HISTOGRAM cannot

be displayed on a discrete axis. If you specify both a BARCHART and a
HISTOGRAM within the same overlay-type layout, then only the primary plot is
displayed and the other plot is rejected from the display.

Axis types must also be the same for plots that must share an axis across the
columns or rows in a multi-cell layout. For example, in a LAYOUT LATTICE, GTL
expects that plots have the same axis type and data ranges if they are to share an
external axis. Otherwise, the external axis cannot be displayed for that row or
column.

Controlling Axis Features

Overview
To enable you to control axis features within each of the layout types, there are
different sets of axis options for the different types of axes:
 1050 Chapter 7 / Axis Features in Layouts

Option Category Statement

Two-dimensional axis options LAYOUT OVERLAY statement

Three-dimensional axis options LAYOUT OVERLAY3D statement

Equated axis options LAYOUT OVERLAYEQUATED statement

Lattice axis options LAYOUT LATTICE statement

DataLattice and DataPanel axes LAYOUT DATALATTICE and LAYOUT

DATAPANEL statements

The options that are available for each layout are documented separately, but it is
worth discussing the following tasks in general for all of the layout types:
n Setting the Axis Type

n Adjusting the Axis View

n Adjusting Axis Thresholds

n Adjusting Axis Offsets

Setting the Axis Type

Within any given layout in the graph display, each plot axis is always of a particular
type. In the default cases, the axis type is always LINEAR, DISCRETE, or TIME.

The axis options for each layout statement include a TYPE= option that enables
you to specify an axis type that overrides the default selection mechanisms. When
you override the default axis type, you must be sure to specify the correct axis type
for the plots that you are defining. For every plot in the template language, the
documentation indicates what axis types it supports. Plots statements that are
specified in the template are ignored if they are incompatible with the axis type.

Each axis type has features specific to that type, and the following axis options
enable you to specify features for the different types:
LINEAROPTS=(linear-suboptions)
DISCRETEOPTS=(discrete-suboptions)
TIMEOPTS=(time-suboptions)
LOGOPTS=(log-suboptions)

One or more of these options can be specified for an axis, but the specified settings
are applied only to the axis type that supports them.

For example, a bar chart has two axes – a TYPE=DISCRETE axis for the X axis and a
TYPE=LINEAR axis for the Y axis. If a numeric column (for example, Age) is
assigned to the X role, then this column’s values are always treated as discrete
values, never as a continuous range of values. You cannot request another axis type
for the X axis, but you can request a different axis type for the Y axis.
 Controlling Axis Features 1051

Sometimes you want a specialized axis type depending on the nature of the data.
For example, if the data have a very large range of values (orders of magnitude
apart), then you could request that the values be displayed on a logarithmic scale.
To set a logarithmic scale, use the TYPE=LOG axis option.

Time series data benefit from displaying the X axis with a TYPE=TIME axis. A TIME
axis type requires that the column values are SAS Date, Time, or Datetime values.

Three-dimensional plots such as BIHISTOGRAM3DPARM and

SURFACEPLOTPARM always use TYPE=LINEAR for X, Y, and Z axes.

Note: Certain plot types or layouts might impose restrictions on what type of axis
can be assigned. The documentation for each plot and layout type identifies any
restrictions that might apply to the axes.

Adjusting the Axis View

The VIEWMIN= and VIEWMAX= axis options can be used to adjust the view of an
axis. You can specify minimum data values to include in the display, maximum data
values, or both (the specified values might be adjusted by the threshold
calculation). By default, the VIEWMIN= value is the minimum data value for the
specified axis and the VIEWMAX= value is the maximum data value for the
specified axis.

A VIEWMIN= value that is greater than the data minimum or a VIEWMAX= value
that is less than the data maximum acts like a “zoom in” operation. The adjusted
view reduces the range of values represented on the axis and can sometimes
exclude markers, lines, or fills that would normally appear.

A VIEWMIN= value that is less than the data minimum or a VIEWMAX= value that
is greater than the data maximum acts like a “zoom out” operation. The adjusted
view extends the range of values represented on the axis and sometimes
compresses the markers, lines, or fills into a smaller area.

The following figure shows how the view settings can affect the tick and data
displays.
 1052 Chapter 7 / Axis Features in Layouts

Adjusting Axis Thresholds

On a continuous, linear axis, the THRESHOLDMIN= and THRESHOLDMAX= axis
options can be used to set a bias for including one more tick mark outside of either
end of the data range (or VIEWMIN to VIEWMAX range). The threshold range is
from 0 (do not include the tick mark) to 1 (include the tick mark). The default is
0.30. The bias at the minimum end of the axis is calculated using the
THRESHOLDMIN= value and the minimum data value (by default) or the
VIEWMIN= value (if set).

The bias at the maximum end of the axis is calculated using the THRESHOLDMAX=
value and the maximum data value (by default) or the VIEWMAX= value (if set).

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents the tick marks

from extending beyond the data range. Specifying THRESHOLDMIN=1 and
THRESHOLDMAX=1 ensures that the data range is bounded by tick marks.

The following figure shows how the threshold settings can affect the tick display on
an axis. In the figure, 8 is the minimum value for the display and 29 is the maximum
value.
 Controlling Axis Features 1053

Adjusting Axis Offsets

The OFFSETMIN= and OFFSETMAX= axis options can be used to reserve an area
at the minimum end of an axis, the maximum end, or both ends. No tick marks are
displayed in the reserved areas.

The offset range is from 0 to 1, and the specified value is used to calculate the
offset as a percentage of the full axis length. The larger the offset area that is
reserved, the less space is available for the tick display area. The default offset
reserves just enough area to fully display markers and other graphical features near
the ends of an axis.

The following figure shows how offset values of 0.08 might compare with the
default offsets for a continuous axis.
1054 Chapter 7 / Axis Features in Layouts

This next figure shows how offset values might affect the discrete axis of a bar
chart.
 1055

8
Axis Options in Layouts

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055
Axis Options for LAYOUT OVERLAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055
Axis Options for LAYOUT OVERLAY3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1119
Row and Column Axis Options for LAYOUT LATTICE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1140
Axis Options for LAYOUT OVERLAYEQUATED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1197
Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL . . . . . . . . . . . . 1218

Dictionary

Axis Options for LAYOUT OVERLAY

Axis options for the plots within an OVERLAY layout.

Interaction: The OVERLAY’s axis options are ignored when the LAYOUT OVERLAY statement is
nested within another layout type that has external axes in effect. For example, the
axis options are ignored when the OVERLAY is nested in a LAYOUT LATTICE with a
COLUMNAXIS= or ROWAXIS= option in effect.
Note: Unless otherwise indicated in an option description, each axis option is available for
the X, Y, X2, and Y2 axis.
See: “LAYOUT OVERLAY Statement” on page 159

Syntax
Axis options for the plots within an OVERLAY layout are specified with the
following
options on a LAYOUT OVERLAY statement:
1056 Chapter 8 / Axis Options in Layouts

XAXISOPTS=(axis-options)

YAXISOPTS=(axis-options)

X2AXISOPTS=(axis-options)

Y2AXISOPTS=(axis-options)

General Options for All Axes in an Overlay

The options that are documented in this section can be used with any of the axis
types that are supported within an OVERLAY layout. Subsequent sections in the
chapter document the axis options that are available only for specific axis types:
discrete, linear, log, or time axes. The following table provides a summary of the
options.

Statement Option Description

DISCRETEOPTS Specifies options for a discrete axis.

DISPLAY Controls which axis features are displayed on the

primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on the

secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the axis

label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should be

included in the displayed axis labels.

LABELSPLITJUSTIFY Specifies the justification of the strings that are

inside the axis label blocks.

LINEAROPTS Specifies features for a standard numeric interval

axis.
 Axis Options for LAYOUT OVERLAY 1057

Statement Option Description

LINEEXTENT Specifies the extent of the axis line.

LOGOPTS Specifies features for a log axis.

NAME Assigns a name to an axis for reference in other

statements.

OFFSETMAX Reserves an area at the maximum end of the axis.

No tick marks are displayed in the reserved area.

OFFSETMIN Reserves an area at the minimum end of the axis.

No tick marks are displayed in the reserved area.

REVERSE Specifies whether the tick values should appear in

the reverse order.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for the
axis length.

TICKSTYLE Specifies the placement of tick marks in relation

to the axis line.

TICKVALUEATTRS Specifies the color and font attributes of the axis

tick values.

TICKVALUEHALIGN Specifies the horizontal alignment for all of the

tick values that are displayed on the Y and Y2
axes.

TICKVALUEVALIGN Specifies the vertical alignment for all of the tick

values that are displayed on the X and X2 axes.

TIMEOPTS Specifies features for a TIME axis.

TYPE Specifies the type of axis to use.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not DISCRETE.

See “Options for Discrete Axes Only” on page 1072 for the options that
you can use for discrete-axis-options.
1058 Chapter 8 / Axis Options in Layouts

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall
chart can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. A secondary
axis is not an independent axis. Rather, it mirrors the primary axis. Thus, for this
option to take effect, all plot statements in the layout must map data to the
same primary axis. For example, a secondary X2 axis can be displayed on top in
the layout, provided all plot statements set XAXIS=X to map data to the
primary X axis (bottom). Similarly, a secondary Y2 axis can be displayed to the
right in the layout, provided all plot statements set YAXIS=Y to map data to the
primary Y axis (left).

NONE
specifies that no axis features are displayed
 Axis Options for LAYOUT OVERLAY 1059

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis

ALL
currently the same as STANDARD

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2,
both the X and X2 axis are primary and a secondary X axis cannot be
displayed. In that case, this option is ignored. The same applies for
the Y axes.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid
lines only. It does not affect the appearance of the minor grid lines.
To control the appearance of the minor grid lines on a log axis, use
the MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template
to absolutely control the display of grid lines or to allow interaction with the
current style to decide whether grid lines are displayed.
1060 Chapter 8 / Axis Options in Layouts

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element
in the current style contains DisplayOpts="ON."

AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in
the current style contains DisplayOpts="OFF."

ON
specifies that grid lines are always displayed.

The current style has no override.

OFF
specifies that grid lines are never displayed.

The current style has no override.

The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines
style element. Most supplied templates use the default setting AUTO_OFF to
indicate a preference for not displaying grid lines, but allowing the style to
override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style
has no preference about grid lines and the graphics template setting for
grid lines is always used.
 Axis Options for LAYOUT OVERLAY 1061

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The
list form implies that all included string literals or dynamic variables will be
concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 1046.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated
by default.

Tips Use the SHORTLABEL= option to specify an alternate axis label to

be used whenever truncation would normally occur.

Use the ODS escape sequence to specify Unicode characters in an

axis label. Here is an example that species degrees Fahrenheit in an
axis label for temperature:
label="Temperature ((*ESC*){unicode '00B0'x}F)"

See “Overriding the Default Axis Label” in SAS Graph Template Language:
User’s Guide

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELFITPOLICY=AUTO | SPLIT | SPLITALWAYS | STACKED

specifies a policy for fitting axis labels in the available space.

AUTO
uses the short label, when specified, instead of the original label.

If the short label does not fit, then it is clipped. When no short label is
specified, the original label is clipped.

SPLIT
splits the axis label at a split character only when necessary in order to make
the label fit the available space.

The split character is specified by the LABELSPLITCHAR= option. The short

label is not used. A split does not occur at a split character if a split is not
needed at that location. If the label does not contain any of the specified
split characters, then it is not split. A label that cannot be split or that does
1062 Chapter 8 / Axis Options in Layouts

not fit the available space even after splitting might overlap the adjoining
space.

SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option.

If the label cannot be split, then it is clipped.

STACKED
displays the Y or Y2 axis label vertically as stacked letters.

Note: This feature applies to SAS 9.4M5 and to later releases.

Valid in YAXISOPTS= and Y2AXISOPTS only

Default AUTO

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

When the overlay layout is nested in a LATTICE layout, SPLIT is

ignored and AUTO is used instead.

Note When LABELPOSITION=CENTER, the available area is the full axis,

including the axis offsets. When LABELPOSITION=DATACENTER,
the available area is the tick display area, excluding the axis offsets.

LABELPOSITION=CENTER | DATACENTER | TOP | BOTTOM | LEFT | RIGHT

specifies the position of the axis label.

CENTER
centers the axis label in the axis area.

For the Y and Y2 axes, the label is oriented vertically and is centered in the
axis area (including the offsets). The label is positioned to the left of the tick
values for the Y axis or to the right of the axis values for the Y2 axis. For the
X and X2 axes, the label is centered in the axis area (including the offsets). It
is positioned below the tick values for the X axis or above the axis values for
the X2 axis.

DATACENTER
centers the axis label in the axis tick display area.

For the Y and Y2 axes, the label is oriented vertically and is centered in the
axis tick display area (excluding the offsets). It is positioned to the left of the
tick values for the Y axis or to the right of the axis values for the Y2 axis. For
the X and X2 axes, the label is centered in the axis tick display area
(excluding the offsets). The label is positioned below the tick values for the
X axis or above the axis values for the X2 axis.

TOP
BOTTOM
orients the label horizontally at the top or bottom of the axis area.
 Axis Options for LAYOUT OVERLAY 1063

The label is right-justified in the axis area for the Y axis and left-justified for
the Y2 axis. If there is not sufficient room in the axis area to display the
label, then the label grows to the right for the Y axis and to the left for the
Y2 axis.

Restriction These options are valid for the Y and Y2 axes only.

Note When TOP or BOTTOM is used, the label might collide with other
graphical features. In that case, use CENTER or DATACENTER
instead.

LEFT
RIGHT
positions the label to the left or right of the axis area.

The label is centered vertically in the axis area.

Restriction These options are valid for the X and X2 axes only.

Note When LEFT or RIGHT is used, the label might collide with other
graphical features.

The following figure shows the CENTER and DATACENTER positions for a blue
Y-axis label Type and a red X-axis label MPG.

In this example, an axis offset is applied to the maximum end of both axes in
order to demonstrate the difference between CENTER and DATACENTER.
CENTER centers the labels on the entire axis area, including the offset.
DATACENTER centers the labels on the tick display areas, which does not
include the offset.

The next figure shows the TOP and LEFT positions, and the BOTTOM and
RIGHT positions for the same axis labels.
1064 Chapter 8 / Axis Options in Layouts

The following figure shows the CENTER and DATACENTER positions for a blue
Y-axis label Qtr and a red X-axis label Close.

The next figure shows the TOP and LEFT positions, and the BOTTOM and
RIGHT positions for the same axis labels.

Default CENTER

Restriction This option does not support collision avoidance. In some cases, axis
label collisions can occur in the axis area.

Interaction When LEFT, RIGHT, TOP, or BOTTOM is in effect, the

SHORTLABEL= option is ignored.

See SHORTLABEL= on page 1068 for information about how short labels
are used.

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the axis label. In that case, all of the specified split characters
together are treated as a single split character.

When LABELFITPOLICY=SPLIT, if the axis label does not fit the available
space, then it is split on a specified split character only if a split is needed at
that point to make the label fit. In this case, a split might not occur on every
split character. When LABELFITPOLICY=SPLITALWAYS, the axis label is split
unconditionally on every occurrence of a split character. If the axis label does
not contain any of the specified split characters, the label is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
labelsplitchar="abc"
 Axis Options for LAYOUT OVERLAY 1065

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.

TRUE
drops the split characters from the axis label display.

FALSE
includes the split characters in the axis label display.

When the label is split while LABELSPLITCHARDROP=FALSE is in effect,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.

justification

CENTER
LEFT
RIGHT
specifies the justification for the X or X2 axis label.

CENTER
TOP
BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.
1066 Chapter 8 / Axis Options in Layouts

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be
enclosed in parentheses. Each option is specified as a name = value pair and
each pair is space separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 1082 for the options that you
can use for linear-axis-options.

LINEEXTENT=FULL | DATA | number

specifies the extent of the axis line.

Note: This feature applies to SAS 9.4M3 and to later releases.

FULL
specifies an axis line that extends along the entire length of the axis.

DATA
specifies an axis line that extends through the data range from the minimum
offset to the maximum offset.

number
specifies how much the axis line extends from DATA toward FULL as a
decimal proportion. A value of 0 is equivalent to DATA, and a value of 1 is
equivalent to FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the
axis line at the midpoint positions of the minimum and maximum
bars. In that case, you can specify a numeric value to lengthen the
axis line so that it extends to the full width of both bars.

The following figure shows a simple example of each value for the X axis and Y-
axis lines. The light-blue dashed lines depict the minimum and maximum offsets
that are set on the axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.
 Axis Options for LAYOUT OVERLAY 1067

Interaction This option overrides the AXISLINEEXTENT= option in the

BEGINGRAPH statement.

Tip The graph wall outline might appear to be an axis line. In that case,
use the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in
the layout statement to suppress the wall outline.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not LOG.

See “Options for Log Axes Only” on page 1097 for the options that you
can use for log-axis-options.

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used
only in an AXISLEGEND statement.

Restriction This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or EXTRACTALWAYS.
In addition, an AXISLEGEND statement must be specified to
generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

1068 Chapter 8 / Axis Options in Layouts

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space. When
LABELPOSITION=CENTER (default), the available space for an axis label is the
full axis, including the axis offsets. When LABELPOSITION=DATACENTER, the
available space for an axis label is the axis tick display area, which excludes the
axis offsets. If the label length exceeds the available space, then the label is
anchored at the left or bottom offset. It extends beyond the opposing offset
until it reaches the end of the axis where it is truncated. An ellipsis designates
the truncation.

Interactions This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

This option has effect only when the LABELPOSITION= option is

set to CENTER or DATACENTER.

Note If the specified label is itself too long for the axis, it is truncated in
the display.

TICKSTYLE=OUTSIDE | INSIDE | ACROSS

specifies the placement of tick marks in relation to the axis line. The figure
shows the tick display for each value.
 Axis Options for LAYOUT OVERLAY 1069

OUTSIDE
displays tick marks outside of the axis frame.
INSIDE
displays tick marks inside the axis frame.

ACROSS
displays tick marks across the axis line.

Default The GraphAxisLines:TickDisplay style reference.

Restriction This option applies to the major tick marks only. It does not affect
the minor tick marks.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick marks.

Note This option has no effect on the placement of the tick values, which
are always outside the axis frame.

Tip You can use the Annotate facility to draw inside or across minor tick
marks in your graphs. See “Creating a Macro That Generates
Annotation Data” in SAS Graph Template Language: User’s Guide.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on
the Y and Y2 axes.
1070 Chapter 8 / Axis Options in Layouts

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on
the X and X2 axes.

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time,
date, or datetime format.

Options must be enclosed in parentheses. Each option is specified

as a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

 Axis Options for LAYOUT OVERLAY 1071

See “Options for Time Axes Only” on page 1107 for the options that
you can use for time-axis-options.

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.

AUTO
requests that the axis type be automatically determined, based on the
overlay contents.

For more information, see “When Plots Share Data and a Common Axis” on
page 1046.

DISCRETE
uses a DISCRETE axis if possible.

The data for discrete axes can be character or numeric. You can add a
DISCRETEOPTS= option list to customize this axis type.

LINEAR
uses a LINEAR axis if possible.

You can add a LINEAROPTS= option list to customize this axis type.

TIME
uses a TIME axis if possible.

Data for this axis must be SAS time, SAS date, or SAS datetime values. You
can add a TIMEOPTS= option list to customize this axis type.

LOG
uses a LOG axis if possible.

You can add a LOGOPTS= option list to customize this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur
for the response axis of a bar chart, line chart, needle plot, or
waterfall chart when a baseline intercept of 0 or less is specified.
It can also occur for the response axis of a waterfall chart when
an initial bar value of 0 or less is specified. To get a log response
axis in those cases, set the baseline intercept or initial bar value
to a positive value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within
the layout are dropped from the display if their data types or data
ranges do not match the axis type requirements. For more
information, see “Plot Axis Types Must Agree on Common Axes” on
page 1049.

After the axis type is determined (whether you set a specific type
or AUTO is in effect), you can use only options that are supported
by that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.
1072 Chapter 8 / Axis Options in Layouts

Options for Discrete Axes Only

The options that are documented in this section can be used with the
DISCRETEOPTS= axis option. The following table provides a summary of the
options.

Discrete Axis Options Description

COLORBANDS Specifies the display of alternating wall-color

bands corresponding to the discrete axis bins.

COLORBANDSATTRS Specifies the appearance of the alternating wall-

color bands.

SORTORDER Specifies how the values on the axis are to be

sorted.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on an axis.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the list of tick values that are displayed

on the axis.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESPLITCHAR Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in or dropped from the displayed tick
value.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are inside

the tick value block.

COLORBANDS=NONE | EVEN | ODD

specifies the display of alternating wall-color bands corresponding to the
discrete axis bins.

Default NONE
 Axis Options for LAYOUT OVERLAY 1073

Restriction This option applies to discrete axes only.

Interaction Specifying this option for more than one axis in the layout might
have unexpected results. The order in which color bands are drawn
might not match the order in which the axis options are specified.

Note The full width of a color band is the distance between midpoints.
When no axis offsets are specified, the first band begins at one-half
of the midpoint distance, and the last band ends at one-half of the
midpoint distance. When axis offsets are specified, the first and last
color bands on the axis might extend into their adjacent offsets by
as much as half the color-band width.

Tips Borders for the color bands can be added by setting TICKTYPE=
INBETWEEN in the DISCRETEOPTS= option, and by setting
GRIDDISPLAY= ON.

Because alternating color bands are drawn on top of the plot wall,
this option can be coordinated with the LAYOUT OVERLAY
statement’s WALLCOLOR= option.

COLORBANDSATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the appearance of the alternating wall-color bands. For the alternating
colors, one set uses the WALLCOLOR= colors that are set in the LAYOUT
OVERLAY statement, and the other set uses the colors set on this option.

Default The GraphBlock style element.

Restriction This option applies to discrete axes only.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

SORTORDER=AUTO | DATA
specifies how the axis values are to be sorted.

Note: This option applies to SAS 9.4M5 and to later releases.

AUTO
for a column in a SAS data set, enables the system to determine the best
sorting for the axis tick values.

For a column in a CAS in-memory table, sorts the tick values in ascending
order by the formatted values.

DATA
arranges the axis tick values in the order in which they appear in the data.

Note In SAS 9.4M6 and SAS 9.4M5, when data stored in a CAS in-memory
table is used and DATA is specified, unpredictable results might occur.
To generate consistent graphs in that case, specify AUTO. Starting
with SAS Viya 3.5 and SAS Studio 5.2, when data stored in a CAS in-
1074 Chapter 8 / Axis Options in Layouts

memory table is used and DATA is specified, DATA is ignored, and

AUTO is used instead.

Default AUTO

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

Example The following example specifies the axis tick values 10, 20, 30,
and 40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT
places the tick marks at the midpoint value location.

INBETWEEN
places the tick marks half way between adjacent midpoint locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

Note Starting with SAS 9.4M2, when TICKTYPE=INBETWEEN, the

outermost tick marks and grid lines at each end of the axis are not
drawn.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y and Y2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.
 Axis Options for LAYOUT OVERLAY 1075

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space.

A split does not occur at a split character if a split is not needed at that
location. If the value does not contain any of the specified split characters,
then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR= on page 1079

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR= on page 1079

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words
do not fit the available space.

SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit
the available space.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.
1076 Chapter 8 / Axis Options in Layouts

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space.

A split does not occur at a split character if a split is not needed at that
location. If the value does not contain any of the specified split characters,
then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR= on page 1079

 Axis Options for LAYOUT OVERLAY 1077

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR= on page 1079

SPLITROTATE
attempts the SPLIT policy, and then the ROTATE policy.

STACKEDALWAYS
always displays the tick values vertically as stacked letters.

Valid in XAXISOPTS= and X2AXISOPTS only

STACKEDALWAYSTHIN
always displays the tick values vertically as stacked letters.

Thinning is performed when the tick values do not fit the available space.

Valid in XAXISOPTS= and X2AXISOPTS only

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.

THIN
eliminates alternate tick values.

TRUNCATE
shortens the tick labels that exceed 12 characters to 12 characters.

A trailing ellipsis indicates that a label is truncated.

Note Data labels might overlap on a crowded axis.

TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.

TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.

TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults THIN for the Y and Y2 axes

ROTATE for the X and X2 axes

Note A note is written to the SAS log when tick value thinning occurs.
1078 Chapter 8 / Axis Options in Layouts

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.

string-list
a space-separated list of values, enclosed in parentheses.

You must enclose each value in the list in quotation marks.

Only the tick values that are included in the string list are displayed on the axis.
The values are displayed in the order in which they are listed. The data values
that are not in the list are dropped. The list can be a subset of the data values. It
can also contain values that are not included in the actual data. A tick value that
is not included in the data appears on the axis, but no data is represented at its
tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated

from adjacent values by a blank space.

Notes If the string list contains duplicate values, then the first
occurrence of the duplicated value in the list is honored and the
remaining instances are ignored.

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan,
Sports, Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30,
and 40:
 Axis Options for LAYOUT OVERLAY 1079

tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.
DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed.
When multiple characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in
the tick value. In that case, all of the specified split characters together are
treated as a single split character.

When TICKVALUESPLITPOLICY=SPLIT, if a tick value collision is detected, then

the tick value is split on a split character only if necessary at that point in order
to avoid collision. In that case, a split might not occur on every split character.
When TICKVALUEFITPOLICY=SPLITALWAYS, the tick value is split
unconditionally on every occurrence of a split character. If the tick value does
not contain any of the specified split characters, then it is not split.

"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
1080 Chapter 8 / Axis Options in Layouts

tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used to

manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether the

split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick
values. The split characters are specified by the TICKVALUESPLITCHAR=
option.

TRUE
drops the split characters from the tick value display.

The following figure shows an example in which

TICKVALUESPLITCHARDROP=TRUE and three-word, asterisk-delimited
tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display.

The fit policy determines how the characters are displayed. If the display
policy is SPLIT or SPLITTHIN and TICKVALUESPLITCHARDROP=FALSE,
then each tick value is split at a split character only where a split is
necessary in order to make the value fit the available space. A split might not
 Axis Options for LAYOUT OVERLAY 1081

occur at every split character in the tick value. At each split point, the split
character remains as the last character in the current line. The characters
that follow the split character, up to and including the split character at the
next split point, are then wrapped to the following line. This process repeats
until the entire data tick value is displayed. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this
case, a split is not needed at the second asterisk.

If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and

TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line. The
following figure shows an example in which
TICKVALUESPLITCHARDROP=FALSE and three-word, asterisk-delimited
tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that a split occurs after each asterisk and each asterisk appears at
the end of the line. In this case, three lines are displayed.

Default TRUE

Interactions The TICKVALUESPLITCHAR= option specifies the split character or

characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not
affect the display of tick values that are not split.
1082 Chapter 8 / Axis Options in Layouts

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Linear Axes Only

The options that are documented in this section can be used only with the
LINEAROPTS= axis option. The following table provides a summary of the options.

Linear Axis Option Description

INCLUDERANGES on page 1083 Specifies one or more ranges for a broken axis.

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

ORIGIN Specifies that the axis perpendicular to the

current axis be drawn at the indicated data value.
 Axis Options for LAYOUT OVERLAY 1083

Linear Axis Option Description

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined by the TICKVALUELIST=
option.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision.

Only the default policy (THIN) is available for a Y
or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a linear

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by start,

end, and increment.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

INCLUDERANGES=(start–end <start2–end2 startN–endN …>)

specifies the ranges for a broken axis.

Note: This option is valid starting in SAS 9.4M1.

start
specifies a numeric value or the keyword MIN.

MIN specifies the minimum data value.

end
specifies a numeric value or the keyword MAX.

MAX specifies the maximum data value.

1084 Chapter 8 / Axis Options in Layouts

The following figure shows a linear axis, broken into ranges 50–52 and 56–73.

The following figure shows a time axis, broken into ranges

'01Jan2001'd–'01May2003'd and '01Jan2005'd–'01Oct2005'd.
As shown in the figure, break lines are drawn to indicate the break in the axis.

Restrictions This option is valid for linear and time axes in an OVERLAY layout
only.

Only one axis can be broken. If this option is specified for both
axes, then it is honored for the vertical axis and ignored for the
horizontal axis.

When plots are associated with the X and X2 axes or with Y and
Y2 axes, neither axis can be broken.

A binned heat map or histogram axis cannot be broken.

A broken axis is not supported in vector graphics output. When a

broken axis is specified and vector graphics output is requested,
the graph is converted into an image instead. A note indicating the
conversion is written to the SAS log. To restore the vector
graphics output in that case, remove the INCLUDERANGES=
option from the LINEAROPTS= or TIMEOPTS= option.

Requirements All of the ranges must be enclosed in parenthesis.

You must specify each range as a starting value, a hyphen, and an

ending value. You must separate adjacent ranges with a space.

Each range must be nonzero. A zero range such as 12–12 is

considered invalid.

Interactions When this option is specified, axis options THRESHOLDMIN=,

THRESHOLDMAX=, VIEWMIN=, VIEWMAX=, and
TICKVALUEPRIORITY= are ignored. Suboption EXTRACTSCALE=
of the TICKVALUEFORMAT= option is also ignored.
 Axis Options for LAYOUT OVERLAY 1085

When this option is specified, data tips are not displayed, and the
plot URL= option is ignored.

Notes When this option is specified, data-clipping might occur for the
following: plot markers and marker characters, box-plot outlier
markers, fixed-position data labels, needle plots and fringe plots
in the X direction, reference lines and drop lines on the broken
axis, axis tables, and relative bubble plots.

Curve label positions are based on the contiguous axis data range.
When curve labels are specified with a broken axis, the curve
label positions might not be ideal.

Tip Starting with SAS 9.4M3, you can use the AXISBREAKTYPE= and
AXISBREAKSYMBOL= options in the BEGINGRAPH statement to
display the break in the axis as only a symbol on the axis line.

See “Creating a Broken Linear Axis” in SAS Graph Template Language:

User’s Guide

Example includeranges=(10-500 1000-5000 10000-50000)

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the TICKVALUEFORMAT=
option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

Note This option is ignored when the data exceeds the limits of integer
values.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
1086 Chapter 8 / Axis Options in Layouts

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.
 Axis Options for LAYOUT OVERLAY 1087

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in SAS 9.4M1 and earlier releases.

One tick with two intervals starting with SAS 9.4M2.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.
1088 Chapter 8 / Axis Options in Layouts

ORIGIN=number
specifies that the axis perpendicular to the current axis is drawn at the indicated
data value. For managing origin settings, GTL treats the X and Y axes as a pair,
and the X2 and Y2 axes as a separate pair. Thus, if you set the Y-axis origin to
200, then the X axis is drawn from that origin point. If the graph also displays an
X2 axis, then it is unaffected and does not move. Similarly, if you set an origin
for the Y2 axis, then the X2 axis moves to that origin point and the X axis is
unaffected.

If you set an origin for the Y2 axis and there is no X2 axis, then the origin setting
for Y2 does not affect the graph display. That is, the X axis does not move to
that origin point.

If you set an origin for an axis and the axis has a tick value at that origin value,
the tick value is not displayed. Suppressing the tick value at the origin prevents
the value from colliding with the axis value on the perpendicular axis. However,
it is possible that the tick values on the orthogonal axes will collide.

Default The axis perpendicular to the current axis is drawn at the minimum
tick value minus the OFFSETMIN= value.

Restriction This option applies to linear axes only.

Interactions If the specified value is outside the data range for the current axis,
then the data range is extended to include the value.

The axis line, ticks, and tick values of the “perpendicular” axis move
to the location indicated by the origin. The axis label is not moved.

Tip This option is often used to create Cartesian axes (axes centered at
ORIGIN=0).

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

This option is ignored when the INCLUDERANGES= option is

specified.

Tips If the threshold is set to 0, the potential tick mark is never

displayed. If the threshold is set to 1, then the tick mark is always
displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.
 Axis Options for LAYOUT OVERLAY 1089

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

This option is ignored when the INCLUDERANGES= option is

specified.

Tips If the threshold is set to 0, the potential tick mark is never

displayed. If the threshold is set to 1, then the tick mark is always
displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option is

ignored.
1090 Chapter 8 / Axis Options in Layouts

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y and Y2 axes, the following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.
 Axis Options for LAYOUT OVERLAY 1091

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.

(format-options)
specifies one or more formatting options for major tick values.

Together, these options provide parameters for determining an optimal

format (w.d, Ew., BESTw.) for displaying major tick values.

MAXWIDTH=integer
specifies the maximum width for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values.

The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.

The scale can be a named scale or a scientific-notation scale. The

EXTRACTSCALETYPE= option specifies the scale type. The scale that is
used is appended to the axis label, as shown in the following example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the
label is truncated, and the scale is appended to the truncated label.
Ellipses indicate that the label was truncated, as shown in the following
example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.
1092 Chapter 8 / Axis Options in Layouts

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

The EXTRACTSCALE= suboption is ignored when the

INCLUDERANGES= option is specified.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis.
In that case, the tick value format might differ from the data
format even when a named format is applied to the data
values.

See “boolean ” on page xii for other Boolean values that you can
use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.

A named scale can be millions, billions, or trillions for values of 999

trillion or less, or a multiple of 10 (denoted as 10^n) for values over
999 trillion. For large tick values, the scale factor is set to ensure that
the absolute value of the largest value is greater than 1. For small
fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for

values greater than 1, or a multiple of 1/10 expressed as 10^–n for
values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621 .
 Axis Options for LAYOUT OVERLAY 1093

Note If you specify a format that significantly reduces precision,

because of tick-value rounding, the plot data elements might not
align properly with the axis tick values. In that case, specify a
tick-value format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option
is used, the only tick values that appear are the tick values in
numeric-list that fall within the explicit data range (set by
VIEWMIN= and VIEWMAX=) or the implicit data range (set by the
actual data minimum and data maximum).

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the DISPLAYSECONDARY=
option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some data
not being displayed. For example, data might not be displayed
when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data
range of the axis, but cannot reduce it.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.
1094 Chapter 8 / Axis Options in Layouts

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored when the INCLUDERANGES= option is

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.
 Axis Options for LAYOUT OVERLAY 1095

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(START=number END=number INCREMENT=number)

specifies the tick values by start, end, and increment.

START=number
specifies the value for the first tick mark.

END=number
specifies the value for the last tick mark.

INCREMENT=number
specifies the increment for intermediate tick marks between the first and
last tick marks.

The END value always controls the last tick mark. The interval between the
last tick mark and the previous tick mark might not necessarily be the
INCREMENT value.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option is
used, the only tick values that appear are those that fall within the
explicit data range (set by VIEWMIN= and VIEWMAX=) or the
implicit data range (set by the actual data minimum and data
maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting
for the TICKVALUEFORMAT= option.
1096 Chapter 8 / Axis Options in Layouts

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the maximum axis tick value that
is displayed. The THRESHOLDMAX= value is used to determine
the maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the minimum axis tick value that
is displayed. The THRESHOLDMIN= value is used to determine
the minimum tick value.
 Axis Options for LAYOUT OVERLAY 1097

This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Options for Log Axes Only

The options that are documented in this section can be used with the LOGOPTS=
axis option. The following table provides a summary of the options.

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for the

axis values.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.

TICKINTERVALSTYLE Specifies how to scale and format the values for

major tick marks.

TICKVALUEFORMAT= Specifies how to format the values for major tick

marks.
1098 Chapter 8 / Axis Options in Layouts

Log Axis Option Description

TICKVALUELIST Specifies the tick values for a log axis as a space-

separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST=

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=, VIEWMAX=,
and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.
 Axis Options for LAYOUT OVERLAY 1099

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.
1100 Chapter 8 / Axis Options in Layouts

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform log tick values such as 10, 20, 30, and 40, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform log tick values such as 1, 10, 100, and 1000 in
the TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to log axes only.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

See “Adjusting Axis Thresholds” on page 1052

 Axis Options for LAYOUT OVERLAY 1101

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to log axes only.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

See “Adjusting Axis Thresholds” on page 1052

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.

AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data.

When the data range is small (within an order of magnitude), a LINEAR

representation is typically used. Data ranges that encompass several orders
of magnitude typically use the LOGEXPAND or LOGEXPONENT
representation.

LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are only the integer exponents for all bases.
1102 Chapter 8 / Axis Options in Layouts

LINEAR
places the major tick marks at non-uniform intervals that cover the range of
the data.

Default AUTO

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format"
on the column is retained. For LINEAR, ticks values are
automatically formatted when the column format is not assigned or
one of w.d, Ew., or BESTw. Other formats (SAS defined or user-
defined) are used if specified.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than
or equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you might

want to include information in the axis label about which base is
used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.
 Axis Options for LAYOUT OVERLAY 1103

When TICKINTERVALSTYLE=LOGEXPAND, this option is honored

for the base 10 and base 2 logarithmic scales, and is ignored for the
base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the
explicit data range set by the VIEWMIN= and VIEWMAX= options
or by the implicit data range set by the actual data minimum and
data maximum are displayed. An internal algorithm determines
the tick marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results
might occur. If VALUESTYPE=EXPANDED is in effect (default),
specify increments of the log base power such as 0.1, 1, 10, 100,
and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify integer increments
of the log base power exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some
data not being displayed. For example, data might not be
displayed when the VIEWMIN= value is greater than the actual
data minimum, or when the VIEWMAX= value is less than actual
data maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some
data might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

If MINORTICKS=TRUE is specified and TICKVALUELIST=

specifies nonuniform log tick values such as 10, 20, 30, and 40,
then MINORTICK=TRUE is ignored. To display minor tick marks in
1104 Chapter 8 / Axis Options in Layouts

that case, specify uniform log tick values such as 1, 10, 100, and
1000 in the TICKVALUELIST= option.

See VIEWMIN= and VIEWMAX= options for controlling the data range

TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the
TICKVALUELIST=, VIEWMAX=, and VIEWMIN= option values. Use this option
to choose your preferred way of specifying log-axis values.

EXPANDED
values are interpreted as powers of the base such as 0.1, 1, 10, 100, and so on,
for base 10, for example.
 Axis Options for LAYOUT OVERLAY 1105

EXPONENT
values are interpreted as integer exponents of the base such as 1, 2, 3, and so
on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base
10 log axis. This results in X-axis tick values of 10, 100, 1000, 10000,
and 100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10
log axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so
on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

1106 Chapter 8 / Axis Options in Layouts

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
default value for VIEWMAX= is less than the value specified by
the VIEWMIN= option, then the VIEWMIN= and VIEWMAX=
values are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 100,000 as an

expanded value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an

exponent value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
 Axis Options for LAYOUT OVERLAY 1107

increment of the log base power exponent such as 1, 2, 3, and so

on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 0.1 as an expanded

value on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent

value on a base 10 log axis:
VIEWMIN=-1

Options for Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS=
axis option. The following table provides a summary of the options.

Time Axis Option Description

INCLUDERANGES on page 1108 Specifies one or more ranges for a broken axis.

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time interval

that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.
1108 Chapter 8 / Axis Options in Layouts

Time Axis Option Description

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor ticks.

MINORTICKS Specifies whether minor ticks are displayed.

SPLITTICKVALUE Specifies whether to split the tick values on an X

or X2 axis, if possible. This option is not available
on a Y or Y2 axis.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on an X or X2 axis. This option is not available on
a Y or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a time

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

INCLUDERANGES=(start–end <start2–end2 startN–endN …>)

specifies the ranges for a broken axis.

Note: This option is valid starting in SAS 9.4M1.

start
specifies a SAS time, date, or date-time constant, or the keyword MIN.

MIN specifies the minimum data value.

end
specifies a SAS time, date, or date-time constant, or the keyword MAX.

MAX specifies the maximum data value.

The following figure shows a linear axis, broken into ranges 50–52 and 56–73.
 Axis Options for LAYOUT OVERLAY 1109

The following figure shows a time axis, broken into ranges

'01Jan2001'd–'01May2003'd and '01Jan2005'd–'01Oct2005'd.

As shown in the figure, break lines are drawn to indicate the break in the axis.

Restrictions This option is valid for linear and time axes in an OVERLAY layout
only.

Only one axis can be broken. If this option is specified for both
axes, then it is honored for the vertical axis and ignored for the
horizontal axis.

When plots are associated with the X and X2 axes or with Y and
Y2 axes, neither axis can be broken.

A binned heat map or histogram axis cannot be broken.

A broken axis is not supported in vector graphics output. When a

broken axis is specified and vector graphics output is requested,
the graph is converted into an image instead. A note indicating the
conversion is written to the SAS log. To restore the vector
graphics output in that case, remove the INCLUDERANGES=
option from the LINEAROPTS= or TIMEOPTS= option.

Requirements All of the ranges must be enclosed in parenthesis.

You must specify each range as a starting value, a hyphen, and an

ending value. You must separate adjacent ranges with a space.

Each range must be nonzero. A zero range such as 12–12 is

considered invalid.

Interactions When this option is specified, axis options THRESHOLDMIN=,

THRESHOLDMAX=, VIEWMIN=, VIEWMAX=, and
TICKVALUEPRIORITY= are ignored. Suboption EXTRACTSCALE=
of the TICKVALUEFORMAT= option is also ignored.
1110 Chapter 8 / Axis Options in Layouts

When this option is specified, data tips are not displayed, and the
plot URL= option is ignored.

Notes When this option is specified, data-clipping might occur for the
following: plot markers and marker characters, box-plot outlier
markers, fixed-position data labels, needle plots and fringe plots
in the X direction, reference lines and drop lines on the broken
axis, axis tables, and relative bubble plots.

Curve label positions are based on the contiguous axis data range.
When curve labels are specified with a broken axis, the curve
label positions might not be ideal.

Tip Starting with SAS 9.4M3, you can use the AXISBREAKTYPE= and
AXISBREAKSYMBOL= options in the BEGINGRAPH statement to
display the break in the axis as only a symbol on the axis line.

See “Creating a Broken Time Axis” in SAS Graph Template Language:

User’s Guide

Example includeranges=('01Jan2001'd-'01May2003'd
'01Jan2005'd-'01Oct2005'd)

INTERVAL=AUTO | SECOND | MINUTE | HOUR | DAY | TENDAY | WEEK |

SEMIMONTH | MONTH | QUARTER | SEMIYEAR | YEAR
specifies the time interval between major ticks.
Table 8.1 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME
 Axis Options for LAYOUT OVERLAY 1111

Default tick value

INTERVAL Unit Tick interval format

SEMIMONTH DATE or 1st and 16th of DATE9.

DATETIME each month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if
the data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10
1112 Chapter 8 / Axis Options in Layouts

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.
 Axis Options for LAYOUT OVERLAY 1113

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.10 on page 1110 for
information about the intervals that you can select. The interval that you select
must be consistent with the axis data duration units such as TIME, DATE, or
DATETIME. For example, if the axis data is in TIME units, then you must specify
AUTO, SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on
the value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.
1114 Chapter 8 / Axis Options in Layouts

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, then MINORTICKS=TRUE is ignored, and minor tick marks
are not displayed.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This
option is not available for a Y or Y2 axis.

TRUE
splits the axis tick values into two lines allowing more tick values to appear.

For example, with INTERVAL= MONTH, this is how tick values are split:

FALSE
does not split the axis tick values.

For example, when this option specifies FALSE, this is how the tick values in
the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option
is not available for the Y and Y2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the
axis. The following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.
 Axis Options for LAYOUT OVERLAY 1115

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= on page 1189TRUE, this option is ignored

and only the THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.

format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed.

This format must be in the same duration units as the data column(s)
mapped to a time axis: TIME, DATE, or DATETIME and should be appropriate
for the value of the INTERVAL= option. For example, if INTERVAL=MONTH
and there are two years of data displayed on the axis, then choosing
TICKVALUEFORMAT=YEAR. would result in several ticks having the same
year value.

DATA
specifies that the SAS date, time, or datetime format associated with the
data column assigned to the axis be used to control how the major tick
values are displayed.

Default The default format used by the INTERVAL= option. The default
does not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

1116 Chapter 8 / Axis Options in Layouts

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range or that specified by the VIEWMIN= or
VIEWMAX= options, then they are used.

Requirement The tick values must be specified as a space-separated list of

values enclosed in parentheses. The items in the list must be in the
same duration units as the data mapped to the axis: TIME, DATE,
or DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format
specified on the TICKVALUEFORMAT= option. If
TICKVALUEFORMAT= is not used, then the values are formatted
according to the column format (the default TICKVALUEFORMAT
value is not applied to these values).

If this option is specified, the SPLITTICKVALUE= and INTERVAL=

options are ignored.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, MINORTICKS=TRUE is ignored, and minor tick marks are
not displayed.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE
 Axis Options for LAYOUT OVERLAY 1117

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

This option is ignored when the INCLUDERANGES= option is

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for XAXISOPTS= and X2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.
1118 Chapter 8 / Axis Options in Layouts

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

This option is ignored when the INCLUDERANGES= option is

specified.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

 Axis Options for LAYOUT OVERLAY3D 1119

Details
The LAYOUT OVERLAY statement provides the XAXISOPTS=, YAXISOPTS=,
X2AXISOPTS=, Y2AXISOPTS= options that enable you to manage the axis display
separately for the X, Y, X2, and Y2 axes. The following example template uses the
YAXISOPTS= option to manage the grid lines, tick marks, and tick values on a Y
axis:

begingraph;
layout overlay /
yaxisopts=(
griddisplay=on
display=(ticks tickvalues)
);
seriesplot x=month y=predict;
endlayout;
endgraph;

Within an OVERLAY layout block, each plot axis is always of a particular type. In
the default cases, the axis type is always DISCRETE, LINEAR, or TIME. The TYPE=
option enables you to specify an axis type that overrides the default. For example,
when appropriate for the data, you can request a LOG axis. When you override the
default axis type, you must be sure to specify the correct axis type for the plot(s)
that you are defining.

Each axis type has features specific to that type, and the following axis options
enable you to specify features for the different types: DISCRETEOPTS= ,
LINEAROPTS= , LOGOPTS= , and TIMEOPTS= . One or more of these options can
be specified for an axis, but the specified settings are applied only to the axis type
that supports them.

Axis Options for LAYOUT OVERLAY3D

Axis options for the plots within an OVERLAY3D layout.

Restriction: The OVERLAY3D layout supports linear and time axes only.
Note: Unless otherwise indicated in an option description, each axis option is available for
the X, Y, and Z axis.
See: “LAYOUT OVERLAY3D Statement” on page 178

Syntax
Axis options for the plots within an OVERLAY3D layout are specified with the
following options on a LAYOUT OVERLAY3D statement:

XAXISOPTS=(axis-options)

YAXISOPTS=(axis-options)
1120 Chapter 8 / Axis Options in Layouts

ZAXISOPTS=(axis-options)

General Options for All Axes in an Overlay3D

The options that are documented in this section can be used with either axis type
that is supported within an OVERLAY3D layout. Subsequent sections in the
chapter document the axis options that are available only for the specific axis type:
linear or time.

Statement Option Description

DISPLAY Controls which axis features are displayed.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the axis

label.

LINEAROPTS Specifies options for a standard numeric interval

axis.

OFFSETMAX Reserves an area at the maximum end of the axis.

No tick marks are displayed in the reserved area.

OFFSETMIN Reserves an area at the minimum end of the axis.

No tick marks are displayed in the reserved area.

TICKVALUEATTRS Specifies the color and font attributes of the axis

tick values.

TIMEOPTS Specifies options for a TIME axis.

TYPE Specifies the type of axis to use.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed
 Axis Options for LAYOUT OVERLAY3D 1121

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall
chart can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template
to absolutely control the display of grid lines or to allow interaction with the
current style to decide whether grid lines are displayed.

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element
in the current style contains DisplayOpts="ON."

AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in
the current style contains DisplayOpts="OFF."
1122 Chapter 8 / Axis Options in Layouts

ON
specifies that grid lines are always displayed.

The current style has no override.

OFF
specifies that grid lines are never displayed.

The current style has no override.

The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines
style element. Most supplied templates use the default setting AUTO_OFF to
indicate a preference for not displaying grid lines, but allowing the style to
override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style
has no preference about grid lines and the graphics template setting for
grid lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The
list form implies that all included string literals or dynamic variables will be
concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 1046.

Interaction This option is ignored if the DISPLAY= option does not display the
axis label.
 Axis Options for LAYOUT OVERLAY3D 1123

Note If the axis label is too long to fit along the axis, then it is truncated
by default.

Tip Use the ODS escape sequence to specify Unicode characters in an

axis label. Here is an example that species degrees Fahrenheit in an
axis label for temperature:
label="Temperature ((*ESC*){unicode '00B0'x}F)"

See “Overriding the Default Axis Label” in SAS Graph Template Language:
User’s Guide

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= option does not display the
axis label.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be
enclosed in parentheses. Each option is specified as a name = value pair and
each pair is space separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 1125 for the options that you
can use for linear-axis-options.

OFFSETMAX=AUTO | number
reserves an area at the maximum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

number
specifies the offset as a decimal proportion of the full axis length.

For a continuous axis, the offset follows the highest data value or highest
tick value, whichever is greater.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

1124 Chapter 8 / Axis Options in Layouts

OFFSETMIN=AUTO | number
reserves an area at the minimum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the minimum end of an axis.

number
specifies the offset as a decimal proportion of the full axis length.

For a continuous axis, the offset precedes the lowest data value or lowest
tick value, whichever is less.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= option does not display tick
values.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time,
date, or datetime format.

Options must be enclosed in parentheses. Each option is specified

as a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

See “Options for Time Axes Only” on page 1134 for the options that
you can use for time-axis-options.

TYPE=AUTO | LINEAR | TIME

specifies the type of axis to use.

AUTO
requests that the axis type be automatically determined by the plot or the
overlay contents.
 Axis Options for LAYOUT OVERLAY3D 1125

LINEAR
uses a LINEAR axis if possible.

You can add a LINEAROPTS= option list to customize this axis type.

TIME
uses a TIME axis if possible.

Data for this axis must be SAS time, SAS date, or SAS datetime values. You
can add a TIMEOPS= option list to customize this axis type.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within
the layout are dropped from the display if their data types or data
ranges do not match the axis type requirements. For more
information, see “Plot Axis Types Must Agree on Common Axes” on
page 1049.

After the axis type is determined (whether you set a specific type
or AUTO is in effect), only options supported by that axis type can
be used. For example, if TYPE=TIME, then only the general
OVERLAY3D axis options and those available on TIMEOPS= are
supported.

Options for Linear Axes Only

This section documents the options that can be used with the LINEAROPTS= axis
option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.
1126 Chapter 8 / Axis Options in Layouts

Linear Axis Option Description

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a linear

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification

(TICKVALUELIST= or TICKVALUESEQUENCE=)
can extend the axis data range.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by start,

end, and increment.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the TICKVALUEFORMAT=
option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

Note This option is ignored when the data exceeds the limits of integer
values.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
 Axis Options for LAYOUT OVERLAY3D 1127

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.
1128 Chapter 8 / Axis Options in Layouts

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in SAS 9.4M1 and earlier releases.

One tick with two intervals starting with SAS 9.4M2.

Interactions The DISPLAY= option specification must include TICKS for this
option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions This option is ignored if the TICKVALUELIST= option is used or if

the tick marks are not enabled by the DISPLAY= option.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.
 Axis Options for LAYOUT OVERLAY3D 1129

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.
1130 Chapter 8 / Axis Options in Layouts

See “Adjusting Axis Thresholds” on page 1052

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option is

ignored.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.

(format-options)
specifies one or more formatting options for major tick values.

Together, these options provide parameters for determining an optimal

format (w.d, Ew., BESTw.) for displaying major tick values.

MAXWIDTH=integer
specifies the maximum width for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values.

The actual number might vary based on other constraints.

 Axis Options for LAYOUT OVERLAY3D 1131

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.

The scale can be a named scale or a scientific-notation scale. The

EXTRACTSCALETYPE= option specifies the scale type. The scale that is
used is appended to the axis label, as shown in the following example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the
label is truncated, and the scale is appended to the truncated label.
Ellipses indicate that the label was truncated, as shown in the following
example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis.
In that case, the tick value format might differ from the data
format even when a named format is applied to the data
values.

See “boolean ” on page xii for other Boolean values that you can
use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.

A named scale can be millions, billions, or trillions for values of 999

trillion or less, or a multiple of 10 (denoted as 10^n) for values over
999 trillion. For large tick values, the scale factor is set to ensure that
the absolute value of the largest value is greater than 1. For small
fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.
1132 Chapter 8 / Axis Options in Layouts

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for

values greater than 1, or a multiple of 1/10 expressed as 10^–n for
values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621 .

Note If you specify a format that significantly reduces precision,

because of tick-value rounding, the plot data elements might not
align properly with the axis tick values. In that case, specify a
tick-value format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick values.

Restriction This option applies to linear axes only.

Requirement The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option does not display tick values.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.
 Axis Options for LAYOUT OVERLAY3D 1133

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the actual data minimum
and data maximum.

Default FALSE

Restriction This option applies to linear axes only.

Interactions This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option does not display the
tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUESEQUENCE=(START=number END=number INCREMENT=number)

specifies the tick values by start, end, and increment.

START=number
specifies the value for the first tick mark.

END=number
specifies the value for the last tick mark.

INCREMENT=number
specifies the increment for intermediate tick marks between the first and
last tick marks.

The END value always controls the last tick mark. The interval between the
last tick mark and the previous tick mark might not necessarily be the
INCREMENT value.

Default An internal algorithm determines the tick marks.

Interactions This option overrides the INTEGER= option.

1134 Chapter 8 / Axis Options in Layouts

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option does not display tick
marks.

Tip The values in the sequence are formatted according to the setting
for the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

Options for Time Axes Only

This section documents the options that can be used with the TIMEOPTS= axis
option. The following table provides a summary of the options.

Time Axis Option Description

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time interval

that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor ticks.

MINORTICKS Specifies whether minor ticks are displayed.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a time

axis as list.

INTERVAL=AUTO | SECOND | MINUTE | HOUR | DAY | TENDAY | WEEK |

SEMIMONTH | MONTH | QUARTER | SEMIYEAR | YEAR
specifies the time interval between major ticks.
 Axis Options for LAYOUT OVERLAY3D 1135

Table 8.2 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of DATE9.

DATETIME each month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if
1136 Chapter 8 / Axis Options in Layouts

the data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.
 Axis Options for LAYOUT OVERLAY3D 1137

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.11 on page 1135 for
information about the intervals that you can select. The interval that you select
must be consistent with the axis data duration units such as TIME, DATE, or
DATETIME. For example, if the axis data is in TIME units, then you must specify
AUTO, SECOND, MINUTE, or HOUR.

Default AUTO
1138 Chapter 8 / Axis Options in Layouts

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on
the value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used or if

the tick marks are not enabled by the DISPLAY= option.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, then MINORTICKS=TRUE is ignored, and minor tick marks
are not displayed.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.

format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed.

This format must be in the same duration units as the data column(s)
mapped to a time axis: TIME, DATE, or DATETIME and should be appropriate
for the value of the INTERVAL= option. For example, if INTERVAL=MONTH
and there are two years of data displayed on the axis, then choosing
TICKVALUEFORMAT=YEAR. would result in several ticks having the same
year value.

DATA
specifies that the SAS date, time, or datetime format associated with the
data column assigned to the axis be used to control how the major tick
values are displayed.

Default The default format used by the INTERVAL= option. The default
does not apply if TICKVALUELIST= is specified.
 Axis Options for LAYOUT OVERLAY3D 1139

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range, then they are used.

Requirement The tick values must be specified as a space-separated list of

values enclosed in parentheses. The items in the list must be in the
same duration units as the data mapped to the axis: TIME, DATE,
or DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format
specified on the TICKVALUEFORMAT= option. If
TICKVALUEFORMAT= is not used, then the values are formatted
according to the column format (the default TICKVALUEFORMAT
value is not applied to these values).

If this option is specified, then the INTERVAL= option is ignored.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, MINORTICKS=TRUE is ignored, and minor tick marks are
not displayed.

Details
The LAYOUT OVERLAY3D statement provides XAXISOPTS=, YAXISOPTS=, and
ZAXISOPTS= options that enable you to manage the axis display separately for the
X, Y, and Z axes. The following example template uses the YAXISOPTS= option to
manage the display of grid lines, tick marks, and tick values on a Y axis:
begingraph;
layout overlay3d /
yaxisopts=(
griddisplay=on
display=(ticks tickvalues)
);
bihistogram3dparm x=rater y=customer z=percent;
endlayout;
endgraph;
 1140 Chapter 8 / Axis Options in Layouts

Within an OVERLAY3D layout block, each plot axis is always either a linear or a
time axis. The default axis-type setting is AUTO, which specifies that the axis type
be automatically determined by the plot or the overlay contents. The TYPE= option
enables you to explicitly specify either a LINEAR or a TIME axis type. When you
override the default axis type, you must be sure to specify the correct axis type for
the plot(s) that you are defining.

Each axis type has features specific to that type, and the axis options
LINEAROPTS= and TIMEOPS= enable you to specify features for a linear or a time
axis. You can combine one or more general axis options with the options for the
specific axis type. However, specified settings are applied only to the axis type that
supports them. For example, if you specify general axis options with time axis
options and the generated graph produces a linear axis type, then the time axis
options are ignored.

Row and Column Axis Options for LAYOUT

LATTICE
Row and column axis options for the plots within a LATTICE layout.

See: “LAYOUT LATTICE Statement” on page 126

Syntax
Axis options for the plots in a LATTICE layout are specified with the following
statements in a LAYOUT LATTICE statement block:

COLUMNAXIS /column-axis-options

ROWAXIS /row-axis-options

General Options for All Row and Column Axes in a Lattice

The options that are documented in this section can be used with any of the row or
column axis types that are supported in a LAYOUT LATTICE statement.
Subsequent sections in the chapter document the axis options that are available
only for specific axis types: discrete, linear, log, or time axes.

Statement Option Description

DISCRETEOPTS Specifies features for a discrete axis.

DISPLAY Controls which axis features are displayed on the

primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on the

secondary axis.
 Row and Column Axis Options for LAYOUT LATTICE 1141

Statement Option Description

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the axis

label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should be

included in the axis labels that are displayed.

LABELSPLITJUSTIFY Specifies the justification of the strings that are

inside the axis label blocks.

LINEAROPTS Specifies features for a standard numeric interval

axis.

LOGOPTS Specifies features for a log axis.

NAME Assigns a name to an axis for reference in other

statements.

OFFSETMAX Reserves an area at the maximum end of the axis.

No tick marks are displayed in the reserved area.

OFFSETMIN Reserves an area at the minimum end of the axis.

No tick marks are displayed in the reserved area.

REVERSE Specifies whether the axis origin should be

reversed.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for the
axis length.

TICKVALUEATTRS Specifies the color and font attributes of the axis

tick values.
1142 Chapter 8 / Axis Options in Layouts

Statement Option Description

TICKVALUEHALIGN Specifies the horizontal alignment for all of the

tick values that are displayed on the Y and Y2
axes.

TICKVALUEVALIGN Specifies the vertical alignment for all of the tick

values that are displayed on the X and X2 axes.

TIMEOPTS Specifies features for a TIME axis.

TYPE Specifies the type of axis to use.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not DISCRETE.

See “Options for Row and Column Discrete Axes Only” on page 1153 for
the options that you can use for discrete-axis-options.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default STANDARD
 Row and Column Axis Options for LAYOUT LATTICE 1143

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall
chart can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

See “Details” on page 1193 for more information about the primary and
secondary axes.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. A secondary
axis is not an independent axis. Rather, it mirrors the primary axis. Thus, for this
option to take effect, all plot statements in the layout must map data to the
same primary axis. For example, a secondary X2 axis can be displayed on top in
the layout, provided all plot statements set XAXIS=X to map data to the
primary X axis (bottom). Similarly, a secondary Y2 axis can be displayed to the
right in the layout, provided all plot statements set YAXIS=Y to map data to the
primary Y axis (left).

NONE
specifies that no axis features are displayed

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis

ALL
currently the same as STANDARD

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2,
both the X and X2 axis are primary and a secondary X axis cannot
1144 Chapter 8 / Axis Options in Layouts

be displayed. In that case, this option is ignored. The same applies

for the Y axes.

Interactions This option is ignored if the COLUMNAXIS statement appears

within a COLUMNAXES block and COLUMN2DATARANGE=UNION
or UNIONALL is in effect.

This option is ignored if the COLUMNAXIS statement appears

within a COLUMN2AXES block and COLUMNDATARANGE=UNION
or UNIONALL is in effect.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.

See “Details” on page 1193 for more information about the primary and
secondary axes.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid
lines only. It does not affect the appearance of the minor grid lines.
To control the appearance of the minor grid lines on a log axis, use
the MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template
to absolutely control the display of grid lines or to allow interaction with the
current style to decide whether grid lines are displayed. When displayed, the
grids appear in all cells.

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element
in the current style contains DisplayOpts="ON."

AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in
the current style contains DisplayOpts="OFF."

ON
specifies that grid lines are always displayed.

The current style has no override.

OFF
specifies that grid lines are never displayed.
 Row and Column Axis Options for LAYOUT LATTICE 1145

The current style has no override.

The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines
style element. Most supplied templates use the default setting AUTO_OFF to
indicate a preference for not displaying grid lines, but allowing the style to
override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style
has no preference about grid lines and the graphics template setting for
grid lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The
list form implies that all included string literals or dynamic variables will be
concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 1046.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated
by default.

Tips Use the SHORTLABEL= option to specify an alternate axis label to

be used whenever truncation would normally occur.
1146 Chapter 8 / Axis Options in Layouts

Use the ODS escape sequence to specify Unicode characters in an

axis label. Here is an example that species degrees Fahrenheit in an
axis label for temperature:
label="Temperature ((*ESC*){unicode '00B0'x}F)"

See “Overriding the Default Axis Label” in SAS Graph Template Language:
User’s Guide

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELFITPOLICY=AUTO | SPLITALWAYS | STACKED

specifies a policy for fitting axis labels in the available space.

AUTO
uses the short label, when specified, instead of the original label.

If the short label does not fit, then it is clipped. When no short label is
specified, the original label is clipped.

SPLIT
splits the axis label at a split character only when necessary in order to make
the label fit the available space.

The split character is specified by the LABELSPLITCHAR= option. The short

label is not used. A split does not occur at a split character if a split is not
needed at that location. If the label does not contain any of the specified
split characters, then it is not split. A label that cannot be split or that does
not fit the available space even after splitting might overlap the adjoining
space.

SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option.

If the label cannot be split, then it is clipped.

STACKED
displays the Y or Y2 axis label vertically as stacked letters.

Note: This feature applies to SAS 9.4M5 and to later releases.

Valid in ROWAXIS= only

 Row and Column Axis Options for LAYOUT LATTICE 1147

Default AUTO

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

Note When LABELPOSITION=CENTER, the available area is the full axis,

including the axis offsets. When LABELPOSITION=DATACENTER,
the available area is the tick display area, excluding the axis offsets.

LABELPOSITION=CENTER | DATACENTER
specifies the position of the axis label.

CENTER
centers each row or column axis label in its axis area.

For the row Y and Y2 axes, the label is oriented vertically and is centered in
the axis area (including the offsets) of its row. It is positioned to the left of
the tick values for the Y axis or to the right of the axis values for the Y2 axis.
For the column X and X2 axes, the label is centered in the axis area
(including the offsets) of its column. It is positioned below the tick values for
the X axis or above the axis values for the X2 axis.

DATACENTER
centers each row or column axis label in its axis tick display area.

For the row Y and Y2 axes, each label is oriented vertically and is centered in
the axis tick display area (excluding the offsets) of its row. The labels are
positioned to the left of the tick values for the Y axis or to the right of the
axis values for the Y2 axis. For the column X and X2 axes, each label is
centered in the axis tick display area (excluding the offsets) of its column.
The labels are positioned below the tick values for the X axis or above the
axis values for the X2 axis.

The following figure shows the CENTER and DATACENTER positions for row Y-
axis labels Open and Close, and column X-axis label Month. An offset is applied
to the maximum end of each axis for demonstration purposes.

Default CENTER
1148 Chapter 8 / Axis Options in Layouts

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the axis label. In that case, all of the specified split characters
together are treated as a single split character.

When LABELFITPOLICY=SPLIT, if the axis label does not fit the available
space, then it is split on a specified split character only if a split is needed at
that point to make the label fit. In this case, a split might not occur on every
split character. When LABELFITPOLICY=SPLITALWAYS, the axis label is split
unconditionally on every occurrence of a split character. If the axis label does
not contain any of the specified split characters, the label is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
labelsplitchar="abc"

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.

TRUE
drops the split characters from the axis label display.

FALSE
includes the split characters in the axis label display.

When the label is split while LABELSPLITCHARDROP=FALSE is in effect,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.
 Row and Column Axis Options for LAYOUT LATTICE 1149

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.

justification

CENTER
LEFT
RIGHT
specifies the justification for the X or X2 axis label.

CENTER
TOP
BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be
enclosed in parentheses. Each option is specified as a name = value pair and
each pair is space separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Row and Column Linear Axes Only” on page 1162 for
the options that you can use for linear-axis-options.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not LOG.

See “Options for Row and Column Log Axes Only” on page 1174 for the
options that you can use for log-axis-options.

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used
only in an AXISLEGEND statement.

Restriction This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.
1150 Chapter 8 / Axis Options in Layouts

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or EXTRACTALWAYS.
In addition, an AXISLEGEND statement must be specified to
generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.
 Row and Column Axis Options for LAYOUT LATTICE 1151

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the specified label is itself too long for the axis, it is truncated in
the display.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on
the Y and Y2 axes.

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on
the X and X2 axes.
1152 Chapter 8 / Axis Options in Layouts

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time,
date, or datetime format.

Options must be enclosed in parentheses. Each option is specified

as a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

See “Options for Row and Column Time Axes Only” on page 1184 for
the options that you can use for time-axis-options.

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.

AUTO
requests that the axis type be automatically determined, based on the
overlay contents.

DISCRETE
uses a DISCRETE axis if possible.

The data for discrete axes can be character or numeric. You can add a
DISCRETEOPTS= option list to customize this axis type.

LINEAR
uses a LINEAR axis if possible.

You can add a LINEAROPTS= option list to customize this axis type.

TIME
uses a TIME axis if possible.
 Row and Column Axis Options for LAYOUT LATTICE 1153

Data for this axis must be SAS time, SAS date, or SAS datetime values. You
can add a TIMEOPTS= option list to customize this axis type.

LOG
uses a LOG axis if possible.

You can add a LOGOPTS= option list to customize this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur
for the response axis of a bar chart, line chart, needle plot, or
waterfall chart when a baseline intercept of 0 or less is specified.
It can also occur for the response axis of a waterfall chart when
an initial bar value of 0 or less is specified. To get a log response
axis in those cases, set the baseline intercept or initial bar value
to a positive value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within
the layout are dropped from the display if their data types or data
ranges do not match the axis type requirements. For more
information, see “Plot Axis Types Must Agree on Common Axes” on
page 1049.

After the axis type is determined (whether you set a specific type
or AUTO is in effect), you can use only options that are supported
by that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.

Options for Row and Column Discrete Axes Only

The options that are documented in this section can be used with the
DISCRETEOPTS= axis option. The following table provides a summary of the
options.

Discrete Axis Option Description

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on an axis.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.
1154 Chapter 8 / Axis Options in Layouts

Discrete Axis Option Description

TICKVALUELIST Specifies the text that is to be displayed for the

tick values that are defined in the
TICKVALUELIST= option.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESPLITCHAR Specifies a list of characters on which the tick

values can be split, if needed.

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are inside

the tick value block.

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

Example The following example specifies the axis tick values 10, 20, 30,
and 40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT
places the tick marks at the midpoint value location.

INBETWEEN
places the tick marks half way between adjacent midpoint locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

 Row and Column Axis Options for LAYOUT LATTICE 1155

Note Starting with SAS 9.4M2, when TICKTYPE=INBETWEEN, the

outermost tick marks and grid lines at each end of the axis are not
drawn.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y and Y2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space.

A split does not occur at a split character if a split is not needed at that
location. If the value does not contain any of the specified split characters,
then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR= on page 1159

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.
1156 Chapter 8 / Axis Options in Layouts

See TICKVALUESPLITCHAR= on page

1159

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words
do not fit the available space.

SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit
the available space.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.
 Row and Column Axis Options for LAYOUT LATTICE 1157

ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space.

A split does not occur at a split character if a split is not needed at that
location. If the value does not contain any of the specified split characters,
then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR= on page 1159

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR= on page 1159

SPLITROTATE
attempts the SPLIT policy, and then the ROTATE policy.

STACKEDALWAYS
always displays the tick values vertically as stacked letters.

Valid in COLUMNAXIS= only

STACKEDALWAYSTHIN
always displays the tick values vertically as stacked letters.

Thinning is performed when the tick values do not fit the available space.

Valid in COLUMNAXIS= only

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.

THIN
eliminates alternate tick values.

TRUNCATE
shortens the tick labels that exceed 12 characters to 12 characters.
1158 Chapter 8 / Axis Options in Layouts

A trailing ellipsis indicates that a label is truncated.

Note Data labels might overlap on a crowded axis.

TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.

TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.

TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults ROTATE for the Y and Y2 axes

THIN for the X and X2 axes

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.

string-list
a space-separated list of values, enclosed in parentheses.

You must enclose each value in the list in quotation marks.

Only the tick values that are included in the string list are displayed on the axis.
The values are displayed in the order in which they are listed. The data values
that are not in the list are dropped. The list can be a subset of the data values. It
can also contain values that are not included in the actual data. A tick value that
is not included in the data appears on the axis, but no data is represented at its
tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated

from adjacent values by a blank space.
 Row and Column Axis Options for LAYOUT LATTICE 1159

Notes If the string list contains duplicate values, then the first
occurrence of the duplicated value in the list is honored and the
remaining instances are ignored.

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan,
Sports, Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30,
and 40:
tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed.
When multiple characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in
1160 Chapter 8 / Axis Options in Layouts

the tick value. In that case, all of the specified split characters together are
treated as a single split character.

When a tick value collision is detected, the tick value is split at each occurrence
of any of the characters in the character list or all. If all of the split characters
occur consecutively in the tick value, then they are treated as a single split
character. If the tick value does not contain any of the specified characters, then
the value is not split.

"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used to

manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether the

split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick
values. The split characters are specified by the TICKVALUESPLITCHAR=
option.

TRUE
drops the split characters from the tick value display.
 Row and Column Axis Options for LAYOUT LATTICE 1161

The following figure shows an example in which

TICKVALUESPLITCHARDROP=TRUE and three-word, asterisk-delimited
tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display.

The fit policy determines how the characters are displayed. If the display
policy is SPLIT or SPLITTHIN and TICKVALUESPLITCHARDROP=FALSE,
then each tick value is split at a split character only where a split is
necessary in order to make the value fit the available space. A split might not
occur at every split character in the tick value. At each split point, the split
character remains as the last character in the current line. The characters
that follow the split character, up to and including the split character at the
next split point, are then wrapped to the following line. This process repeats
until the entire data tick value is displayed. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this
case, a split is not needed at the second asterisk.

If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and

TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line. The
following figure shows an example in which
TICKVALUESPLITCHARDROP=FALSE and three-word, asterisk-delimited
tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that a split occurs after each asterisk and each asterisk appears at
the end of the line. In this case, three lines are displayed.

Default TRUE
1162 Chapter 8 / Axis Options in Layouts

Interactions The TICKVALUESPLITCHAR= option specifies the split character or

characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not
affect the display of tick values that are not split.

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Row and Column Linear Axes Only

The options that are documented in this section can be used with the
LINEAROPTS= axis option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

 Row and Column Axis Options for LAYOUT LATTICE 1163

Linear Axis Option Description

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on an X or X2 axis. This option is not available in
the ROWAXIS statement.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a linear

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by start,

end, and increment.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.
1164 Chapter 8 / Axis Options in Layouts

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the TICKVALUEFORMAT=
option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

Note This option is ignored when the data exceeds the limits of integer
values.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)
 Row and Column Axis Options for LAYOUT LATTICE 1165

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in SAS 9.4M1 and earlier releases.

One tick with two intervals starting with SAS 9.4M2.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.
1166 Chapter 8 / Axis Options in Layouts

Default FALSE

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

 Row and Column Axis Options for LAYOUT LATTICE 1167

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option is

ignored.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y and Y2 axes, the following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

1168 Chapter 8 / Axis Options in Layouts

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.

(format-options)
specifies one or more formatting options for major tick values.

Together, these options provide parameters for determining an optimal

format (w.d, Ew., BESTw.) for displaying major tick values.

MAXWIDTH=integer
specifies the maximum width for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 6
 Row and Column Axis Options for LAYOUT LATTICE 1169

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values.

The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.

The scale can be a named scale or a scientific-notation scale. The

EXTRACTSCALETYPE= option specifies the scale type. The scale that is
used is appended to the axis label, as shown in the following example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the
label is truncated, and the scale is appended to the truncated label.
Ellipses indicate that the label was truncated, as shown in the following
example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis.
In that case, the tick value format might differ from the data
format even when a named format is applied to the data
values.

See “boolean ” on page xii for other Boolean values that you can
use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.
1170 Chapter 8 / Axis Options in Layouts

A named scale can be millions, billions, or trillions for values of 999

trillion or less, or a multiple of 10 (denoted as 10^n) for values over
999 trillion. For large tick values, the scale factor is set to ensure that
the absolute value of the largest value is greater than 1. For small
fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for

values greater than 1, or a multiple of 1/10 expressed as 10^–n for
values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621 .

Note If you specify a format that significantly reduces precision,

because of tick-value rounding, the plot data elements might not
align properly with the axis tick values. In that case, specify a
tick-value format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option
is used, the only tick values that appear are the tick values in
numeric-list that fall within the explicit data range (set by
VIEWMIN= and VIEWMAX=) or the implicit data range (set by the
actual data minimum and data maximum).

Restriction This option applies to linear axes only.

 Row and Column Axis Options for LAYOUT LATTICE 1171

Requirement The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the DISPLAYSECONDARY=
option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some data
not being displayed. For example, data might not be displayed
when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data
range of the axis, but cannot reduce it.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Restriction This option applies to linear axes only.

1172 Chapter 8 / Axis Options in Layouts

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(START=number END=number INCREMENT=number)

specifies the tick values by start, end, and increment.

START=number
specifies the value for the first tick mark.

END=number
specifies the value for the last tick mark.

INCREMENT=number
specifies the increment for intermediate tick marks between the first and
last tick marks.
 Row and Column Axis Options for LAYOUT LATTICE 1173

The END value always controls the last tick mark. The interval between the
last tick mark and the previous tick mark might not necessarily be the
INCREMENT value.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option is
used, the only tick values that appear are those that fall within the
explicit data range (set by VIEWMIN= and VIEWMAX=) or the
implicit data range (set by the actual data minimum and data
maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting
for the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the maximum axis tick value that
is displayed. The THRESHOLDMAX= value is used to determine
the maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
1174 Chapter 8 / Axis Options in Layouts

factors such as thresholds and the tick values computed by the

plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the minimum axis tick value that
is displayed. The THRESHOLDMIN= value is used to determine
the minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Options for Row and Column Log Axes Only

The options that are described in this section can be used with the LOGOPTS= axis
option. The following table provides a summary of the options.

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for the

axis values.

MINORGRID Specifies whether grid lines are displayed at the

minor tick marks.
 Row and Column Axis Options for LAYOUT LATTICE 1175

Log Axis Option Description

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

TICKINTERVALSTYLE Specifies how to scale and format the values for

major tick marks.

TICKVALUEFORMAT= Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the tick values for a log axis as a space-

separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=, VIEWMAX=,
and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
1176 Chapter 8 / Axis Options in Layouts

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.
 Row and Column Axis Options for LAYOUT LATTICE 1177

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Requirement If you set this option to TRUE, you must also specify a value for
MINORTICKCOUNT=. Otherwise, minor tick marks do not appear
on the log axis.
1178 Chapter 8 / Axis Options in Layouts

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform log tick values such as 10, 20, 30, and 40, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform log tick values such as 1, 10, 100, and 1000 in
the TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.

AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data.

When the data range is small (within an order of magnitude), a LINEAR

representation is typically used. Data ranges that encompass several orders
of magnitude typically use the LOGEXPAND or LOGEXPONENT
representation.

LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are only the integer exponents for all bases.

LINEAR
places the major tick marks at non-uniform intervals that cover the range of
the data.

Default AUTO
 Row and Column Axis Options for LAYOUT LATTICE 1179

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format"
on the column is retained. For LINEAR, ticks values are
automatically formatted when the column format is not assigned or
one of w.d, Ew., or BESTw. Other formats (SAS defined or user-
defined) are used if specified.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than
or equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you might

want to include information in the axis label about which base is
used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.

When TICKINTERVALSTYLE=LOGEXPAND, this option is honored

for the base 10 and base 2 logarithmic scales, and is ignored for the
base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=
1180 Chapter 8 / Axis Options in Layouts

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the
explicit data range set by the VIEWMIN= and VIEWMAX= options
or by the implicit data range set by the actual data minimum and
data maximum are displayed. An internal algorithm determines
the tick marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results
might occur. If VALUESTYPE=EXPANDED is in effect (default),
specify increments of the log base power such as 0.1, 1, 10, 100,
and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify integer increments
of the log base power exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some
data not being displayed. For example, data might not be
displayed when the VIEWMIN= value is greater than the actual
data minimum, or when the VIEWMAX= value is less than actual
data maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some
data might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

If MINORTICKS=TRUE is specified and TICKVALUELIST=

specifies nonuniform log tick values such as 10, 20, 30, and 40,
then MINORTICK=TRUE is ignored. To display minor tick marks in
that case, specify uniform log tick values such as 1, 10, 100, and
1000 in the TICKVALUELIST= option.

See TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

 Row and Column Axis Options for LAYOUT LATTICE 1181

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the
TICKVALUELIST=, VIEWMAX=, and VIEWMIN= option values. Use this option
to choose your preferred way of specifying log-axis values.

EXPANDED
values are interpreted as powers of the base such as 0.1, 1, 10, 100, and so on,
for base 10, for example.

EXPONENT
values are interpreted as integer exponents of the base such as 1, 2, 3, and so
on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base
1182 Chapter 8 / Axis Options in Layouts

10 log axis. This results in X-axis tick values of 10, 100, 1000, 10000,
and 100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10
log axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so
on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
default value for VIEWMAX= is less than the value specified by
the VIEWMIN= option, then the VIEWMIN= and VIEWMAX=
values are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
 Row and Column Axis Options for LAYOUT LATTICE 1183

factors such as thresholds and the tick values computed by the

plot statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 100,000 as an

expanded value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an

exponent value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so
on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.
1184 Chapter 8 / Axis Options in Layouts

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 0.1 as an expanded

value on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent

value on a base 10 log axis:
VIEWMIN=-1

Options for Row and Column Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS=
axis option. The following table provides a summary of the options.

Time Axis Option Description

INTERVAL Specifies the time interval between major tick

marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time interval

that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor ticks.

MINORTICKS Specifies whether minor tick marks are displayed.

SPLITTICKVALUE Specifies whether to split the tick values on

column axes, if possible. This option is not
available in the ROWAXIS statement.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on column axes. This option is not available in the
ROWAXIS statement.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.
 Row and Column Axis Options for LAYOUT LATTICE 1185

Time Axis Option Description

TICKVALUELIST Specifies the order of the tick values for a time

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

INTERVAL=AUTO | SECOND | MINUTE | HOUR | DAY | TENDAY | WEEK |

SEMIMONTH | MONTH | QUARTER | SEMIYEAR | YEAR
specifies the time interval between major ticks.
Table 8.3 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of DATE9.

DATETIME each month
1186 Chapter 8 / Axis Options in Layouts

Default tick value

INTERVAL Unit Tick interval format

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if
the data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
 Row and Column Axis Options for LAYOUT LATTICE 1187

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.
1188 Chapter 8 / Axis Options in Layouts

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.12 on page 1185 for
information about the intervals that you can select. The interval that you select
must be consistent with the axis data duration units such as TIME, DATE, or
DATETIME. For example, if the axis data is in TIME units, then you must specify
AUTO, SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on
the value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, then MINORTICKS=TRUE is ignored, and minor tick marks
are not displayed.
 Row and Column Axis Options for LAYOUT LATTICE 1189

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This
option is not available for a Y or Y2 axis.

TRUE
splits the axis tick values into two lines allowing more tick values to appear.

For example, with INTERVAL= MONTH, this is how tick values are split:

FALSE
does not split the axis tick values.

For example, when this option specifies FALSE, this is how the tick values in
the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option
is not available for the X and X2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the
axis. The following policies are valid for the X and X2 axes:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.
1190 Chapter 8 / Axis Options in Layouts

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= on page 1189TRUE, this option is ignored

and only the THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.

format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed.

This format must be in the same duration units as the data column(s)
mapped to a time axis: TIME, DATE, or DATETIME and should be appropriate
for the value of the INTERVAL= option. For example, if INTERVAL=MONTH
and there are two years of data displayed on the axis, then choosing
TICKVALUEFORMAT=YEAR. would result in several ticks having the same
year value.

DATA
specifies that the SAS date, time, or datetime format associated with the
data column assigned to the axis be used to control how the major tick
values are displayed.

Default The default format used by the INTERVAL= option. The default
does not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.
 Row and Column Axis Options for LAYOUT LATTICE 1191

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
default data range or that specified by the VIEWMIN= or
VIEWMAX= options, then they are used.

Requirement The tick values must be specified as a space-separated list of

values enclosed in parentheses. The items in the list must be in the
same duration units as the data mapped to the axis: TIME, DATE,
or DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format
specified on the TICKVALUEFORMAT= option. If
TICKVALUEFORMAT= is not used, then the values are formatted
according to the column format (the default TICKVALUEFORMAT
value is not applied to these values).

If this option is specified, the SPLITTICKVALUE= and INTERVAL=

options are ignored.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, MINORTICKS=TRUE is ignored, and minor tick marks are
not displayed.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.
1192 Chapter 8 / Axis Options in Layouts

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXIS statements only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.
 Row and Column Axis Options for LAYOUT LATTICE 1193

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Details
The LAYOUT LATTICE statement creates a grid of graphs that automatically aligns
plot areas, data display areas, axis labels, and headers across the columns and rows
of the layout. The axis data ranges can be scaled, and the axes for individual cells in
the layout can be managed by row and by column using COLUMNAXIS and
ROWAXIS statements. COLUMNAXIS statements are used within a COLUMNAXES
or COLUMN2AXES block to externalize column axes for the layout. Similarly,
ROWAXIS statements are used within a ROWAXES or ROW2AXES block to
externalize row axes for the layout. Each axis block is used to manage the primary
axis. The axis that is considered primary depends on the settings for the XAXIS=
and YAXIS= options in plot statements that are specified within the layout:
1194 Chapter 8 / Axis Options in Layouts

Option Setting Primary Axis Axis Block to Use

XAXIS=X X (bottom) COLUMNAXES

XAXIS=X2 X2 (top) COLUMN2AXES

YAXIS=Y Y (left) ROWAXES

YAXIS=Y2 Y2 (right) ROW2AXES

For the specifications to take effect,

n UNION or UNIONALL data scaling must be set for the affected columns and
rows. The data scaling is set with the LAYOUT LATTICE statement’s
COLUMNDATARANGE=, COLUMN2DATARANGE=, ROWDATARANGE=, and
ROW2DATARANGE= options.
n Within a COLUMNAXES or COLUMN2AXES block, one COLUMNAXIS
statement should be specified for each column that contains axes that you need
to manage. Both axes blocks can contain a COLUMNAXIS statement for the
same column. For example, to manage the axes in the first column of the layout,
the COLUMNAXES block can contain a COLUMNAXIS statement that manages
the column’s X axes. The COLUMN2AXES block can contain a COLUMNAXIS
statement that manages the column’s X2 axes.
n Within a ROWAXES or ROW2AXES block, one ROWAXIS statement should be
specified for each row that contains axes that you need to manage. Both axes
blocks can contain a ROWAXIS statement for the same row. For example, to
manage the axes in the first row of the layout, the ROWAXES block can contain
a ROWAXIS statement that manages the row’s Y axes. The ROW2AXES block
can contain a ROWAXIS statement that manages the column’s Y2 axes.

In addition to managing the primary axes, you can also display “secondary” axes in
the grid. A secondary axis is not an independent axis. Rather, it mirrors the primary
axis, but it is displayed on the opposite side and can have different display options.
For example, when the X axis (bottom) is primary, you can mirror that axis with a
secondary X axis at the top of the grid. Similarly, when the Y2 axis (right) is primary,
you can mirror that axis with a secondary Y2 axis on the left of the grid. A
secondary axis makes it easier to interpolate values in the cells that are farthest
away from the primary axis. To display a secondary axis, use the
DISPLAYSECONDARY= option.

For general information about managing primary and secondary axes, see “Plot
Data Are Mapped to a Designated Axis” on page 1040. For details about managing
the axes within a LAYOUT LATTICE, see the discussion about the LAYOUT
LATTICE “Row and Column Axis Statements” on page 146.

The following example shows COLUMNAXIS statements for a lattice with two
columns:
layout lattice / columns=2 columndatarange=union;
columnaxes;
columnaxis / griddisplay=on displaysecondary=(ticks tickvalues);
columnaxis / griddisplay=on displaysecondary=(ticks tickvalues);
 Row and Column Axis Options for LAYOUT LATTICE 1195

endcolumnaxes;

/* rest of lattice definition */

endlayout;

COLUMNAXIS and ROWAXIS statements are similar to the XAXISOPTS= and

YAXISOPTS= options for LAYOUT OVERLAY, with the following differences:
n When COLUMNAXIS and ROWAXIS are used, any axis options specified on
plots within the affected columns or rows are ignored. All axis features for the
row and column axes must be specified on the ROWAXIS or COLUMNAXIS
statement.
n When ROWAXIS or COLUMNAXIS are used, any LAYOUT OVERLAYEQUATED
layouts specified for cells in the affected columns or rows are implemented as
LAYOUT OVERLAY layouts. Equated axes are not supported on row and column
axes.

In the default cases for each plot in the layout, the axis type is always DISCRETE,
LINEAR, or TIME. The TYPE= option enables you to specify an axis type that
overrides the default. For example, when appropriate for the data, you can request
a LOG axis. When you override the default axis type, you must be sure to specify
the correct axis type for the plot(s) that you are defining.

Each axis type has features specific to that type, and the following axis options
enable you to specify features for the different types: DISCRETEOPTS= ,
LINEAROPTS= , LOGOPTS= , and TIMEOPTS= . One or more of these options can
be specified for an axis, but the specified settings are applied only to the axis type
that supports them.

Example: Row and Column Axis Options for

LAYOUT LATTICE
This example shows how to externalize axes in a LAYOUT LATTICE and manage
the axis features on primary Y and Y2 axes.
n The first HISTOGRAM statement specifies YAXIS=Y2 to make the Y2 axis the
primary axis for COUNT measures. The second HISTOGRAM statement
specifies YAXIS=Y to make the Y axis the primary axis for PERCENT measures.
n In order to externalize the axes within the layout, the data ranges for the axes
must be unified. In the LAYOUT LATTICE statement, the ROWDATARANGE=
option unifies the data ranges for the Y axes across the row. The
ROW2DATARANGE= option unifies the data ranges for the Y2 axes across the
row.
n The ROWAXIS statement is used to manage axis features for the row axes. To
manage the primary Y axis, a ROWAXIS statement is specified within a
ROWAXES block. To manage the primary Y2 axis, another ROWAXIS statement
is specified within a ROW2AXES block.
n Within the ROWAXES block, the ROWAXIS statement consolidates Y axes in
the row into a single, column Y axis and also displays grid lines. Within the
 1196 Chapter 8 / Axis Options in Layouts

ROW2AXES block, the ROWAXIS statement consolidates Y2 axes in the row

into a single row Y2 axis, but it does not alter the default features of that axis.

The following graph was generated by the “Example Program” on page 1196 :

Example Program
proc template;
define statgraph y2axis;
begingraph;
layout lattice / columns=2 columngutter=10
rowdatarange=union row2datarange=union ;
rowaxes;
rowaxis / griddisplay=on;
endrowaxes;
row2axes;
rowaxis;
endrow2axes;
layout overlay;
histogram height / scale=count yaxis=y2 ;
histogram height / scale=percent yaxis=y ;
densityplot height / normal();
endlayout;
layout overlay;
histogram weight / scale=count yaxis=y2 ;
histogram weight / scale=percent yaxis=y ;
densityplot weight / normal();
endlayout;
endlayout;
endgraph;
 Axis Options for LAYOUT OVERLAYEQUATED 1197

end;

proc sgrender data=sashelp.class template=y2axis;

run;

Axis Options for LAYOUT OVERLAYEQUATED

Axis options for the plots within an OVERLAYEQUATED layout.

Interaction: The OVERLAYEQUATED’s axis options are ignored when the LAYOUT
OVERLAYEQUATED statement is nested within another layout type that has
external axes in effect. For example, the axis options are ignored when the
statement is nested within a LAYOUT LATTICE with a COLUMNAXIS= or
ROWAXIS= option in effect.
See: “LAYOUT OVERLAYEQUATED Statement” on page 168

Syntax
Axis options for the plots within an OVERLAYEQUATED layout are specified with
the
following options on a LAYOUT OVERLAYEQUATED statement:

COMMONAXISOPTS=(common-equated-axis-options)

XAXISOPTS=(equated-axis-options)

YAXISOPTS=(equated-axis-options)

Options That Apply to Both Equated Axes

The options that are documented in this section are specified with the
COMMONAXISOPTS= option and are applied to both the X and Y axes. With the
exception of VIEWMAX and VIEWMIN, these options cannot be applied separately
to an X or Y axis using the XAXISOPTS= or YAXISOPTS= option. See “Options That
Apply Separately to an X or Y Equated Axis” for a list of options that can be applied
to a single axis. The following table provides a summary of the common options.

Equated Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks for all axes.

TICKSTYLE Specifies the placement of tick marks in relation

to the axis line.

TICKVALUELIST Specifies the order of the tick values as list.

1198 Chapter 8 / Axis Options in Layouts

Equated Axis Option Description

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUESEQUENCE Specifies the tick values by start, end, and

increment.

VIEWMAX Specifies the maximum data value to include in

the display on the X and Y axes when the axis
lengths and major tick values are equal.

VIEWMIN Specifies the minimum data value to include in the

display on the X and Y axes when the axis lengths
and major tick values are equal.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the TICKVALUEFORMAT=
option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAM or HISTOGRAMPARM statement.

Note This option is ignored when the data exceeds the limits of integer
values.

See “boolean ” on page xii for other Boolean values that you can use.

TICKSTYLE=OUTSIDE | INSIDE | ACROSS

specifies the placement of tick marks in relation to the axis line. The figure
shows the tick display for each value.

OUTSIDE
displays tick marks outside of the axis frame.
 Axis Options for LAYOUT OVERLAYEQUATED 1199

INSIDE
displays tick marks inside the axis frame.

ACROSS
displays tick marks across the axis line.

Default The GraphAxisLines:TickDisplay style reference.

Restriction This option applies to the major tick marks only. It does not affect
the minor tick marks.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick marks.

Note This option has no effect on the placement of the tick values, which
are always outside the axis frame.

Tip You can use the Annotate facility to draw inside or across minor tick
marks in your graphs. See “Creating a Macro That Generates
Annotation Data” in SAS Graph Template Language: User’s Guide.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option
is used, the only tick values that appear are the tick values in
numeric-list that fall within the explicit data range (set by
VIEWMIN= and VIEWMAX=) or the implicit data range (set by the
actual data minimum and data maximum).

Requirement The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the LAYOUT OVERLAYEQUATED

statement specifies EQUATETYPE=FIT (the default).

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the DISPLAYSECONDARY=
option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some data
not being displayed. For example, data might not be displayed
when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
1200 Chapter 8 / Axis Options in Layouts

numeric-list. The tick numeric-list can extend the implicit data

range of the axis, but cannot reduce it.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Note Unexpected results can occur when TYPE= in the LAYOUT

OVERLAYEQUATED statement is set to anything other than
SQUARE and the tick values in TICKVALUELIST= are not suitable
for the X and Y data ranges. In that case, specify suitable tick
values for X and Y in TICKVALUELIST= or specify TYPE=SQUARE
in the LAYOUT OVERLAYEQUATED statement.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Notes If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

Unexpected results can occur when TYPE= in the LAYOUT

OVERLAYEQUATED statement is set to anything other than
SQUARE and the tick values in TICKVALUELIST= or
TICKVALUESEQUENCE= are not suitable for the X and Y data
ranges. In that case, specify suitable tick values for X and Y in
 Axis Options for LAYOUT OVERLAYEQUATED 1201

TICKVALUELIST= or TICKVALUESEQUENCE=, or specify

TYPE=SQUARE in the LAYOUT OVERLAYEQUATED statement.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUESEQUENCE=(START=number END=number INCREMENT=number)

specifies the tick values by start, end, and increment.

START=number
specifies the value for the first tick mark.

END=number
specifies the value for the last tick mark.

INCREMENT=number
specifies the increment for intermediate tick marks between the first and
last tick marks.

The END value always controls the last tick mark. The interval between the
last tick mark and the previous tick mark might not necessarily be the
INCREMENT value.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option is
used, the only tick values that appear are those that fall within the
explicit data range (set by VIEWMIN= and VIEWMAX=) or the
implicit data range (set by the actual data minimum and data
maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the LAYOUT OVERLAYEQUATED

statement specifies EQUATETYPE=FIT (the default), or if the
DISPLAY= option or the DISPLAYSECONDARY= option does not
display tick marks.

Note Unexpected results can occur when TYPE= in the LAYOUT

OVERLAYEQUATED statement is set to anything other than
SQUARE and the tick values in TICKVALUESEQUENCE= are not
suitable for the X and Y data ranges. In that case, specify suitable
tick values for X and Y in TICKVALUESEQUENCE= or specify
TYPE=SQUARE in the LAYOUT OVERLAYEQUATED statement.

Tip The values in the sequence are formatted according to the setting
for the TICKVALUEFORMAT= option.
1202 Chapter 8 / Axis Options in Layouts

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display on the X and Y axes
when the axis lengths and major tick values are equal. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the X and Y axes.

Restriction This option is honored only when EQUATETYPE=SQUARE.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the maximum axis tick value that
is displayed. The THRESHOLDMAX= value is used to determine
the maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display on the X and Y axes
when the axis lengths and major tick values are equal. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the X and Y axes.

Restriction This option is honored only when EQUATETYPE=SQUARE.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.
 Axis Options for LAYOUT OVERLAYEQUATED 1203

Interactions This option does not determine the minimum axis tick value that
is displayed. The THRESHOLDMIN= value is used to determine
the minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Options That Apply Separately to an X or Y Equated Axis

The options that are documented in this section can be applied to an X axis with
the XAXISOPTS= option, or to the Y axis with the YAXISOPTS= option. See
“Options That Apply to Both Equated Axes” on page 1197 for a list of options that
apply in common to both axes. The following table provides a summary of the
options.

Equated Axis Option Description

DISPLAY Controls which axis features are displayed on the

primary axis.

DISPLAYSECONDARY Controls which axis features are displayed on the

secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies when axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the axis

label.

LABELFITPOLICY on page 1208 Specifies a policy for fitting axis labels in the
available space.

LABELSPLITCHAR on page 1208 Specifies one or more characters on which the

axis labels can be split, if needed.
1204 Chapter 8 / Axis Options in Layouts

Equated Axis Option Description

LABELSPLITCHARDROP on Specifies whether the split characters are

page 1209 included in the displayed axis labels.

LINEEXTENT Specifies the extent of the axis line.

OFFSETMAX Reserves an area at the maximum end of the axis.

No tick marks are displayed in the reserved area.

OFFSETMIN Reserves an area at the minimum end of the axis.

No tick marks are displayed in the reserved area.

REVERSE Specifies whether tick values should appear in the

reverse order.

SHORTLABEL Specifies an alternate axis label to use if the

default or specified axis label is too long for the
axis length.

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.

TICKVALUEATTRS Specifies the color and font attributes of the axis

tick values.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision.

Only the default policy (THIN) is available for a Y
or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for tick marks.

VIEWMAX Specifies the maximum data value to include in

the display on the X or Y axis when the axis
lengths and major tick values are not equal.

VIEWMIN Specifies the minimum data value to include in the

display on the X or Y axis when the axis lengths
and major tick values are not equal.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed
 Axis Options for LAYOUT OVERLAYEQUATED 1205

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default STANDARD

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall
chart can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on the secondary axis. When data are
mapped to the X or Y axis, you can display an X2 or Y2 (secondary) axis using
this option. The secondary axis is a duplicate of the X or Y axis but can have
different display options.

NONE
specifies that no axis features are displayed

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis

ALL
currently the same as STANDARD

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

1206 Chapter 8 / Axis Options in Layouts

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default NONE

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template
to absolutely control the display of grid lines or to allow interaction with the
current style to decide whether grid lines are displayed.

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element
in the current style contains DisplayOpts="ON."

AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in
the current style contains DisplayOpts="OFF."

ON
specifies that grid lines are always displayed.

The current style has no override.

OFF
specifies that grid lines are never displayed.

The current style has no override.

The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines
style element. Most supplied templates use the default setting AUTO_OFF to
indicate a preference for not displaying grid lines, but allowing the style to
override.
 Axis Options for LAYOUT OVERLAYEQUATED 1207

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style
has no preference about grid lines and the graphics template setting for
grid lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The
list form implies that all included string literals or dynamic variables will be
concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis” on
page 1046.

Note If the axis label is too long to fit along the axis, then it is truncated by
default.

Tips Use the SHORTLABEL= option to specify an alternate axis label to be

used whenever truncation would normally occur.

Use the ODS escape sequence to specify Unicode characters in an axis

label. Here is an example that species degrees Fahrenheit in an axis
label for temperature:
label="Temperature ((*ESC*){unicode '00B0'x}F)"

See “Overriding the Default Axis Label” in SAS Graph Template Language:
User’s Guide

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.
1208 Chapter 8 / Axis Options in Layouts

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELFITPOLICY=AUTO | SPLIT | SPLITALWAYS | STACKED

specifies a policy for fitting axis labels in the available space.

AUTO
uses the short label, when specified, instead of the original label.

If the short label does not fit, then it is clipped. When no short label is
specified, the original label is clipped.

SPLIT
splits the axis label at a split character only when necessary in order to make
the label fit the available space.

The split character is specified by the LABELSPLITCHAR= option. The short

label is not used. A split does not occur at a split character if a split is not
needed at that location. If the label does not contain any of the specified
split characters, then it is not split. A label that cannot be split or that does
not fit the available space even after splitting might overlap the adjoining
space.

SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option.

If the label cannot be split, then it is clipped.

STACKED
displays the Y or Y2 axis label vertically as stacked letters.

Note: This feature applies to SAS 9.4M5 and to later releases.

Valid in YAXISOPTS= and Y2AXISOPTS only

Default AUTO

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the axis label. In that case, all of the specified split characters
together are treated as a single split character.

When LABELFITPOLICY=SPLIT, if the axis label does not fit the available
space, then it is split on a specified split character only if a split is needed at
that point to make the label fit. In this case, a split might not occur on every
 Axis Options for LAYOUT OVERLAYEQUATED 1209

split character. When LABELFITPOLICY=SPLITALWAYS, the axis label is split

unconditionally on every occurrence of a split character. If the axis label does
not contain any of the specified split characters, the label is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
labelsplitchar="abc"

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.

TRUE
drops the split characters from the axis label display.

FALSE
includes the split characters in the axis label display.

When the label is split while LABELSPLITCHARDROP=FALSE is in effect,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

LINEEXTENT=FULL | DATA | number

specifies the extent of the axis line.
1210 Chapter 8 / Axis Options in Layouts

Note: This feature applies to SAS 9.4M3 and to later releases.

FULL
specifies an axis line that extends along the entire length of the axis.

DATA
specifies an axis line that extends through the data range from the minimum
offset to the maximum offset.

number
specifies how much the axis line extends from DATA toward FULL as a
decimal proportion. A value of 0 is equivalent to DATA, and a value of 1 is
equivalent to FULL.

Range 0–1

Tip A numeric value is useful for bar charts when DATA terminates the
axis line at the midpoint positions of the minimum and maximum
bars. In that case, you can specify a numeric value to lengthen the
axis line so that it extends to the full width of both bars.

The following figure shows a simple example of each value for the X axis and Y-
axis lines. The light-blue dashed lines depict the minimum and maximum offsets
that are set on the axes.

Default FULL

Restriction This option is valid only in OVERLAY and OVERLAYEQUATED

layouts.

Interactions This option overrides the AXISLINEEXTENT= option in the

BEGINGRAPH statement.

This option overrides the LINEEXTENT= option in

COMMONAXISOPTS=.

Tip The graph wall outline might appear to be an axis line. In that case,
use the WALLDISPLAY=NONE or WALLDISPLAY=(FILL) option in
the layout statement to suppress the wall outline.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in
the reserved area.
 Axis Options for LAYOUT OVERLAYEQUATED 1211

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

Default AUTO

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the specified label is itself too long for the axis, it is truncated in
the display.
1212 Chapter 8 / Axis Options in Layouts

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to the LAYOUT OVERLAYEQUATED

XAXISOPTS= and YAXISOPTS= options only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to the LAYOUT OVERLAYEQUATED

XAXISOPTS= and YAXISOPTS= options only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

 Axis Options for LAYOUT OVERLAYEQUATED 1213

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y axis, the following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

THIN
eliminates alternate tick values.

For the X axis, the following policies are valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.
1214 Chapter 8 / Axis Options in Layouts

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.

(format-options)
specifies one or more formatting options for major tick values.

Together, these options provide parameters for determining an optimal

format (w.d, Ew., BESTw.) for displaying major tick values.

MAXWIDTH=integer
specifies the maximum width for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 6

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values.

The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.

The scale can be a named scale or a scientific-notation scale. The

EXTRACTSCALETYPE= option specifies the scale type. The scale that is
used is appended to the axis label, as shown in the following example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the
label is truncated, and the scale is appended to the truncated label.
Ellipses indicate that the label was truncated, as shown in the following
example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.
 Axis Options for LAYOUT OVERLAYEQUATED 1215

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis.
In that case, the tick value format might differ from the data
format even when a named format is applied to the data
values.

See “boolean ” on page xii for other Boolean values that you can
use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.

A named scale can be millions, billions, or trillions for values of 999

trillion or less, or a multiple of 10 (denoted as 10^n) for values over
999 trillion. For large tick values, the scale factor is set to ensure that
the absolute value of the largest value is greater than 1. For small
fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for

values greater than 1, or a multiple of 1/10 expressed as 10^–n for
values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621 .

Note If you specify a format that significantly reduces precision,

because of tick-value rounding, the plot data elements might not
1216 Chapter 8 / Axis Options in Layouts

align properly with the axis tick values. In that case, specify a
tick-value format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

VIEWMAX=number
specifies the maximum data value to include in the display on the X or Y axes
when the axis lengths and major tick values are equal. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the X and Y axes.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option is ignored when EQUATETYPE=SQUARE.

This option does not determine the maximum axis tick value that
is displayed. The THRESHOLDMAX= value is used to determine
the maximum tick value.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display on the X or Y axes
when the axis lengths and major tick values are equal. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the X and Y axes.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Interactions This option is ignored when EQUATETYPE=SQUARE.

 Axis Options for LAYOUT OVERLAYEQUATED 1217

This option does not determine the minimum axis tick value that
is displayed. The THRESHOLDMIN= value is used to determine
the minimum tick value.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Details
The LAYOUT OVERLAYEQUATED statement is used for equated layouts where the
X and Y axes always have equal increments between tick values. Because the axes
within the equated layout are so closely correlated, some axis adjustments cannot
be applied to one axis without applying them to the other.

For example, the INTEGER= option sets evenly spaced integer values for the axis
tick marks. That setting must be applied to both axes if the correlation between
them is to be maintained. The LAYOUT OVERLAYEQUATED statement provides
the COMMONAXISOPTS= option for specifying the INTEGER= option and other
options whose settings apply in common to both axes.

Despite the close correlation between the axes, some axis adjustments can be
made to one axis without affecting the other. For example, displaying grid lines on
one axis has no impact on the other. The XAXISOPTS= and YAXISOPTS= options
are available for applying settings separately to the X and Y axes.

The following example template sets evenly spaced integer values for the axis tick
marks of both axes. It also specifies the display of grid lines, tick marks, and tick
values for the Y axis:
begingraph;
layout overlayequated /
commonaxisopts=(integer=true);
yaxisopts=(griddisplay=on display=(ticks tickvalues));
seriesplot x=var1 y=var2;
endlayout;
endgraph;
 1218 Chapter 8 / Axis Options in Layouts

Axis Options for LAYOUT DATALATTICE and

LAYOUT DATAPANEL
Axis options for the plots within DATALATTICE and DATAPANEL layouts

See: “LAYOUT DATALATTICE Statement” on page 51

“LAYOUT DATAPANEL Statement” on page 79

Syntax
The X and X2 axis options for the plots within DATALATTICE and DATAPANEL
layouts are specified with the following options:

COLUMNAXISOPTS=(axis-options)
COLUMN2AXISOPTS=(axis-options)

The Y and Y2 axis options for the plots within DATALATTICE and DATAPANEL
layouts are specified with the following options:
ROWAXISOPTS=(axis-options)
ROW2AXISOPTS=(axis-options)

General Options for All Axes in the Layout

The options that are documented in this section can be used with any of the axis
types that are supported within a DATALATTICE or DATAPANEL layout.
Subsequent sections in the chapter document the axis options that are available
only for specific axis types: discrete, linear, log, or time axes. The following table
provides a summary of the general options.

Statement Option Description

ALTDISPLAY Controls which axis features are displayed on

second, fourth, and other even row or column
occurrences of the primary axis.

ALTDISPLAYSECONDARY Controls which features are displayed on second,

fourth, and other even row or column occurrences
of the secondary axis.

DISCRETEOPTS Specifies options for a discrete axis.

DISPLAY Controls which axis features are displayed on

first, third, and other odd row or column
occurrences of the primary axis.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1219

Statement Option Description

DISPLAYSECONDARY Controls which axis features are displayed on

first, third, and other odd row or column
occurrences the secondary axis.

GRIDATTRS Specifies the attributes of the grid lines.

GRIDDISPLAY Specifies whether axis grid lines are displayed.

LABEL Specifies the axis label.

LABELATTRS Specifies the color and font attributes of the axis

label.

LABELFITPOLICY Specifies a policy for fitting axis labels in the

available space.

LABELPOSITION Specifies the position of the axis label.

LABELSPLITCHAR Specifies one or more characters on which the

axis labels can be split, if needed.

LABELSPLITCHARDROP Specifies whether the split characters should be

included in the axis labels that are displayed.

LABELSPLITJUSTIFY Specifies the justification of the strings that are

inside the axis label blocks.

LINEAROPTS Specifies options for a standard numeric interval

axis.

LOGOPTS Specifies options for a log axis.

NAME Assigns a name to an axis for reference in other

statements.

OFFSETMAX Reserves an area at the maximum end of the axis.

No tick marks are displayed in the reserved area.

OFFSETMIN Reserves an area at the minimum end of the axis.

No tick marks are displayed in the reserved area.

REVERSE Specifies whether the tick values should appear in

the reverse order.

SHORTLABEL Specifies an alternate axis label.

1220 Chapter 8 / Axis Options in Layouts

Statement Option Description

TICKVALUEATTRS Specifies the color and font attributes of the axis

tick value labels.

TICKVALUEHALIGN Specifies the horizontal alignment for all of the

tick values that are displayed on the Y and Y2
axes.

TICKVALUEVALIGN Specifies the vertical alignment for all of the tick

values that are displayed on the X and X2 axes.

TIMEOPTS Specifies options for a TIME axis.

TYPE Specifies the type of axis to use.

ALTDISPLAY=STANDARD | ALL | NONE | (display-options

controls which axis features are displayed on second, fourth, and other even row
or column occurrences of the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed

(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL
displays the axis label.

A common label is displayed at the center of the grid, and the label
applies to all the primary axes in the row or column. This label is
overridden by a label specified on the DISPLAY= option.

LINE
displays the axis line.

TICKS
displays the tick marks.

TICKVALUES
displays the values that are represented by the major tick marks.

Default The settings on the DISPLAY= option.

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1221

This option can be used to obtain the alternating axis information as

seen in a ScatterPlotMatrix.

See “Details” on page 1275 for more information about the primary and
secondary axes.

GRIDDISPLAY= and GRIDATTRS= for setting axis grid lines.

ALTDISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which features are displayed on second, fourth, and other even row or
column occurrences of the secondary axis. A secondary axis is not an
independent axis. Rather, it mirrors the primary axis (though it can use different
display features). Thus, for this option to take effect, all plot statements in the
LAYOUT PROTOTYPE must map data to the same primary axis. For example, a
secondary X2 axis can be displayed on top in the layout, provided all plot
statements set XAXIS=X to map data to the primary X axis (bottom). Similarly,
a secondary Y2 axis can be displayed to the right in the layout, provided all plot
statements set YAXIS=Y to map data to the primary Y axis (left).

NONE
specifies that no axis features are displayed

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis

ALL
currently the same as STANDARD

(display-options)
a space-separated list of one or more of the following options enclosed in
parentheses:

LABEL
displays the axis label.

A common label is displayed at the center of the grid, and the label
applies to all the secondary axes in the row or column. This label is
overridden by a label specified on the DISPLAY= option.

LINE
displays the axis line.

TICKS
displays the tick marks.

TICKVALUES
displays the values that are represented by the major tick marks.

Default The settings on the DISPLAYSECONDARY= option.

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2,
then both the X and X2 axis are primary and a secondary X axis
cannot be displayed. In that case, this option is ignored. The same
applies for the Y axes.
1222 Chapter 8 / Axis Options in Layouts

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

This option can be used to obtain the alternating axis information as

seen in a ScatterPlotMatrix.

See “Details” on page 1275 for more information about the primary and
secondary axes.

GRIDDISPLAY= and GRIDATTRS= for setting axis grid lines.

DISCRETEOPTS=(discrete-axis-options)
specifies one or more options for a discrete axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not DISCRETE.

See “Options for Discrete Axes Only” on page 1235 for the options that
you can use for discrete-axis-options.

DISPLAY=STANDARD | ALL | NONE | (display-options)

controls which axis features are displayed on first, third, and other odd row or
column occurrences of the primary axis.

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed

ALL
currently the same as STANDARD

NONE
specifies that no axis features are displayed

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks

Default STANDARD

Note When LABEL is specified, a common label is displayed at the center of

the grid, and the label applies to all the primary axes in the row or
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1223

column. This label overrides a label specified on the ALTDISPLAY=

option.

Tips The default line attributes for the axis line and axis tick marks are
defined in the GraphAxisLine style element.

Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis grid
lines.

When LINE is excluded from the DISPLAY= option, the layout wall
outline or the default baseline of a bar chart, needle plot, or waterfall
chart can appear to be an axis line. To suppress the wall outline, use the
WALLDISPLAY= option in the layout statement. To suppress the plot
baseline, use the BASELINEATTRS= option in the plot statement.

See “Details” on page 1275 for more information about the primary and
secondary axes..

DISPLAYSECONDARY=NONE | ALL | STANDARD | (display-options)

controls which axis features are displayed on first, third, and other odd row or
column occurrences of the secondary axis. A secondary axis is not an
independent axis. Rather, it mirrors the primary axis (though it can use different
display features). Thus, for this option to take effect, all plot statements in the
LAYOUT PROTOTYPE must map data to the same primary axis. For example, a
secondary X2 axis can be displayed on top in the layout, provided all plot
statements set XAXIS=X to map data to the primary X axis (bottom). Similarly,
a secondary Y2 axis can be displayed to the right in the layout, provided all plot
statements set YAXIS=Y to map data to the primary Y axis (left).

NONE
specifies that no axis features are displayed

STANDARD
specifies that the LABEL, LINE, TICKS, and TICKVALUES are displayed on
the secondary axis

ALL
currently the same as STANDARD

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or more of the following options:

LABEL
displays the axis label

LINE
displays the axis line

TICKS
displays the tick marks

TICKVALUES
displays the values that are represented by the major tick marks
1224 Chapter 8 / Axis Options in Layouts

Default NONE

Restriction If some plot statements set XAXIS=X and others set XAXIS=X2,
both the X and X2 axis are primary and a secondary X axis cannot be
displayed. In that case, this option is ignored. The same applies for
the Y axes.

Note When LABEL is specified, a common label is displayed at the center

of the grid, and the label applies to all the primary axes in the row or
column. This label overrides a label specified on the ALTDISPLAY=
option.

Tip Use the GRIDDISPLAY= and GRIDATTRS= options to set the axis
grid lines.

See for more information about the primary and secondary axes.“Details”
on page 1275 .

GRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the grid lines.

Default The GraphGridLines style element.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tip On a log axis, this option affects the appearance of the major grid
lines only. It does not affect the appearance of the minor grid lines.
To control the appearance of the minor grid lines on a log axis, use
the MINORGRIDATTRS= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

GRIDDISPLAY=AUTO_OFF | AUTO_ON | ON | OFF

specifies whether axis grid lines are displayed. This option enables the template
to absolutely control the display of grid lines or to allow interaction with the
current style to decide whether grid lines are displayed.

AUTO_OFF
specifies that grid lines are not displayed unless the GraphGridLines element
in the current style contains DisplayOpts="ON."

AUTO_ON
specifies that grid lines are displayed unless the GraphGridLines element in
the current style contains DisplayOpts="OFF."

ON
specifies that grid lines are always displayed.

The current style has no override.

OFF
specifies that grid lines are never displayed.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1225

The current style has no override.

The following table shows the end results for various combinations of the
GRIDDISPLAY= option and the DisplayOpts= attribute of the GraphGridLines
style element. Most supplied templates use the default setting AUTO_OFF to
indicate a preference for not displaying grid lines, but allowing the style to
override.

DisplayOpts= style
GRIDDISPLAY= option attribute Grid Lines Shown?

AUTO_OFF AUTO no

AUTO_OFF ON yes

AUTO_OFF OFF no

AUTO_ON AUTO yes

AUTO_ON ON yes

AUTO_ON OFF no

ON any value yes

OFF any value no

Default AUTO_OFF

Note Supplied styles use DisplayOpts="AUTO," which means that the style
has no preference about grid lines and the graphics template setting for
grid lines is always used.

LABEL="string" | ("string" …"string")

specifies the axis label. The string can be either a string literal or a dynamic. The
list form implies that all included string literals or dynamic variables will be
concatenated.

Default The default label is derived from the primary plot in the layout. For
more information, see “When Plots Share Data and a Common Axis”
on page 1046.

Interaction This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

Note If the axis label is too long to fit along the axis, then it is truncated
by default.

Tips Use the SHORTLABEL= option to specify an alternate axis label to

be used whenever truncation would normally occur.
1226 Chapter 8 / Axis Options in Layouts

Use the ODS escape sequence to specify Unicode characters in an

axis label. Here is an example that species degrees Fahrenheit in an
axis label for temperature:
label="Temperature ((*ESC*){unicode '00B0'x}F)"

See “Overriding the Default Axis Label” in SAS Graph Template Language:
User’s Guide

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis label.

Default The GraphLabelText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display the axis label.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LABELFITPOLICY=AUTO | SPLITALWAYS | STACKED

specifies a policy for fitting axis labels in the available space.

AUTO
uses the short label, when specified, instead of the original label.

If the short label does not fit, then it is clipped. When no short label is
specified, the original label is clipped.

SPLIT
splits the axis label at a split character only when necessary in order to make
the label fit the available space.

The split character is specified by the LABELSPLITCHAR= option. The short

label is not used. A split does not occur at a split character if a split is not
needed at that location. If the label does not contain any of the specified
split characters, then it is not split. A label that cannot be split or that does
not fit the available space even after splitting might overlap the adjoining
space.

SPLITALWAYS
always split the axis label at every occurrence of a split character, which is
specified by the LABELSPLITCHAR= option.

If the label cannot be split, then it is clipped.

STACKED
displays the Y or Y2 axis label vertically as stacked letters.

Note: This feature applies to SAS 9.4M5 and to later releases.

Valid in ROWAXISOPTS= and ROW2AXISOPTS= only

 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1227

Default AUTO

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

Note When LABELPOSITION=CENTER, the available area is the full axis,

including the axis offsets. When LABELPOSITION=DATACENTER,
the available area is the tick display area, excluding the axis offsets.

LABELPOSITION=CENTER | DATACENTER | TOP | BOTTOM | LEFT | RIGHT

specifies the position of the axis label.

CENTER
centers the axis label in the axis area.

For the Y and Y2 axes, the label is oriented vertically and is centered in the
axis area (including the offsets). The label is positioned to the left of the tick
values for the Y axis or to the right of the axis values for the Y2 axis. For the
X and X2 axes, the label is centered in the axis area (including the offsets). It
is positioned below the tick values for the X axis or above the axis values for
the X2 axis.

DATACENTER
repeats the axis label for each row or column and centers each label in the
axis tick display area of its row or column.

For the Y and Y2 axes, each label is oriented vertically and is centered in the
axis tick display area (excluding the offsets) of its row. The labels are
positioned to the left of the tick values for the Y axis or to the right of the
axis values for the Y2 axis. For the X and X2 axes, each label is centered in
the axis tick display area (excluding the offsets) of its column. The labels are
positioned below the tick values for the X axis or above axis values for the
X2 axis.

TOP
BOTTOM
orients the label horizontally at the top or bottom of the axis area.

The label is right-justified in the axis area for the Y axis and left-justified for
the Y2 axis. If there is not sufficient room in the axis area to display the
label, then the label grows to the right for the Y axis and to the left for the
Y2 axis.

Restriction These options are valid for the Y and Y2 axes only.

Note When TOP or BOTTOM is used, the label might collide with other
graphical features. In that case, use CENTER or DATACENTER
instead.

LEFT
RIGHT
positions the label to the left or right of the axis area.

The label is centered vertically in the axis area.

Restriction These options are valid for the X and X2 axes only.
1228 Chapter 8 / Axis Options in Layouts

Note When LEFT or RIGHT is used, the label might collide with other
graphical features.

The following figure shows the CENTER and DATACENTER positions for a blue
Y-axis label Type and a red X-axis label MPG.

The next figure shows the TOP and LEFT positions, and the BOTTOM and
RIGHT positions for the same axis labels.

The following figure shows the CENTER and DATACENTER positions for a blue
Y-axis label Qtr and a red X-axis label Close.

The next figure shows the TOP and LEFT positions, and the BOTTOM and
RIGHT positions for the same axis labels.

Default CENTER

Restriction This option does not support collision avoidance. In some cases, axis
label collisions can occur in the axis area.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1229

Interaction When LEFT, RIGHT, TOP, or BOTTOM is in effect, the

SHORTLABEL= option is ignored.

See SHORTLABEL= on page 1232 for information about how short labels
are used.

LABELSPLITCHAR="character-list"
specifies one or more characters on which the axis labels can be split, if needed.
When multiple split characters are specified, each character in the list is treated
as a separate split character unless the specified characters appear
consecutively in the axis label. In that case, all of the specified split characters
together are treated as a single split character.

When LABELFITPOLICY=SPLIT, if the axis label does not fit the available
space, then it is split on a specified split character only if a split is needed at
that point to make the label fit. In this case, a split might not occur on every
split character. When LABELFITPOLICY=SPLITALWAYS, the axis label is split
unconditionally on every occurrence of a split character. If the axis label does
not contain any of the specified split characters, the label is not split.

"character-list"
one or more characters with no space between each character and enclosed
in quotation marks.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no space between

them. For example, to specify the split characters a, b, and c, use
the following option:
labelsplitchar="abc"

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHARDROP= option specifies whether the split

characters are included in the displayed data label or are dropped.

Notes When multiple characters are specified, the order of the

characters in the list is not significant.

The split characters are case sensitive.

Tip Use the LABELSPLITJUSTIFY= option to specify the justification

of the strings in the axis label block.

LABELSPLITCHARDROP=TRUE | FALSE
specifies whether the split characters are included in the displayed axis labels.

TRUE
drops the split characters from the axis label display.

FALSE
includes the split characters in the axis label display.
1230 Chapter 8 / Axis Options in Layouts

When the label is split while LABELSPLITCHARDROP=FALSE is in effect,

each split character remains as the last character in the current line. The
characters that follow the split character, up to and including the next split
character, are then wrapped to the next line.

Default TRUE. The split characters are dropped from the axis label.

Interactions This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

The LABELSPLITCHAR= option specifies the split characters.

See “boolean ” on page xii for other Boolean values that you can use.

LABELSPLITJUSTIFY=justification
specifies the justification of the strings that are inside the axis label blocks.

justification

CENTER
LEFT
RIGHT
specifies the justification for the X or X2 axis label.

CENTER
TOP
BOTTOM
specifies the justification for the Y or Y2 axis label.

Default CENTER

Interaction This option has effect only when LABELPOSITION= is CENTER or

DATACENTER.

LINEAROPTS=(linear-axis-options)
specifies one or more options for a numeric interval axis. Options must be
enclosed in parentheses. Each option is specified as a name = value pair and
each pair is space separated.

Interaction This option is ignored if the axis type is not LINEAR.

See “Options for Linear Axes Only” on page 1243 for the options that you
can use for linear-axis-options.

LOGOPTS=(log-axis-options)
specifies one or more options for a log axis. Options must be enclosed in
parentheses. Each option is specified as a name = value pair and each pair is
space separated.

Interaction This option is ignored if the axis type is not LOG.

See “Options for Log Axes Only” on page 1255 for the options that you
can use for log-axis-options.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1231

NAME="string"
assigns a name to an axis for reference in other statements. Currently, it is used
only in an AXISLEGEND statement.

Restriction This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Interactions This option is ignored unless the axis is discrete. The axis can be
discrete by default, or explicitly set to discrete with a TYPE=
DISCRETE setting.

For this option to take effect, an axis legend must be enabled. To

enable an axis legend, the DISCRETEOPTS= option must set the
TICKVALUEFITPOLICY to either EXTRACT or EXTRACTALWAYS.
In addition, an AXISLEGEND statement must be specified to
generate the axis legend.

OFFSETMAX=AUTO | AUTOCOMPRESS | number

reserves an area at the maximum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.

For a continuous axis, the offset follows the highest data value or highest
tick value, whichever is greater.

Default AUTOCOMPRESS

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053

OFFSETMIN=AUTO | AUTOCOMPRESS | number

reserves an area at the minimum end of the axis. No tick marks are displayed in
the reserved area.

AUTO
reserves just enough area to fully display markers and other graphical
features near the maximum end of an axis.

AUTOCOMPRESS
applies an automatic offset that prevents axis labels and tick values from
extending beyond the axis length.

number
specifies the offset as a decimal proportion of the full axis length.
1232 Chapter 8 / Axis Options in Layouts

For a continuous axis, the offset precedes the lowest data value or lowest
tick value, whichever is less.

Default AUTOCOMPRESS

Range 0–1. The sum of OFFSETMAX= and OFFSETMIN= should not be more
than 1.

See “Adjusting Axis Offsets” on page 1053.

REVERSE=TRUE | FALSE
specifies whether tick values should appear in the reverse order.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

SHORTLABEL="string"
specifies an alternate axis label to display when the default label or the label
specified by the LABEL= option is too long to fit the available space. When
LABELPOSITION=CENTER (default), the available space for an axis label is the
full axis, including the axis offsets. When LABELPOSITION=DATACENTER, the
available space for an axis label is the axis tick display area, which excludes the
axis offsets. If the label length exceeds the available space, then the label is
anchored at the left or bottom offset. It extends beyond the opposing offset
until it reaches the end of the axis where it is truncated. An ellipsis designates
the truncation.

Interactions This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the axis label.

This option has effect only when the LABELPOSITION= option is

set to CENTER or DATACENTER.

Note If the specified label is itself too long for the grid length or the grid
width, then it is truncated in the display.

TICKVALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the axis tick values.

Default The GraphValueText style element.

Interaction This option is ignored if the DISPLAY= or DISPLAYSECONDARY=

option does not display tick values.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TICKVALUEHALIGN=LEFT | CENTER | RIGHT

specifies the horizontal alignment for all of the tick values that are displayed on
the Y and Y2 axes.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1233

Defaults RIGHT for a Y axis

LEFT for a Y2 axis

Restriction This option is valid for the Y and Y2 axes only.

TICKVALUEVALIGN=TOP | CENTER | BOTTOM

specifies the vertical alignment for all of the tick values that are displayed on
the X and X2 axes.

Defaults TOP for an X axis

BOTTOM for an X2 axis

Restriction This option is valid for the X and X2 axes only.

TIMEOPTS=(time-axis-options)
specifies one or more options for a time axis.

Requirements Columns associated with a time axis must be in SAS time, SAS
date, or SAS datetime units and have an associated SAS time,
date, or datetime format.

Options must be enclosed in parentheses. Each option is specified

as a name = value pair and each pair is space separated.

Interaction This option is ignored if the axis type is not TIME.

1234 Chapter 8 / Axis Options in Layouts

See “Options for Time Axes Only” on page 1265 for the options that
you can use for time-axis-options.

TYPE=AUTO | DISCRETE | LINEAR | TIME | LOG

specifies the type of axis to use.

AUTO
requests that the axis type be automatically determined, based on the
overlay contents.

DISCRETE
uses a DISCRETE axis if possible.

The data for discrete axes can be character or numeric. You can add a
DISCRETEOPTS= option list to customize this axis type.

LINEAR
uses a LINEAR axis if possible.

You can add a LINEAROPTS= option list to customize this axis type.

TIME
uses a TIME axis if possible.

Data for this axis must be SAS time, SAS date, or SAS datetime values. You
can add a TIMEOPTS= option list to customize this axis type.

LOG
uses a LOG axis if possible.

You can add a LOGOPTS= option list to customize this axis type.

Interaction If a log axis is requested and the axis data contains 0 or negative
values, the axis reverts to a linear axis. This outcome can occur
for the response axis of a bar chart, line chart, needle plot, or
waterfall chart when a baseline intercept of 0 or less is specified.
It can also occur for the response axis of a waterfall chart when
an initial bar value of 0 or less is specified. To get a log response
axis in those cases, set the baseline intercept or initial bar value
to a positive value.

Default AUTO

Interactions If this option is set to anything other than AUTO, then plots within
the layout are dropped from the display if their data types or data
ranges do not match the axis type requirements. For more
information, see “Plot Axis Types Must Agree on Common Axes” on
page 1049.

After the axis type is determined (whether you set a specific type
or AUTO is in effect), you can use only options that are supported
by that axis type. For example, if TYPE=TIME, then only the general
OVERLAY axis options and those available on TIMEOPTS= are
supported.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1235

Options for Discrete Axes Only

The options that are documented in this section can be used with the
DISCRETEOPTS= axis option. The following table provides a summary of the
options.

Discrete Axis Option Description

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKTYPE Specifies the position of the axis tick mark.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the list of tick values that are displayed

on the axis.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESPLITCHAR Specifies a list of characters on which the tick

values can be split, if needed.

TICKVALUESPLITCHARDROP Specifies whether the split characters are

included in the displayed tick values.

TICKVALUESPLITJUSTIFY Specifies justification of the strings that are inside

the tick value block.

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default Determined by the system or by the TICKVALUELIST= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.
1236 Chapter 8 / Axis Options in Layouts

Example The following example specifies the axis tick values 10, 20, 30,
and 40, and the tick display values A, B, C, and D:
tickvaluelist=("10" "20" "30" "40");
tickdisplaylist=("A" "B" "C" "D");

TICKTYPE=MIDPOINT | INBETWEEN
specifies the position of the axis tick marks.

MIDPOINT
places the tick marks at the midpoint value location.

INBETWEEN
places the tick marks half way between adjacent midpoint locations.

Default MIDPOINT

Restriction This option applies to discrete axes only.

Note Starting with SAS 9.4M2, when TICKTYPE=INBETWEEN, the

outermost tick marks and grid lines at each end of the axis are not
drawn.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used.

For the Y and Y2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1237

SPLIT
splits the tick value at a split character, which is specified by the
TICKVALUESPLITCHAR= option, only when necessary in order to make the
value fit the available space.

A split does not occur at a split character if a split is not needed at that
location. If the value does not contain any of the specified split characters,
then the value is not split. Values that are not split or that do not fit the
available space even after splitting might overlap the adjoining space.

See TICKVALUESPLITCHAR= on page 1240

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR= on page 1240

SPLITALWAYSTHIN
same as SPLITALWAYS, except that thinning is performed when long words
do not fit the available space.

SPLITTHIN
same as SPLIT, except that thinning is performed when long words do not fit
the available space.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

EXTRACT
displays consecutive integers along the axis instead of the actual tick values
in order to represent those tick values.

In most cases, this policy is implemented if the system estimates that a

collision might occur. If no collision occurs, then the actual tick values are
displayed on the axis in the normal manner.

Requirement The EXTRACT policy must be used with an AXISLEGEND

statement. For more information, see “Extracting Discrete Axis
Tick Values into a Legend” in SAS Graph Template Language:
User’s Guide.

EXTRACTALWAYS
same as EXTRACT, except that the extraction is implemented regardless of
whether collision occurs.

Requirement The EXTRACTALWAYS policy must be used with an

AXISLEGEND statement. For more information, see “Extracting
Discrete Axis Tick Values into a Legend” in SAS Graph Template
Language: User’s Guide.

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

1238 Chapter 8 / Axis Options in Layouts

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

SPLITALWAYS
always splits the axis tick value at every occurrence of a split character that
is specified by the TICKVALUESPLITCHAR= option.

See TICKVALUESPLITCHAR= on page 1240

STACKEDALWAYS
always displays the tick values vertically as stacked letters.

Valid in COLUMNAXISOPTS= and COLUMN2AXISOPTS= only

STACKEDALWAYSTHIN
always displays the tick values vertically as stacked letters.

Thinning is performed when the tick values do not fit the available space.

Valid in COLUMNAXISOPTS= and COLUMN2AXISOPTS= only

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

STAGGERTRUNCATE
attempts the STAGGER policy, and then the TRUNCATE policy.

THIN
eliminates alternate tick values.

TRUNCATE
shortens the tick labels that exceed 12 characters to 12 characters.

A trailing ellipsis indicates that a label is truncated.

 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1239

Note Data labels might overlap on a crowded axis.

TRUNCATEROTATE
attempts the TRUNCATE policy, and then the ROTATE policy.

TRUNCATESTAGGER
attempts the TRUNCATE policy, and then the STAGGER policy.

TRUNCATETHIN
attempts the TRUNCATE policy, and then the THIN policy.

Defaults ROTATE for the Y and Y2 axes

THIN for the X and X2 axes

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

Restrictions This option applies only to discrete axes.

Only character formats are supported.

Interaction This option is ignored when the axis tick values are extracted to an
axis legend. See TICKVALUEFITPOLICY=EXTRACT |
EXTRACTALWAYS.

Tip Use this option when you want to duplicate tick values on an axis.

TICKVALUELIST=(string-list)
specifies the list of tick values that are to be displayed on the axis.

string-list
a space-separated list of values, enclosed in parentheses.

You must enclose each value in the list in quotation marks.

Only the tick values that are included in the string list are displayed on the axis.
The values are displayed in the order in which they are listed. The data values
that are not in the list are dropped. The list can be a subset of the data values. It
can also contain values that are not included in the actual data. A tick value that
is not included in the data appears on the axis, but no data is represented at its
tick mark.

Requirements The list of values must be enclosed in parentheses.

Each value must be enclosed in quotation marks and separated

from adjacent values by a blank space.

Notes If the string list contains duplicate values, then the first
occurrence of the duplicated value in the list is honored and the
remaining instances are ignored.
1240 Chapter 8 / Axis Options in Layouts

When the values specified in the list are compared with the actual
data values, leading blanks are honored and trailing blanks are
ignored.

Tips You can use this option to subset the axis values or to display the
values in a specific order.

You can use this option to display values on the axis that are not
contained in the data.

Examples The following example specifies the axis tick values Sedan,
Sports, Wagon, and SUV:
tickvaluelist=("Sedan" "Sports" "Wagon" "SUV")

The following example specifies the axis tick values 10, 20, 30,
and 40:
tickvaluelist=("10" "20" "30" "40")

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESPLITCHAR="character-list"
specifies a list of characters on which the tick values can be split, if needed.
When multiple characters are specified, each character in the list is treated as a
separate split character unless the specified characters appear consecutively in
the tick value. In that case, all of the specified split characters together are
treated as a single split character.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1241

When TICKVALUESPLITPOLICY=SPLIT, if a tick value collision is detected, then

the tick value is split on a split character only if necessary at that point in order
to avoid collision. In that case, a split might not occur on every split character.
When TICKVALUEFITPOLICY=SPLITALWAYS, the tick value is split
unconditionally on every occurrence of a split character. If the tick value does
not contain any of the specified split characters, then it is not split.

"character-list"
one or more characters with no delimiter between each character.

Default A blank space

Requirements The list of characters must be enclosed in quotation marks.

Multiple characters must be specified with no delimiters. For

example, to specify the characters a, b, and c, use the following
option:
tickvaluesplitchar="abc"

Interactions This option is ignored unless option TICKVALUEFITPOLICY= is

set to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

The TICKVALUEFITPOLICY= option sets the policy that is used to

manage the split behavior of the tick values.

The TICKVALUESPLITCHARDROP= option specifies whether the

split characters are displayed or dropped from the display.

Notes When multiple characters are specified, the order of characters in

the list is not significant.

The split characters are case sensitive.

Tips Use the TICKVALUESPLITJUSTIFY= option to specify the

justification of the strings in the tick value block.

For the X and X2 axis tick values, use the TICKVALUEVALIGN=

option to specify the vertical alignment of the tick values.

For the Y and Y2 axis tick values, use the TICKVALUEHALIGN=

option to specify the horizontal alignment of the tick values.

Example The following example specifies a blank space, a comma, and an

underscore as split characters:
tickvaluesplitchar=" ,_"

TICKVALUESPLITCHARDROP=TRUE | FALSE
specifies whether the split characters should be included in the displayed tick
values. The split characters are specified by the TICKVALUESPLITCHAR=
option.

TRUE
drops the split characters from the tick value display.

The following figure shows an example in which

TICKVALUESPLITCHARDROP=TRUE and three-word, asterisk-delimited
1242 Chapter 8 / Axis Options in Layouts

tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that the asterisk delimiter is not displayed.

FALSE
includes the split characters in the tick value display.

The fit policy determines how the characters are displayed. If the display
policy is SPLIT or SPLITTHIN and TICKVALUESPLITCHARDROP=FALSE,
then each tick value is split at a split character only where a split is
necessary in order to make the value fit the available space. A split might not
occur at every split character in the tick value. At each split point, the split
character remains as the last character in the current line. The characters
that follow the split character, up to and including the split character at the
next split point, are then wrapped to the following line. This process repeats
until the entire data tick value is displayed. The following figure shows an
example in which TICKVALUESPLITCHARDROP=FALSE and three-word,
asterisk-delimited tick values are split on the asterisk character by using the
SPLIT policy.

Notice that a split occurs on the first asterisk and not at the second. In this
case, a split is not needed at the second asterisk.

If the fit policy is SPLITALWAYS or SPLITALWAYSTHIN, and

TICKVALUESPLITCHARDROP=FALSE, then each tick value is split at every
instance of a split character in the value regardless of whether a split is
actually needed. Each split character remains as the last character in the
current line. The characters that follow each split character, up to and
including the next split character, are then wrapped to the next line. The
following figure shows an example in which
TICKVALUESPLITCHARDROP=FALSE and three-word, asterisk-delimited
tick values are split on the asterisk character by using the SPLITALWAYS
policy.

Notice that a split occurs after each asterisk and each asterisk appears at
the end of the line. In this case, three lines are displayed.

Default TRUE
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1243

Interactions The TICKVALUESPLITCHAR= option specifies the split character or

characters.

This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUESPLITJUSTIFY=CENTER | LEFT | RIGHT

specifies justification of the strings that are inside the tick value block. The
justification is relative to an individual tick value’s display area and does not
affect the display of tick values that are not split.

Defaults CENTER for an X or X2 axis

RIGHT for a Y axis

LEFT for a Y2 axis

Interaction This option is ignored unless option TICKVALUEFITPOLICY= is set

to SPLIT, SPLITALWAYS, SPLITTHIN, or SPLITALWAYSTHIN.

Options for Linear Axes Only

The options that are documented in this section can be used with the
LINEAROPTS= axis option. The following table provides a summary of the options.

Linear Axis Option Description

INTEGER Specifies that evenly spaced integer values are

used for tick marks.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

1244 Chapter 8 / Axis Options in Layouts

Linear Axis Option Description

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether the minor tick marks are

displayed on the axis.

THRESHOLDMAX Specifies a bias for including one more tick mark

at the maximum end of the axis.

THRESHOLDMIN Specifies a bias for including one more tick mark

at the minimum end of the axis.

TICKDISPLAYLIST Specifies the text that is displayed for the tick

values that are defined in the TICKVALUELIST=
option.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision.

Only the default policy (THIN) is available for a Y
or Y2 axis.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a linear

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification can

extend the axis data range.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

TICKVALUESEQUENCE Specifies the tick values for a linear axis by start,

end, and increment.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

INTEGER=TRUE | FALSE
specifies that evenly spaced integer values are used for tick marks.

Default FALSE

Interactions This option is overridden by the TICKVALUELIST= or

TICKVALUESEQUENCE= option.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1245

This option overrides the MAXDECIMALS= and

PREFERREDDECIMALS= suboptions of the TICKVALUEFORMAT=
option.

INTEGER=TRUE is ignored for the X or X2 axis when a histogram

plot is the primary plot and BINAXIS=TRUE is specified in the
HISTOGRAMPARM statement.

Note This option is ignored when the data exceeds the limits of integer
values.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)
1246 Chapter 8 / Axis Options in Layouts

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Defaults Four ticks with five intervals in SAS 9.4M1 and earlier releases.

One tick with two intervals starting with SAS 9.4M2.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1247

Default FALSE

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

THRESHOLDMAX=number
specifies a bias for including one more tick mark at the maximum end of the
axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

THRESHOLDMIN=number
specifies a bias for including one more tick mark at the minimum end of the axis.

Default 0.30

Range 0–1

Restriction This option applies to linear axes only.

1248 Chapter 8 / Axis Options in Layouts

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is used.

Tips If the threshold is set to 0, the potential tick mark is never displayed.
If the threshold is set to 1, then the tick mark is always displayed.

Specifying THRESHOLDMIN=0 and THRESHOLDMAX=0 prevents

the tick marks from extending beyond the data range.

Specifying THRESHOLDMIN=1 and THRESHOLDMAX=1 ensures

that the data range is bounded by tick marks.

For the minimum axis length, set the THRESHOLDMIN= and

THRESHOLDMAX= options to 0.

See “Adjusting Axis Thresholds” on page 1052

TICKDISPLAYLIST=(string-list)
specifies the text that is displayed for the tick values that are defined in the
TICKVALUELIST= option. The string list is a space-separated list of string
values that are displayed on the axis in place of the values in the
TICKVALUELIST= option. The strings map one-to-one positionally with the
values that are listed in the TICKVALUELIST= option.

Default The display of tick values is controlled by the

TICKVALUEFORMAT= option.

Requirements The list of values must be enclosed in parentheses.

Each value (character and numeric) must be enclosed in quotation

marks and separated from adjacent values by a blank space.

Interaction When this option is specified, the TICKVALUEFORMAT= option is

ignored.

Tip This option should be used with the TICKVALUELIST= option. The
number of items in the list for this option should equal the
number of items in the list for the TICKVALUELIST= option.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an axis. The effectiveness
of a collision-avoidance policy depends on the number of tick values, their
length, and the length of the axis. Which policies are valid depends on the axis
on which this option is used. For the Y and Y2 axes, the following policies are
valid:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

THIN
eliminates alternate tick values.

For the X and X2 axes, the following policies are valid:

 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1249

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=(format-options) | DATA | format

specifies how to format the values for major tick marks.

(format-options)
specifies one or more formatting options for major tick values.

Together, these options provide parameters for determining an optimal

format (w.d, Ew., BESTw.) for displaying major tick values.

MAXWIDTH=integer
specifies the maximum width for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 8

MAXDECIMALS=integer
specifies the maximum number of decimals for displayed tick values.

Values might be rounded or converted to E notation to fit into this width.

Default 6
1250 Chapter 8 / Axis Options in Layouts

Note The MAXWIDTH= option value should be greater than the

MAXDECIMALS= option value.

PREFERREDDECIMALS=integer
specifies the number of decimal places that you want to display for most
values.

The actual number might vary based on other constraints.

Default 2

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.

The scale can be a named scale or a scientific-notation scale. The

EXTRACTSCALETYPE= option specifies the scale type. The scale that is
used is appended to the axis label, as shown in the following example.
Total Sales (millions)

For long axis labels, if the scale does not fit the available space, then the
label is truncated, and the scale is appended to the truncated label.
Ellipses indicate that the label was truncated, as shown in the following
example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases in which the scale does not fit even with truncation, the
entire axis is dropped.

Default FALSE

Restriction The scale that is extracted by the EXTRACTSCALE= option

is derived from the English locale for all locales.

Interactions The scale type is determined by the EXTRACTSCALETYPE=

option.

If the axis label is not displayed, then the

EXTRACTSCALE=TRUE option is ignored.

Note When EXTRACTSCALE=TRUE and a scale is extracted, the

tick values are formatted to provide the best fit on the axis.
In that case, the tick value format might differ from the data
format even when a named format is applied to the data
values.

See “boolean ” on page xii for other Boolean values that you can
use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1251

A named scale can be millions, billions, or trillions for values of 999

trillion or less, or a multiple of 10 (denoted as 10^n) for values over
999 trillion. For large tick values, the scale factor is set to ensure that
the absolute value of the largest value is greater than 1. For small
fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a
multiple of 1/10 (10^–n) for values less than 1 trillionth.

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for

values greater than 1, or a multiple of 1/10 expressed as 10^–n for
values less than 1.

Default DEFAULT

Restriction The scale is derived from the English locale for all locales.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621 .

Note If you specify a format that significantly reduces precision,

because of tick-value rounding, the plot data elements might not
align properly with the axis tick values. In that case, specify a
tick-value format with a higher precision.

Default (MAXWIDTH=8, MAXDECIMALS=6, PREFERREDDECIMALS=2,

EXTRACTSCALE=FALSE, EXTRACTSCALETYPE=DEFAULT)

Interaction This option is ignored when the TICKDISPLAYLIST= option is

specified.

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option
is used, the only tick values that appear are the tick values in
numeric-list that fall within the explicit data range (set by
VIEWMIN= and VIEWMAX=) or the implicit data range (set by the
actual data minimum and data maximum).

Restriction This option applies to linear axes only.

1252 Chapter 8 / Axis Options in Layouts

Requirement The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

Interactions This option overrides the INTEGER= option.

This option is ignored if the TICKVALUESEQUENCE= option is

specified, or if the DISPLAY= option or the DISPLAYSECONDARY=
option does not display tick values.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some data
not being displayed. For example, data might not be displayed
when the VIEWMIN= value is greater than the actual data
minimum, or when the VIEWMAX= value is less than actual data
maximum.

If TICKVALUEPRIORITY= TRUE, then the VIEWMIN= and

VIEWMAX= options are ignored if they are fully enclosed by the
numeric-list. The tick numeric-list can extend the implicit data
range of the axis, but cannot reduce it.

If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform tick values such as 3, 5, 11, and 23, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform tick values such as 0, 10, 20, and 30 in the
TICKVALUELIST= option.

Tip The values in the list are formatted according to the setting for the
TICKVALUEFORMAT= option.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether an axis tick specification (TICKVALUELIST= or
TICKVALUESEQUENCE=) can extend the axis data range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by either the TICKVALUELIST= or
TICKVALUESEQUENCE= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Restriction This option applies to linear axes only.

 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1253

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the TICKVALUELIST= or

TICKVALUESEQUENCE= option is not specified.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

TICKVALUESEQUENCE=(START=number END=number INCREMENT=number)

specifies the tick values by start, end, and increment.

START=number
specifies the value for the first tick mark.

END=number
specifies the value for the last tick mark.

INCREMENT=number
specifies the increment for intermediate tick marks between the first and
last tick marks.
1254 Chapter 8 / Axis Options in Layouts

The END value always controls the last tick mark. The interval between the
last tick mark and the previous tick mark might not necessarily be the
INCREMENT value.

Default An internal algorithm determines the tick marks, based on the

actual axis data range or the data range established by the
VIEWMIN= and VIEWMAX= options. By default, when this option is
used, the only tick values that appear are those that fall within the
explicit data range (set by VIEWMIN= and VIEWMAX=) or the
implicit data range (set by the actual data minimum and data
maximum).

Interactions This option overrides the INTEGER= option.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the START= option value and the
VIEWMAX= option is set to the END= option value, then all ticks in
the tick sequence are displayed.

If TICKVALUEPRIORITY= TRUE, then the tick sequence might

extend the explicit data range of the axis, but never reduce it.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display tick marks.

Tip The values in the sequence are formatted according to the setting
for the TICKVALUEFORMAT= option.

See TICKVALUELIST= option as an alternative for customizing tick

marks.

VIEWMAX=number
specifies the maximum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the maximum axis tick value that
is displayed. The THRESHOLDMAX= value is used to determine
the maximum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1255

factors such as thresholds and the tick values computed by the

plot statement, are used to determine the axis tick values.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display. The value might be
adjusted by the threshold calculation.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Interactions This option does not determine the minimum axis tick value that
is displayed. The THRESHOLDMIN= value is used to determine
the minimum tick value.

This option is ignored when TICKVALUEPRIORITY= TRUE.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Options for Log Axes Only

The options that are documented in this section can be used with the LOGOPTS=
axis option. The following table provides a summary of the options.

Log Axis Option Description

BASE Specifies the base of the logarithmic scale for the

axis values.

MINORGRID Specifies whether grid lines are displayed at the

minor tick marks.
1256 Chapter 8 / Axis Options in Layouts

Log Axis Option Description

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKCOUNT Specifies the number of minor ticks that are

displayed on the axis.

MINORTICKS Specifies whether minor ticks are displayed.

TICKINTERVALSTYLE Specifies how to scale and format the values for

major tick marks.

TICKVALUEFORMAT= Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the tick values for a log axis as a space-

separated list.

TICKVALUEPRIORITY Specifies whether the TICKVALUELIST

specification can extend the axis data range.

VALUESTYPE Specifies the scale that the system uses when

interpreting the TICKVALUELIST=, VIEWMAX=,
and VIEWMIN= option values.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

BASE=10 | 2 | E
specifies the base of the logarithmic scale for the axis values.

Default 10

Restriction This option applies to log axes only.

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1257

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.
1258 Chapter 8 / Axis Options in Layouts

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);

MINORTICKCOUNT=positive-integer
specifies the number of minor ticks that are displayed on the axis.

Default Eight ticks with nine intervals (BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT. ).

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Interactions The DISPLAY= or DISPLAYSECONDARY= option specification

must include TICKS for this option to have any effect.

The MINORTICKS= option must specify TRUE for this option to

have any effect.

Tip To display n intervals between major ticks, use

MINORTICKCOUNT=n-1.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Restriction Minor ticks can be displayed only when BASE=10 and

TICKINTERVALSTYLE= is LOGEXPAND or LOGEXPONENT.

Requirement If you set this option to TRUE, you must also specify a value for
MINORTICKCOUNT=. Otherwise, minor tick marks do not appear
on the log axis.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1259

Interaction If MINORTICKS=TRUE is specified and TICKVALUELIST= specifies

nonuniform log tick values such as 10, 20, 30, and 40, then
MINORTICK=TRUE is ignored. To display minor tick marks in that
case, specify uniform log tick values such as 1, 10, 100, and 1000 in
the TICKVALUELIST= option.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

TICKINTERVALSTYLE=AUTO | LOGEXPAND | LOGEXPONENT | LINEAR

specifies how to scale and format the values for major tick marks.

AUTO
selects a LOGEXPAND, LOGEXPONENT, or LINEAR representation
automatically based on the range of the data.

When the data range is small (within an order of magnitude), a LINEAR

representation is typically used. Data ranges that encompass several orders
of magnitude typically use the LOGEXPAND or LOGEXPONENT
representation.

LOGEXPAND
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are expanded as follows:

Base=10

Base=2

Base=E

LOGEXPONENT
places the major tick marks at uniform intervals at integer powers of the
base.

The tick values are only the integer exponents for all bases.

LINEAR
places the major tick marks at non-uniform intervals that cover the range of
the data.

Default AUTO
1260 Chapter 8 / Axis Options in Layouts

Restrictions This option applies to log axes only.

For LOGEXPONENT, formats on data columns contributing to the

axis are ignored. For LOGEXPAND, formats on data columns
contributing to the axis are ignored, although any "named format"
on the column is retained. For LINEAR, ticks values are
automatically formatted when the column format is not assigned or
one of w.d, Ew., or BESTw. Other formats (SAS defined or user-
defined) are used if specified.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

Note When BASE=10 and LOGEXPAND or LOGEXPONENT is used, an

intermediate tick is added whenever the axis data range is less than
or equal to 1.5 powers of 10.

Tip If you use TICKINTERVALSTYLE=LOGEXPONENT, then you might

want to include information in the axis label about which base is
used.

TICKVALUEFORMAT=DATA | format
specifies how to format the values for major tick marks.

Note: This feature applies to SAS 9.4M3 and to later releases.

DATA
uses the format that has been assigned to the column that is contributing to
the axis (or BEST6 if no format is assigned) in order to control the formatting
of the major tick values.

format
specifies a format to apply to the major tick values.

Restriction GTL currently honors most, but not every, SAS format. For
details, see Appendix 4, “Format Support in ODS Graphics,” on
page 1621.

Restriction This option applies to log axes only.

Interactions This option is ignored when

TICKINTERVALSTYLE=LOGEXPONENT.

When TICKINTERVALSTYLE=LOGEXPAND, this option is honored

for the base 10 and base 2 logarithmic scales, and is ignored for the
base E scale.

When TICKINTERVALSTYLE=LINEAR, this option is honored for

the base 10, base 2, and base E logarithmic scales.

See BASE=

TICKINTERVALSTYLE=
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1261

TICKVALUELIST=(numeric-list)
specifies the tick values for a linear axis as a list.

Default Only the tick values specified in the list that fall within the
explicit data range set by the VIEWMIN= and VIEWMAX= options
or by the implicit data range set by the actual data minimum and
data maximum are displayed. An internal algorithm determines
the tick marks.

Requirements The tick values must be specified as a space-separated list of

numeric values, enclosed in parentheses.

The values that you specify must be appropriate for the

VALUESTYPE= specification. Otherwise, unexpected results
might occur. If VALUESTYPE=EXPANDED is in effect (default),
specify increments of the log base power such as 0.1, 1, 10, 100,
and so on, on a base 10 log axis, for example. If
VALUESTYPE=EXPONENT is in effect, specify integer increments
of the log base power exponent such as 1, 2, 3, and so on.

Interactions The VALUESTYPE= option determines how the values in the list
are interpreted.

The VIEWMIN= and VIEWMAX= options alter the axis data range.
If the VIEWMIN= option is set to the minimum tick list value and
the VIEWMAX= option is set to the maximum tick list value, then
all ticks in the tick list are displayed. This might result in some
data not being displayed. For example, data might not be
displayed when the VIEWMIN= value is greater than the actual
data minimum, or when the VIEWMAX= value is less than actual
data maximum.

If the VIEWMIN= value is greater than the actual data minimum or

the VIEWMAX= value is less than actual data maximum, some
data might not be displayed.

This option is ignored if the DISPLAY= or the

DISPLAYSECONDARY= option does not display the tick values.

If MINORTICKS=TRUE is specified and TICKVALUELIST=

specifies nonuniform log tick values such as 10, 20, 30, and 40,
then MINORTICK=TRUE is ignored. To display minor tick marks in
that case, specify uniform log tick values such as 1, 10, 100, and
1000 in the TICKVALUELIST= option.

See VIEWMIN= and VIEWMAX= options for controlling the data range

TICKINTERVALSTYLE= for specifying the scale and format of the

major tick values

TICKVALUEPRIORITY= for controlling the behavior of the

TICKVALUELIST= option

BASE= for specifying the log base

1262 Chapter 8 / Axis Options in Layouts

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

VALUESTYPE=EXPANDED | EXPONENT
specifies the scale that the system uses when interpreting the
TICKVALUELIST=, VIEWMAX=, and VIEWMIN= option values. Use this option
to choose your preferred way of specifying log-axis values.

EXPANDED
values are interpreted as powers of the base such as 0.1, 1, 10, 100, and so on,
for base 10, for example.

EXPONENT
values are interpreted as integer exponents of the base such as 1, 2, 3, and so
on, for base 10, base 2, and base E.

Default EXPANDED

Note This option does not change the style of the tick values that are
displayed on the axis. It changes only how the VIEWMIN=,
VIEWMAX=, and TICKVALUELIST= option values are interpreted by
the system.

Tip This option is particularly useful when BASE=E.

Examples The following example specifies VIEWMIN= and VIEWMAX= as

exponent values instead of as expanded values on an expanded Base
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1263

10 log axis. This results in X-axis tick values of 10, 100, 1000, 10000,
and 100000.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexpand
valuestype=exponent
viewmin=1 viewmax=5));

The following example specifies TICKVALUELIST= as a list of

expanded values instead of exponent values on an exponent Base 10
log axis. This results in X-axis tick values of 1, 2, 3, 4, and 5.
xaxisopts=(type=log
logopts=(base=10
tickintervalstyle=logexponent
tickvaluepriority=true
valuestype=expanded
tickvaluelist=(10 100 1000 10000 100000)));

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so
on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

If an invalid value is specified for the VIEWMAX= option, the

default value for VIEWMAX= is used instead. In that case, if the
default value for VIEWMAX= is less than the value specified by
the VIEWMIN= option, then the VIEWMIN= and VIEWMAX=
values are swapped.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
1264 Chapter 8 / Axis Options in Layouts

factors such as thresholds and the tick values computed by the

plot statement, are used to determine the axis tick values.

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 100,000 as an

expanded value on a base 10 log axis:
VIEWMAX=100000

The following example specifies a value of 100,000 as an

exponent value on a base 10 log axis:
VIEWMAX=5

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

The value that you specify must be appropriate for the

VALUESTYPE= specification and the log base. Otherwise,
unexpected results might occur. If VALUESTYPE=EXPANDED is
in effect (default), specify an increment of the log base power
such as 0.1, 1, 10, 100, and so on, on a base 10 log axis, for example.
If VALUESTYPE=EXPONENT is in effect, specify an integer
increment of the log base power exponent such as 1, 2, 3, and so
on.

Interactions This option is ignored when TICKVALUEPRIORITY= TRUE.

The VALUESTYPE= option determines how the value is

interpreted.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1265

When BASE=10 and TICKINTERVALSTYLE=LOGEXPAND or

TICKINTERVALSTYLE=LOGEXPONENT is used, an intermediate
tick is added whenever the axis data range is less than or equal to
1.5 powers of 10.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Examples The following example specifies a value of 0.1 as an expanded

value on a base 10 log axis:
VIEWMIN=0.1

The following example specifies a value of 0.1 as an exponent

value on a base 10 log axis:
VIEWMIN=-1

Options for Time Axes Only

The options that are documented in this section can be used with the TIMEOPTS=
axis option. Some of the options are axis-specific. The following tables provide a
summary of the time-axes options based on the layout axis option in which it is
used.

Table 8.4 Time-Axis Options for the COLUMNAXIS= or COLUMN2AXIS=Layout

Option

Time Axis Option Description

INTERVAL Specifies that evenly spaced integer values are

used for tick marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time interval

that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor ticks.

MINORTICKS Specifies whether minor ticks are displayed.

SPLITTICKVALUE Specifies whether to split the tick values on an X

or X2 axis, if possible.

TICKVALUEFITPOLICY Specifies a policy for avoiding tick value collision

on an X or X2 axis.
1266 Chapter 8 / Axis Options in Layouts

Time Axis Option Description

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a time

axis as list.

TICKVALUEROTATION Specifies how the tick values are rotated on the X

and X2 axes.

VIEWMAX Specifies the maximum data value to include in

the display.

VIEWMIN Specifies the minimum data value to include in the

display.

Table 8.5 Time-Axis Options for the ROWAXIS= or ROW2AXIS=Layout Option

Time Axis Option Description

INTERVAL Specifies that evenly spaced integer values are

used for tick marks.

INTERVALMULTIPLIER Specifies a multiplier to apply to the time interval

that is in effect for the axis.

MINORGRID Specifies whether grid lines are displayed at the

minor tick values.

MINORGRIDATTRS Specifies the attributes of the minor grid lines.

MINORTICKINTERVAL Specifies the time interval between minor ticks.

MINORTICKS Specifies whether minor ticks are displayed.

TICKVALUEFORMAT Specifies how to format the values for major tick

marks.

TICKVALUELIST Specifies the order of the tick values for a time

axis as list.

TICKVALUEPRIORITY Specifies whether an axis tick specification that is

specified with the TICKVALUELIST= option can
extend the axis data range.

VIEWMAX Specifies the maximum data value to include in

the display.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1267

Time Axis Option Description

VIEWMIN Specifies the minimum data value to include in the

display.

INTERVAL=AUTO | SECOND | MINUTE | HOUR | DAY | TENDAY | WEEK |

SEMIMONTH | MONTH | QUARTER | SEMIYEAR | YEAR
specifies the time interval between major ticks.
Table 8.6 Time Intervals

Default tick value

INTERVAL Unit Tick interval format

AUTO DATE, TIME, or automatically automatically

DATETIME chosen chosen

SECOND TIME or second TIME8.

DATETIME

MINUTE TIME or minute TIME8.

DATETIME

HOUR TIME or hour TIME8.

DATETIME

DAY DATE or day DATE9.

DATETIME

TENDAY DATE or 10 days DATE9.

DATETIME

WEEK DATE or 7 days DATE9.

DATETIME

SEMIMONTH DATE or 1st and 16th of DATE9.

DATETIME each month

MONTH DATE or month MONYY7.

DATETIME

QUARTER DATE or 3 months YYQC6.

DATETIME

SEMIYEAR DATE or 6 months MONYY7.

DATETIME

YEAR DATE or year YEAR4.

DATETIME
1268 Chapter 8 / Axis Options in Layouts

Default AUTO. An appropriate interval is chosen based on the data and the
column date, date-time, or time format.

Restriction This option applies to time axes only.

Requirement The data column(s) mapped to a time axis must be in the same
duration units: TIME, DATE, or DATETIME. The selection of an
interval must be consistent with the duration unit. For example, if
the data are in time units, you can specify only AUTO, SECOND,
MINUTE, HOUR.

Interaction This option is ignored if the TICKVALUELIST= option is used.

INTERVALMULTIPLIER=positive-integer
specifies a multiplier to apply to the time interval that is in effect for the axis.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1

Restriction This option applies to time axes only.

Interaction This option is ignored if the TICKVALUELIST= option is used.

Tip Use the INTERVAL= option to specify a different time interval.

Examples To specify 3-month intervals:

INTERVAL=MONTH INTERVALMULTIPLIER=3

To specify 10-year intervals:

INTERVAL=YEAR INTERVALMULTIPLIER=10

MINORGRID=TRUE | FALSE
specifies whether grid lines are displayed at the minor tick marks.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphMinorGridLines:DisplayOpts style reference starting with

SAS 9.4M2. If attribute DisplayOpts is not defined in the active
style, then FALSE is the default value.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1269

Interaction This option is ignored if the GRIDDISPLAY= option does not display
the grid lines.

Tips The GRIDATTRS= option does not affect the appearance of the
minor grid lines. To control the minor grid line appearance, use the
MINORGRIDATTRS= option.

Use the MINORTICKS= option to display the minor tick marks on the
axis.

See “boolean ” on page xii for other Boolean values that you can use.

MINORGRIDATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the minor grid lines. This option does not affect the
major grid lines.

The following figure shows the minor grid lines set to light blue, dotted lines on
a linear axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a base-10 log axis. (See the example.)

The following figure shows the minor grid lines set to light blue, dotted lines on
a time axis. (See the example.)

Defaults The GraphGridLines style element is used starting with SAS 9.4.

The GraphMinorGridLines style element is used starting with SAS

9.4M2.

Interaction This option is ignored when MINORTICKS=FALSE.

Note When style-element is specified, only the style element’s

ContrastColor, LineStyle, and LineThickness attributes are used.

Tip Use the GRIDATTRS= option to control the appearance of the major
grid lines.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Line Options” on page 1602 for available line options.

Example Here is an example that specifies light blue, dotted lines for the
minor grid.
minorgridattrs=(color=lightblue pattern=dot);
1270 Chapter 8 / Axis Options in Layouts

MINORTICKINTERVAL=interval
specifies the time interval between minor ticks. See Table 8.15 on page 1267 for
information about the intervals that you can select. The interval that you select
must be consistent with the axis data duration units such as TIME, DATE, or
DATETIME. For example, if the axis data is in TIME units, then you must specify
AUTO, SECOND, MINUTE, or HOUR.

Default AUTO

Interactions This option is ignored if the TICKVALUELIST= option is used.

This option is ignored if the MINORTICKINTERVAL= setting is

greater than the INTERVAL= setting.

MINORTICKS=TRUE | FALSE
specifies whether minor ticks are displayed. When MINORTICKS=TRUE, the
minor tick marks are displayed on the axis as shown in the following figure.

Default FALSE

Interactions The number of minor ticks is dependent on the value of the

MINORTICKINTERVAL= option, if specified. If
MINORTICKINTERVAL= is not specified, then it is dependent on
the value of the INTERVAL= option.

This option is ignored if the TICKVALUELIST= option is used, or if

the DISPLAY= or DISPLAYSECONDARY= option does not display
the tick marks.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, then MINORTICKS=TRUE is ignored, and minor tick marks
are not displayed.

Tip Use the MINORGRID= option to display grid lines at the minor tick
values.

See “boolean ” on page xii for other Boolean values that you can use.

SPLITTICKVALUE=TRUE | FALSE
specifies whether to split the tick values on an X or X2 axis, if possible. This
option is not available for a Y or Y2 axis.

TRUE
splits the axis tick values into two lines allowing more tick values to appear.

For example, with INTERVAL= MONTH, this is how tick values are split:
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1271

FALSE
does not split the axis tick values.

For example, when this option specifies FALSE, this is how the tick values in
the previous example appear:

Typically, fewer tick values fit, causing thinning, rotation, or staggering of the
values. See the TICKVALUEFITPOLICY= option.

Default TRUE

Restrictions This option applies to time axes only.

This option is valid only in the COLUMNAXIS= and

COLUMN2AXIS= layout axis options.

Interaction This option is ignored if the TICKVALUELIST= or

TICKVALUEFORMAT= option is used.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEFITPOLICY=policy
specifies a policy for avoiding tick value collision on an X or X2 axis. This option
is not available for the Y and Y2 axes. The effectiveness of a collision-avoidance
policy depends on the number of tick values, their length, and the length of the
axis. The following policies are valid for the X and X2 axes:

NONE
makes no attempt to avoid collisions between tick values.

Tick values are displayed even when they collide.

ROTATE
rotates the tick values if a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYS
rotates the tick values regardless of whether a collision occurs.

The TICKVALUEROTATION= option specifies whether the values are rotated

to a 45-degree diagonal or a 90-degree vertical position. By default, the
values are rotated to a 45-degree diagonal position.

ROTATEALWAYSDROP
attempts the ROTATEALWAYS policy, and then drops the tick values if
collisions still occur.

ROTATETHIN
attempts the ROTATE policy, and then the THIN policy.

STAGGER
alternates the tick values between two rows.
1272 Chapter 8 / Axis Options in Layouts

STAGGERROTATE
attempts the STAGGER policy, and then the ROTATE policy.

STAGGERTHIN
attempts the STAGGER policy, and then the THIN policy.

THIN
eliminates alternate tick values.

Default THIN

Restriction This option is valid only for the X and X2 axes.

Interaction When SPLITTICKVALUE= on page 1270TRUE, this option is ignored

and only the THIN policy is used.

Note A note is written to the SAS log when tick value thinning occurs.

TICKVALUEFORMAT=format | DATA
specifies how to format the values for major tick marks.

format
specifies a SAS date, time, or datetime format to control how the major tick
values are displayed.

This format must be in the same duration units as the data column(s)
mapped to a time axis: TIME, DATE, or DATETIME and should be appropriate
for the value of the INTERVAL= option. For example, if INTERVAL=MONTH
and there are two years of data displayed on the axis, then choosing
TICKVALUEFORMAT=YEAR. would result in several ticks having the same
year value.

DATA
specifies that the SAS date, time, or datetime format associated with the
data column assigned to the axis be used to control how the major tick
values are displayed.

Default The default format used by the INTERVAL= option. The default
does not apply if TICKVALUELIST= is specified.

Restrictions This option applies to time axes only.

GTL currently honors most but not every SAS format. For details,
see Appendix 4, “Format Support in ODS Graphics,” on page 1621.

TICKVALUELIST=(time-constant-list | date-constant-list | datetime-constant-list |

numeric-list)
specifies the tick values for a time axis as list.

Default An internal algorithm determines the tick values.

Restrictions This option applies to time axes only.

If TICKVALUEPRIORITY= is set to FALSE, then this option does

not extend the data range of the axis. If the values fall within the
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1273

default data range or that specified by the VIEWMIN= or

VIEWMAX= options, then they are used.

Requirement The tick values must be specified as a space-separated list of

values enclosed in parentheses. The items in the list must be in the
same duration units as the data mapped to the axis: TIME, DATE,
or DATETIME. The values can be expressed as SAS TIME, DATE, or
DATETIME constants (for example, "13:23"T, "11MAY06"D, or
"11MAY06:13:23"DT) or their numeric equivalents.

Interactions The values in the list are formatted according to the format
specified on the TICKVALUEFORMAT= option. If
TICKVALUEFORMAT= is not used, then the values are formatted
according to the column format (the default TICKVALUEFORMAT
value is not applied to these values).

If this option is specified, the SPLITTICKVALUE= and INTERVAL=

options are ignored.

If MINORTICKS=TRUE is specified with the TICKVALUELIST=

option, MINORTICKS=TRUE is ignored, and minor tick marks are
not displayed.

TICKVALUEPRIORITY=TRUE | FALSE
specifies whether the TICKVALUELIST= specification can extend the axis data
range.

TRUE
extends the axis data range (but does not reduce it) to include the minimum
and maximum values that are specified by the TICKVALUELIST= option.

If the minimum and maximum of the user-specified values are within the
data range, this option has no effect.

FALSE
displays only the tick values that are specified by the TICKVALUELIST=
option that fall within the explicit data range set by the VIEWMIN= and
VIEWMAX= options or by the implicit data range set by the actual data
minimum and data maximum.

Default FALSE

Restriction This option is valid only in the ROWAXIS= and ROW2AXIS= layout
axis options.

Interactions When this option is set to TRUE, the VIEWMIN= and VIEWMAX=
options are ignored.

This option is ignored if the DISPLAY= option or the

DISPLAYSECONDARY= option does not display the tick values.

This option is ignored if the TICKVALUELIST= option is not

specified.
1274 Chapter 8 / Axis Options in Layouts

Note If the minimum and maximum of the specified values are within the
data range, then this option has no effect.

See “boolean ” on page xii for other Boolean values that you can use.

TICKVALUEROTATION=DIAGONAL | DIAGONAL2 | VERTICAL

specifies how the tick values are rotated on the X and X2 axes.

DIAGONAL
rotates the tick values to a 45-degree diagonal position.

The X labels read left to right in a downward direction. The X2 labels read
left to right in an upward direction.

DIAGONAL2
rotates the tick values to a –45-degree diagonal position.

The X labels read left to right in an upward direction. The X2 labels read left
to right in a downward direction.

Note: This feature applies to SAS 9.4M5 and to later releases.

VERTICAL
rotates the labels to a 90-degree vertical position.

The labels are always drawn from bottom to top.

Default DIAGONAL

Restriction This option is valid for COLUMNAXISOPTS= and

COLUMN2AXISOPTS= only.

Interaction The TICKVALUEFITPOLICY= option must be set to ROTATE or

ROTATEALWAYS for this option to have any effect.

VIEWMAX=number
specifies the maximum data value to include in the display.

Default The maximum value in the data for the specified axis.

Requirements If VIEWMIN= is also specified, VIEWMAX= must be greater than

VIEWMIN=. Otherwise, the default values are used for
VIEWMIN= and VIEWMAX=.

If VIEWMIN= is not specified, VIEWMAX= must be less than the

data maximum. Otherwise, a blank graph is drawn.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The maximum axis tick value might differ from the VIEWMAX=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1275

Tip To display the VIEWMAX= value as the maximum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

VIEWMIN=number
specifies the minimum data value to include in the display.

Default The minimum value in the data for the specified axis.

Requirements If VIEWMAX= is also specified, VIEWMIN= must be less than

VIEWMAX=. Otherwise, the default values are used for
VIEWMAX= and VIEWMIN=.

If VIEWMAX= is not specified, VIEWMIN= must be greater than

the data minimum. Otherwise, a blank graph is drawn.

Notes Setting a VIEWMAX= or VIEWMIN= value does not alter the

original data or any calculations on it.

The minimum axis tick value might differ from the VIEWMIN=
value. The VIEWMIN= and VIEWMAX= values, and additional
factors such as thresholds and the tick values computed by the
plot statement, are used to determine the axis tick values.

Tip To display the VIEWMIN= value as the minimum tick value, use
the TICKVALUELIST= option.

See “Adjusting the Axis View” on page 1051

Details
The LAYOUT DATALATTICE and LAYOUT DATAPANEL statements each create a
grid of graphs based on the values of one or more classification variables. In the
grid, the axes are always external and displayed on the “primary” axes by default.
The axes that are considered primary depend on the settings for the XAXIS= and
YAXIS= options in plot statements that are specified within the LAYOUT
PROTOTYPE. For managing the primary axes, both the LAYOUT DATALATTICE and
LAYOUT DATAPANEL statements have COLUMNAXISOPTS=,
COLUMN2AXISOPTS=, ROWAXISOPTS=, and ROW2AXISOPTS= options that
manage the axis features separately for columns and rows. The settings that are
available can manage odd and even columns and rows separately, enabling you to
simplify the axis display within the grid.

The following table shows which axis is primary for the XAXIS= and YAXIS=
settings, and which axis option to use to manage that primary axis.

Option Setting Primary Axis Axis Option to Use

XAXIS=X X (bottom) COLUMNAXISOPTS=

1276 Chapter 8 / Axis Options in Layouts

Option Setting Primary Axis Axis Option to Use

XAXIS=X2 X2 (top) COLUMN2AXISOPTS=

YAXIS=Y Y (left) ROWAXISOPTS=

YAXIS=Y2 Y2 (right) ROW2AXISOPTS=

The settings that are available for the axis options can manage odd and even
columns and rows separately, enabling you to simplify the axis display within the
grid.
n To manage the first, third, and odd occurrences of a primary axis, use the
DISPLAY= option.
n To manage the second, fourth, and even occurrences of a primary axis, use the
ALTDISPLAY= option.

You can also display “secondary” axes in the grid. A secondary axis is not an
independent axis. Rather, it mirrors the primary axis, but it is displayed on the
opposite side and can have different display options. For example, when the X axis
(bottom) is primary, you can mirror that axis with a secondary X axis at the top of
the grid. Similarly, when the Y2 axis (right) is primary, you can mirror that axis with
a secondary Y2 axis on the left of the grid. A secondary axis makes it easier to
interpolate values in the cells that are farthest away from the primary axis.

Secondary axes can be displayed in the graph, provided all plot statements in the
LAYOUT PROTOTYPE map data to the same primary axis. For example, a
secondary X axis can be displayed at the top of the layout, provided all plot
statements set XAXIS=X to map data to the primary X axis (bottom). Similarly, a
secondary Y2 axis can be displayed to the left in the layout, provided all plot
statements set YAXIS=Y2 to map data to the primary Y2 axis (right). If all plot
statements in the LAYOUT PROTOTYPE do not map data to the same primary axis,
then the secondary axes are not displayed.

To display secondary axes in the grid, use the DISPLAYSECONDARY= and

ALTDISPLAYSECONDARY= options. As with the options for the primary axes, the
DISPLAYSECONDARY= option manages the first, third, and odd occurrences of a
secondary axis. The ALTDISPLAYSECONDARY= option manages the second,
fourth, and even occurrences of a secondary axis.

In the default cases for the plots within the layout, the axis type is always
DISCRETE, LINEAR, or TIME. The TYPE= option enables you to specify an axis type
that overrides the default. For example, when appropriate for the data, you can
request a LOG axis. When you override the default axis type, you must be sure to
specify the correct axis type for the plot(s) that you are defining.

Each axis type has features specific to that type, and the following axis options
enable you to specify features for the different types: DISCRETEOPTS= ,
LINEAROPTS= , LOGOPTS= , and TIMEOPTS= . One or more of these options can
be specified for an axis, but the specified settings are applied only to the axis type
that supports them.
 Axis Options for LAYOUT DATALATTICE and LAYOUT DATAPANEL 1277

Note: Certain plot types or layouts sometimes impose restrictions on what type of
axis can be assigned. See the plot or layout documentation for default axis types
and any restrictions that might apply.

Example: Axis Options for LAYOUT DATALATTICE

and LAYOUT DATAPANEL
The following graph was generated by the “Example Program” on page 1277:

Example Program
This example shows how axis attributes can be managed separately for even and
odd columns and rows in the layout grid. In this case, the ROWAXISOPTS= option is
used to stagger the Y-axes attributes. On the primary (left) Y axis, DISPLAY=
displays TICKS and TICKVALUES on the first and third rows while ALTDISPLAY=
displays just TICKS on the second row. On the secondary (right) Y axis,
DISPLAYSECONDARY= displays just TICKS on the first and third rows, while
ALTDISPLAYSECONDARY= displays TICKS and TICKVALUES on the second row.
This alternating pattern could also have been set for the column axes. The pattern
is independent of the number of rows and columns.
proc template;
define statgraph layoutdatalattice;
begingraph;
1278 Chapter 8 / Axis Options in Layouts

entrytitle "Annual Furniture Sales Comparisons";

layout datalattice rowvar=country columnvar=year /
rowdatarange=union
headerlabellocation=inside
headerlabeldisplay=value
headerbackgroundcolor=GraphAltBlock:color
rowaxisopts=(griddisplay=on
display=(tickvalues)
altdisplay=(ticks)
displaysecondary=(ticks)
altdisplaysecondary=(ticks tickvalues)
linearopts=(tickvalueformat=dollar12.))
columnaxisopts=(display=(tickvalues)
timeopts=(tickvalueformat=monname3.));
layout prototype / cycleattrs=true;
seriesplot x=month y=TotalActual / name="Actual";
seriesplot x=month y=TotalPredict / name="Predict";
endlayout;
sidebar / align=bottom;
discretelegend "Actual" "Predict" / border=false;
endsidebar;
endlayout;
endgraph;
end;
run;

proc summary data=sashelp.prdsal2 nway;

class country year month;
var actual predict;
output out=prdsal2 sum=TotalActual TotalPredict;
run;

proc sgrender data=prdsal2 template=layoutdatalattice;

run;
 1279

PART 6

Legend Statements

Chapter 9
Legend Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
1280
 1281

9
Legend Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
AXISLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
CONTINUOUSLEGEND Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1291
DISCRETELEGEND, MERGEDLEGEND Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1303
LEGENDITEM Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1322
LEGENDTEXTITEMS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1329

Dictionary

AXISLEGEND Statement
Generates consecutive integers for display as axis-tick values in the graph, and creates a legend that
correlates those integers with the actual tick values that they represent.

Requirements: The axis must be named with the axis NAME= option.
The axis must be a discrete axis (TYPE=DISCRETE).
The axis must use the TICKVALUEFITPOLICY=EXTRACT or
TICKVALUEFITPOLICY=EXTRACTALWAYS suboption of the DISCRETEOPTS= axis
option.
Interactions: If the total legend area in the graph exceeds the percentage that is set by the
MAXLEGENDAREA= option in an ODS GRAPHICS statement that is in effect for
the output destination, the legend might be dropped
If DISPLAYCLIPPED=FALSE and the full legend cannot be displayed, the legend
might be dropped.
1282 Chapter 9 / Legend Statements

Syntax
AXISLEGEND "axis-name" < /options >;

Summary of Optional Arguments

Appearance options
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before
the next row begins.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the legend.
DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend
cannot be rendered entirely.
DOWN=positive-integer
specifies the number of legend entries that are placed vertically before
the next column begins.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or
transparent (FALSE).
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend
border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent
layout when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.
LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when
the legend is specified within an overlay-type or REGION layout.
 AXISLEGEND Statement 1283

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend title.

Required Argument
"axis-name"
references an axis’s name. The name controls which axis determines the legend
entries.

Requirement The axis name must be enclosed in quotation marks.

Optional Arguments
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the
next row begins.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Interaction This option is ignored if ORDER= COLUMNMAJOR

Tip When a DISCRETELEGEND statement is placed in a nested layout,

include the ACROSS= option to specify the number of columns.
Otherwise, the legend items might wrap in an unexpected fashion.

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.

NONE
does not automatically align the legend within its parent layout.

The legend’s position is set by the HALIGN= and VALIGN= options.

AUTO
within the parent layout, attempts to center the legend in the area that is
farthest from any surrounding data point markers.

(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the parent layout’s other graphics features.

The location-list is space-separated and can contain any of these locations:

TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT,
BOTTOM, and BOTTOMRIGHT.
1284 Chapter 9 / Legend Statements

Default NONE

Restriction AUTO is available only when the parent layout contains a scatter
plot. Otherwise, it is ignored.

Interactions This option has no effect unless LOCATION= INSIDE.

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type layout.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The Color attribute of the GraphBackground style element.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By

default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.

Defaults TRUE in SAS 9.4M1 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting

with SAS 9.4M2. If attribute FrameBorder is not defined in the active
style, then TRUE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

“Line Options” on page 1602 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend cannot
be rendered entirely. Based on the legend contents and placement, or when the
graph size is reduced, parts of the legend (title, legend symbol, or legend value)
 AXISLEGEND Statement 1285

might be clipped (truncated). When clipping occurs and this option is set to
FALSE, the entire legend is removed from the graph. The space for that legend is
then reclaimed by the remainder of the graph. When this option is set to TRUE,
the legend always appears, even if some parts of it have been clipped.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the next
column begins.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Restriction This option is ignored if ORDER= ROWMAJOR

HALIGN=CENTER | LEFT | RIGHT | number

specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.

number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is
all the way to the left and 1 is all the way to the right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Default CENTER

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN= options

cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or REGION layout.

Default OUTSIDE
1286 Chapter 9 / Legend Statements

Restriction This option has effect only when the legend statement appears
within an overlay-type or REGION layout and at least one
standalone plot statement is referenced by the parent layout.

Interactions The actual position is determined by the settings for the

LOCATION=, AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the legend is
positioned outside of the plot area.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.

Default ROWMAJOR

Interaction If ORDER=ROWMAJOR, then use the ACROSS= option to limit the

number of entries in a row. If ORDER=COLUMNMAJOR, then use the
DOWN= option to limit the number of entries in a column.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.
 AXISLEGEND Statement 1287

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend perimeter.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page xiii

TITLE="string"
specifies the title of the legend.

Default No title
1288 Chapter 9 / Legend Statements

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
between the title and the legend. If TITLEBORDER=FALSE is in
effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.

number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is
on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds

to the containing layout's bottom edge plus half the legend
height. Similarly, one corresponds to the containing layout's top
edge minus half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout. It is ignored if the parent layout
is not an overlay-type or REGION layout.
 AXISLEGEND Statement 1289

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN= options

cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

Details
An AXISLEGEND statement is used to consolidate the axis display so that a large
number of tick values can be clearly displayed without collision. In an axis legend,
the legend symbol is a positive integer that is generated to represent a specific axis
tick position, and the legend value displays the axis tick value. In the “Example:
AXISLEGEND Statement” on page 1290 , the first legend symbol is the integer 1 and
the first legend value is Neon SE 4dr.

To implement an axis legend, all of the following requirements must be met:

n The axis must be a discrete axis. The axis can be discrete by default, or
explicitly set to discrete by setting TYPE=DISCRETE among the parent layout’s
axis-option settings.
n The parent layout’s axis options must specify NAME= to assign a name to the
axis.
n The parent layout’s axis options must use the DISCRETEOPTS= option to
specify either TICKVALUEFITPOLICY=EXTRACT or
TICKVALUEFITPOLICY=EXTRACTALWAYS. EXTRACT implements an axis
legend if the system estimates that a collision might occur. EXTRACTALWAYS
implements an axis legend regardless of whether a collision occurs.
1290 Chapter 9 / Legend Statements

n The AXISLEGEND statement must reference the axis name that was specified
in the parent layout’s NAME= axis option.

When all of these requirements are satisfied, the tick values of the named discrete
axis can be replaced with consecutive positive integers. The axis legend displays
those integer values and the corresponding tick values that the integers represent.

Within an overlay-type layout, when an axis legend is placed inside the plot area
with LOCATION= INSIDE,
n The axis legend is always placed on top of plot lines and markers.

n By default, its background is fully transparent ( OPAQUE= FALSE), meaning that

underlying lines, markers, and data labels show through the legend.
n Its position can be controlled with the AUTOALIGN= option, or with the
HALIGN= and VALIGN= options.

Within an overlay-type layout, when an axis legend is placed outside the plot area
with LOCATION=OUTSIDE,
n By default, its background is fully opaque (OPAQUE=TRUE).

n Its position can be controlled with the HALIGN= and VALIGN= options.

When an axis legend is placed within nested layouts, it might be necessary to do

one of the following to obtain the desired legend organization:
n use the ACROSS= option and also set ORDER= ROWMAJOR

n use the DOWN= option and also set ORDER=COLUMNMAJOR

Example: AXISLEGEND Statement

The following graph was generated by the “Example Program” on page 1291 . The
LAYOUT OVERLAY statement’s XAXISOPTS= option uses NAME= to assign a
name to the X axis so that it can be referenced in an AXISLEGEND statement. The
LAYOUT OVERLAY statement’s DISCRETEOPTS= option specifies
TICKVALUEFITPOLICY=EXTRACT, which implements the axis legend if a collision
occurs in the tick-value display for the X axis. The AXISLEGEND statement
references the name that was assigned to the X axis.
 CONTINUOUSLEGEND Statement 1291

Example Program
proc template;
define statgraph axislegend ;
begingraph;
entrytitle "Mileage for Vehicles Made by Dodge";
layout overlay / xaxisopts=(name="xaxis"
discreteopts=(tickvaluefitpolicy=extract)) ;
barchart category=model response=mpg_highway / stat=mean ;
axislegend "xaxis";
endlayout;
endgraph;
end;

proc sort data=sashelp.cars out=dodge; by descending mpg_highway;

where make="Dodge";

proc sgrender data=dodge template=axislegend;

run;

CONTINUOUSLEGEND Statement
Creates a legend with a color ramp corresponding to a range of values.
1292 Chapter 9 / Legend Statements

Syntax
CONTINUOUSLEGEND "graph-name" < /options >;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the legend.
INTEGER=TRUE | FALSE
specifies whether only integer tick values are used in the continuous
legend.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or
transparent (FALSE).
ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the legend.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend
border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Data range options

VALUECOUNTHINT=positive-integer
recommends a number of values for the continuous legend to use to label
the data range.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent
layout when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.
LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when
the legend is specified within an overlay-type or REGION layout.
VALIGN=CENTER | TOP | BOTTOM | number
 CONTINUOUSLEGEND Statement 1293

specifies the vertical alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.

Scale options
EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it
to reduce the tick value width.
EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend title.

Required Argument
"graph-name"
specifies the plot to be represented by the legend. The plot is identified by the
name that is assigned to it on the plot statement’s NAME= option.

Restriction Unlike the DISCRETELEGEND statement, only a single graph-name

can be specified.

Requirement graph-name must be enclosed in quotation marks.

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.

NONE
does not automatically align the legend within its parent layout.

The legend’s position is set by the HALIGN= and VALIGN= options.

AUTO
within the parent layout, attempts to center the legend in the area that is
farthest from any surrounding data point markers.

(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the parent layout’s other graphics features.

The location-list is space-separated and can contain any of these locations:

TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT,
BOTTOM, and BOTTOMRIGHT.

Default NONE

Interactions This option has no effect unless LOCATION= INSIDE.

1294 Chapter 9 / Legend Statements

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type layout.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphLegendBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By

default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.

Defaults FALSE in SAS 9.4M1 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting

with SAS 9.4M2. If attribute FrameBorder is not defined in the active
style, then FALSE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

Tip The color of the frame around the color ramp and the color ramp tick
lines is controlled by the ContrastColor attribute of the
GraphAxisLines style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

“Line Options” on page 1602 for available line-options.

EXTRACTSCALE=TRUE | FALSE
specifies whether to extract a scale factor from the tick values and use it to
reduce the tick value width. The scale can be a named scale or a scientific-
notation scale. The EXTRACTSCALETYPE= option specifies the scale type. The
scale used is appended to the legend title as shown in the following example.
Total Sales (millions)
 CONTINUOUSLEGEND Statement 1295

For long legend titles, if the scale does not fit the available space, then the title
is truncated, and the scale is appended to the truncated title. Ellipses indicate
that the label was truncated as shown in the following example.
Total Sales for the Fourth Quarter Of ... (millions)

In extreme cases where the title does not fit even with truncation, the title is
dropped.

Restriction The scale that is extracted by the EXTRACTSCALE= option is

derived from the English locale.

Interactions The scale type is determined by the EXTRACTSCALETYPE= option.

If the axis label is not displayed, then the EXTRACTSCALE=TRUE

option is ignored.

See “boolean ” on page xii for other Boolean values that you can use.

EXTRACTSCALETYPE=DEFAULT | SCIENTIFIC
specifies whether to extract a named scale or a scientific-notation scale.

DEFAULT
extracts a named scale.

A named scale can be millions, billions, or trillions for values of 999 trillion or
less, or a multiple of 10 (denoted as 10^n) for values over 999 trillion. For
small fractional tick values, the scale factor is set to ensure that the
absolute value of the smallest value is greater than 1. The scale can be
millionth, billionth, or trillionth for values of 1 trillionth or more, or a multiple
of 1/10 (10^–n) for values less than 1 trillionth.

SCIENTIFIC
extracts a scientific-notation scale.

A scientific-notation scale is a multiple of 10 expressed as 10^n for values

greater than 1, or a multiple of 1/10 expressed as 10^–n for values less than 1.

HALIGN=CENTER | LEFT | RIGHT | number

specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.

number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is
all the way to the left and 1 is all the way to the right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. HALIGN=0

corresponds to the containing layout's left edge plus half the
legend width. Similarly, HALIGN=1 corresponds to the containing
layout's right edge minus half the legend width.
1296 Chapter 9 / Legend Statements

Defaults If LOCATION=OUTSIDE, then the default is RIGHT.

If LOCATION=INSIDE, then the default is CENTER.

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN= options

cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

INTEGER=TRUE | FALSE
specifies whether only integer tick values are used in the continuous legend.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default FALSE

Restriction This option applies only to smooth color gradients. For leveled
gradients such as those used with contour plots of type FILL,
LINEFILL or LABELEDLINEFILL, this option is ignored. See
CONTOURTYPE=.

Interaction INTEGER=TRUE is ignored when there are no integer values

between the minimum and maximum legend values.

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or REGION layout.

Default OUTSIDE

Restriction This option has effect only when the legend statement appears
within a 2-D overlay-type layout and there is at least one
standalone plot statement with XY axes that is referenced by the
legend.

Interactions The actual position is determined by the settings for the

LOCATION=, AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
number setting for the alignment is invalid when the legend is
positioned outside of the plot area.

Within an overlay-type layout, if the ORIENT= option is not set,

then the orientation changes depending on the actual position. If
 CONTINUOUSLEGEND Statement 1297

LOCATION=OUTSIDE and the legend is right or left of the plot,

then the orientation is vertical. If LOCATION=OUTSIDE and the
legend is above or below the plot, then the orientation is horizontal.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORIENT=VERTICAL | HORIZONTAL
specifies the orientation of the legend.

Default VERTICAL

Restriction In order to use ORIENT=HORIZONTAL when the continuous legend

is in an OVERLAY or REGION layout, you must place the legend
inside the graph area (LOCATION=INSIDE).

Tip To orient the legend horizontally outside of the graph area in an

OVERLAY or REGION layout, use LOCATION=OUTSIDE and
VALIGN=BOTTOM instead.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.
1298 Chapter 9 / Legend Statements

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend perimeter.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page xiii

TITLE="string"
specifies the title of the legend. The title is placed below the legend body.

Default No title

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
 CONTINUOUSLEGEND Statement 1299

between the title and the legend. If TITLEBORDER=FALSE is in

effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.

number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is
on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds

to the containing layout's bottom edge plus half the legend
height. Similarly, one corresponds to the containing layout's top
edge minus half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout. It is ignored if the parent layout
is not an overlay-type or REGION layout.

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN= options

cannot both be set to CENTER.
1300 Chapter 9 / Legend Statements

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VALUECOUNTHINT=positive-integer
recommends a number of values for the continuous legend to use to label the
data range.

Default 6

Restriction The associated plot must be displayed with smooth gradient for
this option to have any effect. For example, in a contour plot,
CONTOURTYPE must be set to GRADIENT or LINEGRADIENT.

Requirement positive-integer must be greater than zero.

Note The legend uses even intervals to label the range.

Details
A continuous legend consists of a color ramp and a numeric scale indicating color
values.

In the following figure, the continuous legend references a contour plot with a fixed
number of levels. The color ramp and legend values automatically reflect these
discrete levels. The legend option VALUECOUNTHINT= has no effect.
 CONTINUOUSLEGEND Statement 1301

In this next figure, the continuous legend references a contour plot with a
continuous gradient. The number of legend values displayed is automatically
determined by the legend, or decided by the contributing plot, such as the
CONTOURPLOTPARM with CONTOURTYPE=FILL. For plots with a continuous
gradient, you can use the legend option VALUECOUNTHINT= to control how many
legend values appear. (This option is ignored if there is no gradient.)

Within an overlay-type layout, when a continuous legend is placed inside the plot
area with LOCATION= INSIDE,
n It is always placed on top of plot lines and markers.

n By default, its background is fully transparent ( OPAQUE= FALSE), meaning that

underlying lines, markers, and data labels show through the legend.
n Its position can be controlled with the AUTOALIGN= option, or with the
HALIGN= and VALIGN= options.

Within an overlay-type layout, when a continuous legend is placed outside the plot
area with LOCATION=OUTSIDE,
 1302 Chapter 9 / Legend Statements

n By default, its background is fully opaque (OPAQUE=TRUE).

n Its position can be controlled with the HALIGN= and VALIGN= options.

Example: CONTINUOUSLEGEND Statement

The following graph was generated by the “Example Program” on page 1302:

Example Program
proc template;
define statgraph continuouslegend;
begingraph;
entrytitle "Height and Weight Distribution";
layout overlay;
scatterplot x=height y=weight /
markercolorgradient=density
markerattrs=(symbol=squarefilled size=6px)
name="scatter";
continuouslegend "scatter" / orient=vertical
location=outside valign=center halign=right
valuecounthint=10 title="Density";
endlayout;
endgraph;
end;
run;
ods graphics / reset width=475px;
 DISCRETELEGEND, MERGEDLEGEND Statements 1303

proc sgrender data=sashelp.gridded(where=(count>0))

template=continuouslegend;
run;

DISCRETELEGEND, MERGEDLEGEND Statements

Creates a legend with entries that refer to plots, or group values, or both legend items.

Restrictions: The MERGEDLEGEND statement can be used for grouped plots only.
The MERGEDLEGEND statement supports only line and marker overlays.
Notes: Often the data in a plot is classified by a group variable. Or, multiple columns of
data are plotted in the same graph. These groups or columns are represented in the
graph by different color or line patterns or marker symbols. In these cases, a
DISCRETELEGEND can be added to the graph to help decode the data. The
MERGEDLEGEND statement can be used to consolidate legend entries when the
graph displays grouped data for two plots. The MERGEDLEGEND statement must
specify exactly two names that reference the source for the legend entry values.
Prior to SAS 9.4M3, when a discrete attribute map is used for group values in a plot
that contributes to a discrete legend and attributes are overridden in the plot
statement, the attributes of some plot features and their corresponding legend
items might not match. Starting with SAS 9.4M3, the attributes of the legend items
always match the attributes of the corresponding plot features.
See: “LEGENDITEM Statement” on page 1322

Syntax
DISCRETELEGEND "graph-name-1" | "legend-item-name-1" | "discrete-attr-name-1"
<"graph-name-2" | "legend-item-name-2" | "discrete-attr-name-2"…> </options>;

MERGEDLEGEND "graph-name" | "legend-item-name" | "discrete-attr-name"

"graph-name" | "legend-item-name" | "discrete-attr-name"
</options>;

Summary of Optional Arguments

Appearance options
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before
the next row begins.
AUTOITEMSIZE=TRUE | FALSE
specifies that all markers, lines, and filled symbols in the legend are sized
in proportion to the font size used for the legend entry labels.
BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
1304 Chapter 9 / Legend Statements

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the legend.
DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend
cannot be rendered entirely.
DOWN=positive-integer
specifies the number of legend entries that are placed vertically before
the next column begins.
FILLITEMOUTLINE=AUTO | ON
specifies whether the fill swatches are outlined only when enabled by the
contributing statements or are always outlined.
ITEMSIZE=(size-options)
specifies the size of specific types of items that are in a discrete or
merged legend.
OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or
transparent (FALSE).
ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space that is added outside the legend
border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend
perimeter.
SORTBY=LABEL | TEXT
specifies whether text legend items are sorted by label or by text.
SORTORDER=AUTO | REVERSEAUTO | ASCENDINGFORMATTED |
DESCENDINGFORMATTED
specifies the sort order to use for the legend entry labels.
TITLEBORDER=TRUE | FALSE
specifies a border around the legend title that separates it from the
legend body.
VALUEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend values.

Content options
ADDITIONALNAMES=("graph-name" | "legend-item-name" | "discrete-attr-name"
<"graph-name" | "legend-item-name" | "discrete-attr-name"> …)
specifies additional legend items that are to be added to the two items
that are required in the MERGEDLEGEND statement.
EXCLUDE=(item-names)
specifies a list of legend entries to exclude from the display.
TYPE=ALL | FILL | FILLCOLOR | LINE | LINECOLOR | LINEPATTERN | MARKER
| MARKERCOLOR | MARKERSYMBOL | TEXT
specifies which visual attributes to display for legend entries in the
legend.
 DISCRETELEGEND, MERGEDLEGEND Statements 1305

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the legend is automatically aligned within its parent
layout when nested within an overlay-type layout.
HALIGN=CENTER | LEFT | RIGHT | number
specifies the horizontal alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.
LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when
the legend is specified within an overlay-type or REGION layout.
VALIGN=CENTER | TOP | BOTTOM | number
specifies the vertical alignment of the legend within its parent layout
when nested within an overlay-type or REGION layout.

Text options
TITLE="string"
specifies the title of the legend.
TITLEATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font attributes of the legend title.

Required Arguments
"graph-name"
references one or more unique names that are specified by the NAME= option of
a plot statement. These names control which plots contribute to the legend, and
the order in which the legend entries occur. If a contributing plot uses a
GROUP= option, then there is a legend entry for each group value that is present
in the data.

Requirements Each graph-name must be enclosed in quotation marks.

Each plot that is referenced must be able to be identified by the

legend entries. For example, a filled CONTOURPLOTPARM plot
cannot be specified in the DISCRETELEGEND or
MERGEDLEGEND statement because it requires a continuous
legend.

"legend-item-name"
references one or more unique values specified by the NAME= option of a
LEGENDITEM statement. Each legend-item-name must be enclosed in quotation
marks.

See “LEGENDITEM Statement” on page 1322

"discrete-attr-name"
references one or more unique values that are specified by the NAME= option in
a DISCRETEATTRMAP statement. The discrete attribute map that the name
references contributes all of its value statements as legend items, regardless of
whether they match the data.

Restriction When one or more discrete attribute map names (discrete-attr-

name) are specified in the DISCRETELEGEND statement, TYPE=ALL
1306 Chapter 9 / Legend Statements

on page 1317 (default) is not supported. If TYPE=ALL is in effect in

that case, a discrete legend is not generated. To create a discrete
legend for a discrete attribute map, you must also specify an
attribute other than ALL in the TYPE= option in the
DISCRETELEGEND statement.

See “DISCRETEATTRMAP Statement” on page 1505

Optional Arguments
ACROSS=positive-integer
specifies the number of legend entries that are placed horizontally before the
next row begins. A legend entry typically consists of two parts, such as a marker
symbol and an associated value.

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Interactions This option is ignored if ORDER= COLUMNMAJOR

This option is ignored when the DISCRETELEGEND statement is in

a LAYOUT GLOBALLEGEND block and the TYPE=ROW option is in
effect for the layout.

Tip When a DISCRETELEGEND statement is placed in a nested layout,

include the ACROSS= option to specify the number of columns.
Otherwise, the legend items might wrap in an unexpected fashion.

ADDITIONALNAMES=("graph-name" | "legend-item-name" | "discrete-attr-name"

<"graph-name" | "legend-item-name" | "discrete-attr-name"> …)
specifies additional legend items that are to be added to the two items that are
required in the MERGEDLEGEND statement.

Note: This option is valid starting in SAS 9.4M1.

The items from the sources that are specified in this option are not included in
the merging of the legend items. They are appended to the legend after the
items from the sources that are specified in the required arguments are merged.

Restriction This option applies to the MERGEDLEGEND statement only.

Requirements The list of names must be enclosed in parentheses.

Each name must be enclosed in quotation marks and separated

from adjacent values by a blank space.

See “"graph-name"” on page 1305

“"legend-item-name"” on page 1305

“"discrete-attr-name"” on page 1305

 DISCRETELEGEND, MERGEDLEGEND Statements 1307

AUTOALIGN=NONE | AUTO | (location-list)

specifies whether the legend is automatically aligned within its parent layout
when nested within an overlay-type layout.

NONE
does not automatically align the legend within its parent layout.

The legend’s position is set by the HALIGN= and VALIGN= options.

AUTO
within the parent layout, attempts to center the legend in the area that is
farthest from any surrounding data point markers.

(location-list)
within the parent layout, restricts the legend’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the parent layout’s other graphics features.

The location-list is space-separated and can contain any of these locations:

TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT,
BOTTOM, and BOTTOMRIGHT.

Default NONE

Restriction AUTO is available only when the parent layout contains a scatter
plot. Otherwise, it is ignored.

Interactions This option has no effect unless LOCATION= INSIDE.

When LOCATION=INSIDE and AUTOALIGN= is not NONE, the

HALIGN= and VALIGN= options are ignored.

Note Prior to SAS 9.4M5, the legend might overlap other legends or
insets. Starting with SAS 9.4M5, alignments that cause the legend
to overlap another legend or inset are avoided.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type layout.

AUTOITEMSIZE=TRUE | FALSE
specifies that all markers, lines, and filled symbols in the legend are sized in
proportion to the font size used for the legend entry labels. These proportional
sizes take effect regardless of the size reported by the plot or LEGENDITEM.
The line segments are drawn as deemed appropriate by the legend, regardless of
the line thickness reported by the plot.

Default FALSE

Interaction When set to TRUE, this setting considers the font size in effect from
the VALUEATTRS= option.

See “boolean ” on page xii for other Boolean values that you can use.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the legend background.
1308 Chapter 9 / Legend Statements

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphLegendBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect for the color to be seen. By

default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the legend.

Defaults TRUE in SAS 9.4M1 and earlier releases.

The GraphLegendBackground:FrameBorder style reference starting

with SAS 9.4M2. If attribute FrameBorder is not defined in the active
style, then TRUE is the default value.

Tip The BORDERATTRS= option controls the appearance of the legend

border.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the legend.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element

“Line Options” on page 1602 for available line-options.

DISPLAYCLIPPED=TRUE | FALSE
specifies whether the legend is displayed when any portion of the legend cannot
be rendered entirely. Based on the legend contents and placement, or when the
graph size is reduced, parts of the legend (title, legend symbol, or legend value)
might be clipped (truncated). When clipping occurs and this option is set to
FALSE, the entire legend is removed from the graph. The space for that legend is
then reclaimed by the remainder of the graph. When this option is set to TRUE,
the legend always appears, even if some parts of it have been clipped.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

DOWN=positive-integer
specifies the number of legend entries that are placed vertically before the next
column begins. A legend entry typically consists of two parts, such as a marker
symbol and an associated value.
 DISCRETELEGEND, MERGEDLEGEND Statements 1309

Default The entries are placed to best fit the available area. This “best fit”
approach works only when the legend is nested in the template’s
outermost layout.

Restriction This option is ignored if ORDER= ROWMAJOR

Interaction This option is ignored when the DISCRETELEGEND statement is in a

LAYOUT GLOBALLEGEND block.

EXCLUDE=(item-names)
specifies a list of legend entries to exclude from the display.

Default No items are excluded.

Requirement Each item name must be enclosed in quotation marks and

separated from adjacent names by a space.

Note When the specified names are compared with the legend entry
names, leading blanks are honored and trailing blanks are ignored.

Tip For plots with groups, you can exclude specific group values.

Example The following example excludes items Truck and Wagon from the
legend.
exclude=("Truck" "Wagon")

FILLITEMOUTLINE=AUTO | ON
specifies whether the fill swatches are outlined only when enabled by the
contributing statements or are always outlined.

Note: This feature applies to SAS 9.4M3 and to later releases.

AUTO
honors the DISPLAY= option settings for the contributing plot or the
FILLDISPLAY= setting of the legend item.

ON
always outlines the fill swatches.

Default ON

Restriction This option is valid only in the DISCRETELEGEND statement.

Requirement TYPE= on page 1317 must be FILL, FILLCOLOR, or ALL for this
option to have any effect.

Note The outlines are always 1 pixel wide with a solid pattern.

Tip To display color swatches with no outline when DISPLAY=

includes OUTLINE, specify FILLITEMOUTLINE=AUTO and
TYPE=FILLCOLOR.

HALIGN=CENTER | LEFT | RIGHT | number

specifies the horizontal alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.
1310 Chapter 9 / Legend Statements

number
specifies an explicit position in the containing layout.

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s width, where 0 is
all the way to the left and 1 is all the way to the right.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. HALIGN=0

corresponds to the containing layout's left edge plus half the
legend width. Similarly, HALIGN=1 corresponds to the containing
layout's right edge minus half the legend width.

Default CENTER

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout.

Interactions If LOCATION= OUTSIDE, then the HALIGN= and VALIGN= options

cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

ITEMSIZE=(size-options)
specifies the size of specific types of items that are in a discrete or merged
legend.

Note: This option is valid starting in SAS 9.4M1.

The following size options are supported:

FILLASPECTRATIO=AUTO | GOLDEN | positive-number

specifies the aspect ratio for the fill swatches.

Note: This feature applies to SAS 9.4M3 and to later releases.

AUTO
uses an equal width and height for color swatches with solid color fills, or
uses the golden ratio for swatches with pattern fills.

GOLDEN
specifies the golden ratio of 1.618 (width = 1.618 * height) for both solid
color and pattern fill swatches.
 DISCRETELEGEND, MERGEDLEGEND Statements 1311

positive-number
specifies a custom aspect ratio.

Default AUTO

Restrictions This option is valid only in the DISCRETELEGEND statement.

This option does not apply to bubble plot fill color swatches
(filled circle).

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=TRUE.

When FILLASPECTRATIO= is specified but neither

FILLHEIGHT= nor HEIGHTSCALE= is specified, the height for
color swatches with a solid fill and color swatches with a
pattern fill differ.

Note FILLHEIGHT=, HEIGHTSCALE=, and FILLASPECTRATIO= also

apply to fill-pattern color swatches with no outline.

Tip Use FILLHEIGHT= or HEIGHTSCALE= to specify the height.

FILLHEIGHT=AUTO | BIG | dimension | style-reference

specifies the height of the fill swatches.

Note: This feature applies to SAS 9.4M3 and to later releases.

AUTO
enables the SAS system to determine the fill color swatch height.

BIG
specifies sizes that were developed as defaults for SAS Visual Analytics,
which results in color swatches that are larger than the SAS system
defaults.

dimension
specifies a custom height for the fill swatches.

See “dimension” on page xiii

style-reference
specifies a style attribute that controls the height of the fill swatches.

The style reference must provide a valid dimensional value.

See “style-reference ” on page xv

Default AUTO

Restriction This option is valid only in the DISCRETELEGEND statement.

1312 Chapter 9 / Legend Statements

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=TRUE.

Tip Use FILLASPECTRATIO= to specify the width.

HEIGHTSCALE=positive-number
specifies a scale factor that is to be applied to the fill swatch height.

Values greater than 1 increase the height of the fill swatches, and values less
than 1 reduce the height.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default 1

Restriction This option is valid only in the DISCRETELEGEND statement.

Interactions The legend must display fill entries for this option to have any
effect.

This option is ignored when AUTOITEMSIZE=TRUE.

Tips Use FILLHEIGHT= to change the base height.

Use FILLASPECTRATIO= to specify the width.

LINELENGTH=dimension
specifies the length of the line glyph for line entries in the legend.

Default 48 pixels

Interaction The legend must display line entries for this option to have any
effect.

See “dimension” on page xiii

LOCATION=OUTSIDE | INSIDE
specifies whether the legend appears inside or outside the plot area when the
legend is specified within an overlay-type or REGION layout.

Default OUTSIDE

Restriction This option has effect only when the legend statement appears
within an overlay-type or REGION layout and at least one
standalone plot statement is referenced by the parent layout.

Interactions The actual position is determined by the settings for the

LOCATION=, AUTOALIGN= , HALIGN= , and VALIGN= options.

If this option is set to OUTSIDE, then the HALIGN= and VALIGN=

options must specify a keyword (LEFT, RIGHT, or CENTER). The
 DISCRETELEGEND, MERGEDLEGEND Statements 1313

number setting for the alignment is invalid when the legend is

positioned outside of the plot area.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

OPAQUE=TRUE | FALSE
specifies whether the legend background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

ORDER=ROWMAJOR | COLUMNMAJOR
specifies whether legend entries are organized into rows or into columns.

Default ROWMAJOR

Interaction If ORDER=ROWMAJOR, then use the ACROSS= option to limit the

number of entries in a row. If ORDER=COLUMNMAJOR, then use the
DOWN= option to limit the number of entries in a column.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space that is added outside the legend border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.
1314 Chapter 9 / Legend Statements

Tip Use pad-options to create non-uniform padding.

Default AUTO

Notes The default units for dimension are pixels.

Starting with SAS 9.4M1, the default padding between the discrete
legend and the plot area (including the axes) is increased to 10 pixels,
depending on the context. If the new default padding is not desirable,
then use the OUTERPAD= option to adjust it.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the legend perimeter.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the legend perimeter.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default Padding for all sides is 0.

Note The default units for dimension are pixels.

See “dimension” on page xiii

SORTBY=LABEL | TEXT
specifies whether text legend items are sorted by label or by text.

Note: This feature applies to SAS 9.4M3 and to later releases.

Default LABEL
 DISCRETELEGEND, MERGEDLEGEND Statements 1315

Restriction This option is valid only in the DISCRETELEGEND statement.

Interaction This option has an effect only when sorting is performed,

TYPE=TEXT is specified in this DISCRETELEGEND statement, and
the legend items are defined in either a LEGENDTEXTITEM
statement or in a LEGENDITEM statement that specifies
TYPE=TEXT.

SORTORDER=AUTO | REVERSEAUTO | ASCENDINGFORMATTED |

DESCENDINGFORMATTED
specifies the sort order to use for the legend entry labels.

AUTO
presents groups of legend entries in the order in which they are listed in the
legend statement.

The internal ordering of the entries is derived from the constituent plot-
statement options.

REVERSEAUTO
presents groups of legend entries in the reverse order in which they are listed
in the legend statement.

Note: This feature applies to SAS 9.4M5 and to later releases.

Tip This option is useful for grouped, stacked vertical bar charts when you
want the legend entries to be drawn in the same order as the bar group
segments.

ASCENDINGFORMATTED
DESCENDINGFORMATTED
performs a linguistic sort on the group items. It has the same effect as
sorting the input data, but the data is not changed.

Entries from separate plots, and separate legend items are combined and
ordered as a single list.

Default AUTO. Groups of legend entries are presented in the order in which
they are listed in the legend statement. The internal ordering of the
entries is derived from the constituent plot-statement options.

Interaction This option overrides the order that is set by any constituent plot
statement’s GROUPORDER= option.

TITLE="string"
specifies the title of the legend. The title is placed to the left of the legend body,
except in the following cases:
n The legend contains two or more rows of items.

n The legend is in a nested layout.

n The legend is in an OVERLAYEQUATED layout.

n The legend title length exceeds the space that is available on the left side of
the legend.
1316 Chapter 9 / Legend Statements

In those cases, the title is placed above the legend body.

Default No title

Requirement string must be enclosed in quotation marks.

Interaction When the title is placed to the left of the legend, if

TITLEBORDER=TRUE is in effect, no separator is displayed
between the title and the legend. If TITLEBORDER=FALSE is in
effect in that case, a separator is displayed.

TITLEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend title.

Default The GraphLabelText style element.

Interactions For this option to have any effect, the TITLE= option must also be
specified.

If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

TITLEBORDER=TRUE | FALSE
specifies a border around the legend title that separates it from the legend
body.

Default FALSE

Interactions Prior to SAS 9.4M3, when LOCATION=INSIDE, BORDER=TRUE and

the TITLE= option must also be in effect for this option to have any
effect. When LOCATION=OUTSIDE, the TITLE= option must also
be in effect. The BORDER= option has no effect on the title border
in that case.

Starting with SAS 9.4M3, BORDER=TRUE and the TITLE= option

must also be in effect for this option to have any effect. The
LOCATION= option has no effect on the title border in that case.

Tip The line attributes of the title border are set by the
BORDERATTRS= options.

See “boolean ” on page xii for other Boolean values that you can use.
 DISCRETELEGEND, MERGEDLEGEND Statements 1317

TYPE=ALL | FILL | FILLCOLOR | LINE | LINECOLOR | LINEPATTERN | MARKER |

MARKERCOLOR | MARKERSYMBOL | TEXT
specifies which visual attributes to display for legend entries in the legend.

Note: TEXT is valid starting with SAS 9.4M3.

The TYPE= option can be used as a filter. If a statement contributing to the

legend does not have any visual attributes that match the TYPE specified, then
the legend does not display any entries from that statement.

Some keywords can be used to create specialized legends that display a single
visual attribute. For example, keywords FILLCOLOR or MARKERSYMBOL result
in the display of a single attribute. Other keywords (for example, FILL, LINE, or
MARKER) result in legends that display a set of visual attributes. For example,
keyword LINE results in the display of both line color and line pattern for legend
entries that include lines in their display.

If this option is set to LINEPATTERN or MARKERSYMBOL, then a filled symbol

is drawn using the same text color as the color used for the legend entry labels.
The symbol is sized automatically, as if the AUTOITEMSIZE= option is set to
TRUE. For keywords FILLCOLOR, LINECOLOR, and MARKERCOLOR, the filled
symbols are drawn as outlined color swatches. The outline is 1 pixel wide, and
its color is controlled by the CONTRASTCOLOR attribute of the GraphOutlines
style element.

Default ALL

Restrictions This option is valid only in the DISCRETELEGEND statement.

When one or more discrete attribute map names (discrete-attr-

name on page 1305) are specified in the DISCRETELEGEND
statement, TYPE=ALL (default) is not supported. If TYPE=ALL is in
effect in that case, a discrete legend is not generated. To create a
discrete legend for a discrete attribute map, you must also specify
an attribute other than ALL in the TYPE= option in the
DISCRETELEGEND statement.

A LEGENDITEM statement can be referenced only from a discrete

legend of the same attribute type or of an overlapping attribute
type. For legends that display multiple visual attributes (use both
colors and marker symbols, for example), the default visual
properties are derived from the GraphDataDefault style-element.

Starting with SAS 9.4M3, the LEGENDTEXTITEMS statement can

be referenced only from a discrete legend of type TEXT or ALL.

VALIGN=CENTER | TOP | BOTTOM | number

specifies the vertical alignment of the legend within its parent layout when
nested within an overlay-type or REGION layout.

number
specifies an explicit position in the containing layout.
1318 Chapter 9 / Legend Statements

Range The number specification can range from 0 to 1. The number

represents a fraction of the parent container’s height, where 0 is
on the bottom and 1 is on the top.

Interaction For a number setting to take effect, LOCATION=INSIDE must be

set. A number setting is invalid on this option when
LOCATION=OUTSIDE.

Tip The legend is effectively anchored at its center. Zero corresponds

to the containing layout's bottom edge plus half the legend
height. Similarly, one corresponds to the containing layout's top
edge minus half the legend height.

Default CENTER

Restriction This option is available only when this statement is nested within
an overlay-type or REGION layout. It is ignored if the parent layout
is not an overlay-type or REGION layout.

Interactions If LOCATION= OUTSIDE, then the VALIGN= and HALIGN= options

cannot both be set to CENTER.

This option is ignored when LOCATION=INSIDE, the AUTOALIGN=

option is not NONE, and the parent layout is an overlay-type
layout.

See LAYOUT OVERLAY on page 159 for more information about how
child positions are determined in an overlay-type or REGION
layout.

VALUEATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend values.

Default The GraphValueText style element.

Interaction If one or more text options are specified and they do not include all
the font properties (color, family, size, weight, style), then non-
specified properties will be derived from the GraphLabelText style
element.

Note When you specify style-element, only the style attributes COLOR,
FONTFAMILY, FONTSIZE, FONTSTYLE, and FONTWEIGHT are
used.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

 DISCRETELEGEND, MERGEDLEGEND Statements 1319

Details
A discrete legend consists of one or more units called legend entries. Each legend
entry consists of a legend symbol and a legend value. The legend symbol is
typically a marker, line, or filled symbol that represents a specific area in the plot.
The legend value is descriptive text that is derived from group values, or that is
assigned by the plot’s LEGENDLABEL= option or by the LEGENDITEM statement’s
LABEL= option. To specify a discrete legend, you can use the DISCRETELEGEND
statement. To consolidate legend entries for common grouped values that are
represented by two separate plots, you can use the MERGEDLEGEND statement.

For grouped plots, a discrete legend represents all of the values that are present in
the data. To ensure that legend entries are displayed for group values, regardless of
whether those values are present in the data, you can use a discrete attribute map
as described in “Displaying Legend Entries for Group Values That Are Not in the
Data” on page 1512.

In the DISCRETELEGEND statement, you can specify one or more names that
reference the source for the legend entry values. You can use the TYPE= option to
control the visual attributes for the legend display.

For legend items that represent fills or colors without a specific shape, a filled
symbol with a one-pixel-wide, solid outline is used to represent the data values. If
the feature being represented by the legend item has an outline, then the default
outline color is derived from the feature’s outline color. If the feature being
represented is an outline only or is a filled outlined marker, then the outline
thickness is derived from the plot. If the feature does not have an outline, then the
default outline color for its corresponding legend entry is derived from the
GraphOutline style element.

You can use the MERGEDLEGEND statement to consolidate lines and marker
symbols from discrete legend entries when a graph displays grouped values for
exactly two plots. With a MERGEDLEGEND statement, the legend values from the
group variables in two plots are compared. For each common value, the
corresponding legend lines and marker symbols are combined, and only one legend
entry is created for each matching set of group values. You can use the
MERGEDLEGEND statement only for grouped plots. You must specify two names
that reference the source for the legend entry values.

When a discrete legend is placed inside the plot area of an overlay type layout and
LOCATION= INSIDE, the following conditions apply:
n The discrete legend is always placed on top of plot lines and markers.

n By default, the background of the discrete legend is fully transparent

( OPAQUE= FALSE). Underlying lines, markers, and data labels show through
the legend.
n You can control the position of the discrete legend with the AUTOALIGN=
option, or with the HALIGN= and VALIGN= options. (The AUTOALIGN= option
is not available in a LAYOUT OVERLAY3D statement.)

When a discrete legend is placed outside the plot area of an overlay type layout
and LOCATION=OUTSIDE, the following conditions apply:
1320 Chapter 9 / Legend Statements

n By default, the background of the discrete legend is fully opaque

(OPAQUE=TRUE).
n You can control the position of the discrete legend with the HALIGN= and
VALIGN= options.

When a discrete legend is placed within nested layouts, you might need to do one
of the following to obtain the legend organization that you want:
n use the ACROSS= option, and also set ORDER= ROWMAJOR

n use the DOWN= option, and also set ORDER=COLUMNMAJOR

A legend might be dropped if the total legend area in the graph exceeds the
percentage that is set by the MAXLEGENDAREA= option in an ODS GRAPHICS ON
statement that is in effect for the output destination. A legend might also be
dropped if DISPLAYCLIPPED= FALSE and the full legend cannot be displayed.

Examples

Example 1: DISCRETELEGEND Statement

The following graph using the DISCRETELEGEND statement was generated by
“Example Program” . It displays two discrete legends, one that shows the
confidence limits for two ellipses and a second that shows the values for a
GROUP= variable:
 DISCRETELEGEND, MERGEDLEGEND Statements 1321

Example Program
proc template;
define statgraph discretelegend;
begingraph;
entrytitle "Prediction Ellipses";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth /
group=species name="s";
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.2
name="p80" legendlabel="80%"
outlineattrs=graphconfidence;
ellipse x=petallength y=petalwidth /
type=predicted alpha=0.05
name="p95" legendlabel="95%"
outlineattrs=graphconfidence2;
discretelegend "s" / title="Species:" ;
discretelegend "p80" "p95" / across=1
autoalign=(topleft) location=inside ;
endlayout;
entryfootnote halign=left "Fisher's Iris Data";
endgraph;
end;

proc sgrender data=sashelp.iris template=discretelegend;

run;

Example 2: MERGEDLEGEND Statement

The following graph using the MERGEDLEGEND statement was generated by
“Example Program” . In the template definition, a grouped scatter plot is overlaid
with a series plot for each group, and the two plots are referenced by a single
merged legend:
 1322 Chapter 9 / Legend Statements

Example Program
proc sort data=sashelp.class
out =class;
by sex;
run;

proc template;
define statgraph mergedLegend;
begingraph;
entrytitle "Linear Regression By Gender";
layout overlay;
scatterplot x=height y=weight / group=sex name="scat";
regressionplot x=height y=weight/ group=sex name="reg";
mergedlegend "scat" "reg" / border=true;
endlayout;
endgraph;
end;

proc sgrender data=class template=mergedLegend;

run;

LEGENDITEM Statement
Creates the definition for a legend item that can be included in a discrete legend.
 LEGENDITEM Statement 1323

Restriction: The LEGENDITEM statement is used with the DISCRETELEGEND and

MERGEDLEGEND statements only.
Requirements: The LEGENDITEM statement must be a child of the BEGINGRAPH block. It cannot
be embedded in any other GTL statement block.
The LEGENDITEM statement must appear in the global definition area of the
template between the BEGINGRAPH statement and the first LAYOUT statement.
Note: The LEGENDITEM statement creates the definition for a legend item that can be
included in a discrete legend.
See: “DISCRETELEGEND, MERGEDLEGEND Statements” on page 1303

Syntax
LEGENDITEM TYPE=type NAME="string" </options>;

Summary of Optional Arguments

Appearance options
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the fill when TYPE= is set to FILL.
FILLDISPLAY=STANDARD | ALL | (fill-option)
specifies whether the fill swatch for this legend item displays fill only or
displays fill and outline.
FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the appearance of the line when TYPE= is set to LINE or
MARKERLINE.
MARKERATTRS=style-element | style-element (marker-options) | (marker-
options)
specifies the appearance of the marker when TYPE= is set to MARKER or
MARKERLINE.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the color of the outline when TYPE= is set to FILL.
TEXT="string"
specifies the text to display when TYPE= is set to TEXT.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
specifies the font and color attributes of the string that is specified on
the TEXT= option when TYPE= is set to TEXT.

Label options
LABEL="string"
specifies a label to be used for the legend item.
LABELATTRS
specifies the color and font attributes of the legend item’s label.
1324 Chapter 9 / Legend Statements

Required Arguments
TYPE=FILL | MARKER | MARKERLINE | LINE | TEXT
specifies a type for the legend item.
FILL
specifies a fill (displayed as a filled square).

The appearance can be managed with the FILLATTRS= and

OUTLINEATTRS= options.

MARKER
specifies a marker.

The appearance can be managed with the MARKERATTRS= option.

MARKERLINE
specifies a marker and a line.

The appearance can be managed with the MARKERATTRS= and

LINEATTRS= options.

LINE
specifies a line. The appearance can be managed with the LINEATTRS=
option.

TEXT
specifies text that can be displayed in the legend area.

The text string is defined with the TEXT= option, and the text appearance
can be managed with the TEXTATTRS= option.

NAME="string"
assigns a name to the legend item for reference in a DISCRETELEGEND or
MERGEDLEGEND statement.

Restrictions The string is case sensitive and must define a unique name within
the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Optional Arguments
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the color of the fill when TYPE= is set to FILL.

Default The GraphDataDefault style element.

Interaction The Transparency attribute cannot be derived from the style

element, but it can be set with this option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

 LEGENDITEM Statement 1325

FILLDISPLAY=STANDARD | ALL | (fill-option)

specifies whether the fill swatch for this legend item displays fill only or
displays fill and outline.

Note: This feature applies to SAS 9.4M3 and to later releases.

STANDARD
ALL
displays the fill and outline.

(fill-option)
specifies a list of one or more fill options.

The list can include one or both of the following options:

FILL
displays only the fill.

OUTLINE
displays only the outline.

Default STANDARD

Requirement You must specify TYPE=FILL for this option to have any effect.

FILLEDOUTLINEDMARKERS=TRUE | FALSE
specifies whether markers are drawn with both fill and an outline.

TRUE
draws filled markers (marker symbols with the suffix FILLED) using both fill
and an outline.

Interaction When this option is TRUE, the marker fill is drawn using the
FILLATTRS= specification, and the outline is drawn using the
OUTLINEATTRS= specification.

FALSE
draws the markers using fill or an outline, but not both.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

LABEL="string"
specifies a label to be used for the legend item.

Default No label is displayed

Tip The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS
specifies the color and font attributes of the legend item’s label.

Default The GraphValueText style element.

1326 Chapter 9 / Legend Statements

Interaction For this option to take effect, the LABEL= option must also be
specified.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the appearance of the line when TYPE= is set to LINE or
MARKERLINE.

Default The GraphDataDefault style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the appearance of the marker when TYPE= is set to MARKER or
MARKERLINE.

Default The GraphDataDefault style element.

Interaction The Weight attribute cannot be derived from the style element, but
it can be set with this option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the color of the outline when TYPE= is set to FILL.

Default The GraphDataDefault style element.

Restriction This option uses only the color specification in the style element or
line options. The line pattern and line thickness specifications are
ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

TEXT="string"
specifies the text to display when TYPE= is set to TEXT. The font and color
attributes for the text are specified by the TEXTATTRS= option.

Default a blank space

 LEGENDITEM Statement 1327

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the font and color attributes of the string that is specified on the
TEXT= option when TYPE= is set to TEXT.

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

Details
The LEGENDITEM statement creates a definition for a legend item that can be
included in a legend. The item that you define is independent of the data and
enables you to customize the legend to enhance or replace a standard legend. For
example, to display annotation text within the legend area, you can define a
LEGENDITEM statement with TYPE=TEXT and specify the text in the TEXT=
option.

As demonstrated in the “Example Program” on page 1328, multiple LEGENDITEM

statements can be used to replace a plot statement’s data-driven legend by
defining custom legend items to display in the legend. This use enables you to
communicate information that is not in the data. For this use, you would define one
or more LEGENDITEM statements to specify legend attributes, and then reference
those items in your legend statement. You must not reference the plot statement
itself in the legend statement. Although no direct connection would exist between
the plot data and the legend, you could communicate the connection by managing
the visual attributes in both the plot and the legend.

The LEGENDITEM statement must be located within the BEGINGRAPH block but
outside of the outermost layout block. You can use multiple LEGENDITEM
statements to define multiple definitions. In that case, each definition specifies a
single legend entry and each item name must be referenced in the legend
statement.

Note: A single legend statement can reference multiple item names and also
multiple plot names.

When specifying attribute options for a particular LEGENDITEM statement,

options that do not apply to the specified TYPE= value are ignored. For example,
the MARKERATTRS= option sets the attributes for a marker and is useful if you set
TYPE=MARKER or TYPE=MARKERLINE. However, if TYPE= is set to a value that
does not display a marker symbol, then the MARKERATTRS= option is ignored.
 1328 Chapter 9 / Legend Statements

Example: LEGENDITEM Statement

The following graph was generated by the “Example Program” on page 1328. The
example specifies three LEGENDITEM statements to define graphical properties
for two marker symbols and a text string. The NAME= option in each LEGENDITEM
statement assigns a name to the definition. Those names are referenced in a
DISCRETELEGEND statement so that the definitions are displayed in the graph
legend. To correlate the legend with the scatter plot, the example creates an
attribute map that matches values M and F to the same graphical properties that
are specified in the LEGENDITEM statements. That attribute map is referenced in
the scatter plot.

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Team Tryouts: Height and Weight by Sex";
discreteattrmap name="symbols" / ignorecase=true
trimleading=true;
value "m" / markerattrs=(color=blue symbol=diamondfilled);
value "f" / markerattrs=(color=red symbol=circlefilled);
enddiscreteattrmap;
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";
legendItem type=marker name="m_marker" /
markerattrs=(color=blue symbol=diamondfilled)
label="70% of boys present" ;
legendItem type=marker name="f_marker" /
 LEGENDTEXTITEMS Statement 1329

markerattrs=(color=red symbol=circlefilled)
label="75% of girls present" ;
legendItem type=text name="status" /
text=" (5 health forms missing)" ;
layout overlay;
scatterplot x=height y=weight / group=groupmarkers;
discretelegend "m_marker" "f_marker" "status" /
autoitemsize=true;
endlayout;
endgraph;
end;
proc sgrender data=sashelp.class template=scatterplot;
run;

LEGENDTEXTITEMS Statement
Creates the definition for data-driven text items that can be included in a discrete legend.

Restrictions: The LEGENDTEXTITEMS statement is used only with the DISCRETELEGEND

statement.
Grouping is not supported.
The maximum number of items that the LEGENDTEXTITEMS statement can
contribute to the legend is 100. If this limit is exceeded, the LEGENDTEXTITEMS
statement is dropped. If no other statements contribute to the legend in that case,
the legend is not drawn.
Requirements: The LEGENDTEXTITEMS statement must be a child of the BEGINGRAPH block. It
cannot be embedded in any other GTL statement block.
The LEGENDTEXTITEMS statement must appear in the global definition area of
the template between the BEGINGRAPH statement and the first LAYOUT
statement.
The TYPE= option in the legend statement that references this statement must be
set to ALL or TEXT.
Note: This feature applies to SAS 9.4M3 and to later releases.
See: “DISCRETELEGEND, MERGEDLEGEND Statements” on page 1303

Syntax
LEGENDTEXTITEMS NAME="string" TEXT=column </options>;

Required Arguments
NAME="string"
assigns a name to the legend items for reference in a DISCRETELEGEND or
MERGEDLEGEND statement.
1330 Chapter 9 / Legend Statements

Restriction This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

TEXT=column
specifies the column in the plot data set that contains the text items.

Notes One item is added for each observation. Grouping is not supported.

The TEXT column should not contain missing values. A missing TEXT
column value is treated as if no text is specified. If an observation
contains a missing TEXT column value and a valid LABEL column value,
only the label value is added to the legend for that observation. If both
values are missing, nothing is added to the legend.

If the plot data is a SAS data set, the text items are arranged in data
order in the plot legend. If the plot data is a CAS in-memory table, the
text items are arranged in ascending order of the text column values.

Optional Arguments
LABEL=column
specifies the column that contains the labels for the legend items.

Default No labels are displayed for the items

Note Each observation that has a LABEL column value should have a
corresponding TEXT column value. If an observation contains a valid
LABEL column value and a missing TEXT column value, only the label
value is added to the legend for that observation.

Tip The font and color attributes for the label are specified by the
LABELATTRS= option.

LABELATTRS=style-element | style-element (text-options) | (text-options)

specifies the color and font attributes of the legend item labels.

Default The GraphValueText style element.

Note Space is reserved in the legend for the height of the legend label text
regardless of whether the LABEL= option is specified.

Tip If you are not using the LABEL= option, specify

LABELATTRS=(SIZE=0pt) to reclaim the space that is reserved for the
label text height.

See “General Syntax for Attribute Options” on page 1599 for the syntax for
using a style-element value.

“Text Options” on page 1606 for available text-options.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the font and color attributes of the text values that is specified on the
TEXT= option.
 LEGENDTEXTITEMS Statement 1331

Default The GraphValueText style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax for
using a style-element value.

“Text Options” on page 1606 for available text-options.

Details
The LEGENDTEXTITEMS statement creates a definition for data-driven text legend
items that can be included in a discrete legend. The items that you define in the
data are independent of the plot and enable you to customize the legend in order to
enhance or replace a standard legend. The text items are stored in the plot data set
in the column that is specified by the TEXT= option. The LEGENDTEXTITEMS
statement must be located within the BEGINGRAPH block but outside of the
outermost layout block. You reference the LEGENDTEXTITEMS statement in your
legend statement by the name specified in the NAME= option. You must not
reference the plot statement itself in the legend statement. To add items from
multiple columns, specify one LEGENDTEXTITEMS statement for each column,
and then reference all of the LEGENDTEXTITEMS statements in your legend
statement. Although no direct connection would exist between the plot data and
the legend, you could communicate the connection by managing the visual
attributes in both the plot and the legend.

Example: LEGENDTEXTITEMS Statement

The following graph was generated by the “Example Program” on page 1332. The
graph shows a scatter plot of height and weight by sex for 19 participants in a team
tryout. To reduce clutter in the plot, a numeric ID is used to label the marker for
each individual in the plot. A legend in the right margin of the graph displays the
participant name for each ID.
 1332 Chapter 9 / Legend Statements

Example Program
In this example program, a LEGENDTEXTITEMS statement is used to create the
legend of IDs and names in the right margin of the graph. The name IDLEGEND is
assigned to the LEGENDTEXTITEMS statement. The ID column provides the legend
text items, and the Name column provides a name as the label for each text item.
Here is the SAS code.
/* Assign a numeric ID to each name */
data class;
set sashelp.class;
id=_N_;
run;

/* Define the template for the graph */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Team Tryouts: Height and Weight by Sex";
legendtextItems name="idlegend" text=id / label=name;
layout lattice / columns=2 columnweights=(85 15);
layout overlay /
xaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot))
yaxisopts=(griddisplay=on
gridattrs=(color=lightgray pattern=dot));
scatterplot x=height y=weight / name="scatter"
group=sex datalabel=id;
discretelegend "scatter" / location=inside
autoalign=(bottom bottomright);
endlayout;
layout overlay;
 LEGENDTEXTITEMS Statement 1333

discretelegend "idlegend" /
title="ID/Name" titleattrs=(weight=bold)
valign=top border=false
order=rowmajor across=1;
endlayout;
endlayout;
endgraph;
end;

/* Render the graph */

proc sgrender data=class template=scatterplot;
run;
1334 Chapter 9 / Legend Statements
 1335

PART 7

Text Statements

Chapter 10
Managing Text Items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1337

Chapter 11
Text Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347
1336
 1337

10
Managing Text Items

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1337
Using Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1338
Available Prefix Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1338
Managing Horizontal Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1338
Managing Font Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339
Using Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341
Available Text Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341
Subscripting and Superscripting Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341
Using UNICODE Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1342
Rules for Unicode and Special Character Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . 1342
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1343
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1343
Lowercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1344
Uppercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1345
Special Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1346

Overview
The ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements all use the same
syntax for specifying one or more pieces of text called text items. For example, here
is the ENTRYTITLE syntax:

ENTRYTITLE text-item <…<text-item>> </options>;

Each text item can be specified using the following syntax:

<prefix-option…<prefix-option>> "string" | dynamic | character-expression | {text-

command}

The simplest specification for each statement is to specify a string in quotation

marks. For example, you might specify a graph title as follows:

entrytitle "Height and Weight by Sex";

 1338 Chapter 10 / Managing Text Items

In this example, the string “Height and Weight by Sex” is formatted as the graph
title and displayed in the title position. If multiple strings are specified, then they
are concatenated into a single line of text. This second specification generates the
same title:

entrytitle "Height " "and " "Weight " "by " "Sex";

To provide control over the text, multiple prefix options can precede each text item,
and the text item can be a string literal, a dynamic, or a text command. All text
items with the same HALIGN= setting are concatenated into one string. Up to three
strings with different horizontal alignment can result for each ENTRY statement.
Leading and trailing blanks in the concatenated string are always used.
n a string must be enclosed in quotation marks.

n a character expression must be enclosed in an EVAL function.

n a text command must be enclosed in braces. (See “Using Text Commands” on

page 1341).

Using Prefix Options

Available Prefix Options

The following prefix options are available on ENTRY, ENTRYTITLE, and
ENTRYFOOTNOTE statements:
HALIGN
specifies the horizontal alignment of a text item.
TEXTATTRS
specifies font attributes for a text item.

When used, a prefix option applies not only to immediately following piece of text
but also to ALL subsequent text strings and text-commands. If the same prefix
option appears more than once, then each use overrides the last used prefix of the
same name.

Managing Horizontal Alignment

For the ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements, the default
horizontal alignment is CENTER.
 Using Prefix Options 1339

entry "One" " Two" " Three" " Four" " Five" " Six";

The HALIGN= option can be used to change the horizontal alignment to LEFT,
CENTER, or RIGHT. The following specification left-justifies the text:

entry halign=left
"One" " Two" " Three" " Four" " Five" " Six";

Text items are positionally concatenated by alignment area. For example, the
following specification left-justifies the first three strings and right-justifies the last
three strings:

entry halign=left "One" " Two" " Three"

halign=right "Four" " Five" " Six";

Even if the HALIGN= specifications are jumbled, the final text is nevertheless
positionally concatenated by alignment area:

entry halign=right "Five" halign=left "One"

halign=right " Six" halign=center "Three"
halign=left " Two" halign=center " Four";

Note: When long strings are used or the bounding container is constrained, the
alignment areas might overlap.

Managing Font Attributes

For the ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements, the default font
attributes are determined by the active ODS style. The TEXTATTRS= option can be
used to change the default font attributes. The TEXTATTRS= option syntax is as
follows (see “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style element and “Text Options” on page 1606 for available text options):

TEXTATTRS=style-element | style-element(text-options) | (text-options)

For example, the following ENTRYTITLE statement uses prefix options to create
this title line:
1340 Chapter 10 / Managing Text Items

entrytitle textattrs=(color=black) "Center"

textattrs=(color=red) " Text"
halign=right textattrs=(color=black size=10pt) "Right "
textattrs=(color=red size=10pt) "side"
halign=left
textattrs=(color=black style=italic size=10pt) "Left"
textattrs=(color=red style=italic size=10pt) " side";

n The TEXTATTRS= options are reset each time a new TEXTATTRS= appears—
there is no “carry over” or accumulation of the settings.
n Blanks (spaces) must be provided as needed to achieve the desired
concatenation. When concatenating dynamic variables that are stripped of
leading and trailing blanks, a literal space must be inserted to separate them, as
shown in the following example:

entry _DYN1 " " _DYN2;

Alternatively, font attributes can be specified by overriding the style element

defaults. The following specification overrides the settings of the GraphTitleText
style element, which sets the default attributes for ENTRYTITLE text:

entrytitle
halign=left
textattrs=GraphTitleText(color=black style=italic) "Left"
textattrs=GraphTitleText(color=red style=italic) " side"
halign=center
textattrs=(color=black) "Center"
textattrs=(color=red) " Text"
halign=right
textattrs=GraphTitleText(color=black weight=bold) "Right "
textattrs=GraphTitleText(color=red weight=bold) "side" ;

Dynamic variables can also be used in the text strings. In the following
ENTRYTITLE statement, assume that _DEPLABEL and _MODELLABEL are dynamic
variables that are specified on PROC TEMPLATE’s DYNAMIC statement:

entrytitle "Residual by Predicted for " _DEPLABEL

halign=left textattrs=GraphTitleText _MODELLABEL /
pad=(bottom=5);

n The default style element for ENTRYTITLE is GraphTitleText, so all three text
items (one literal and two dynamic variables) get these font properties as a
starting point.
n The text "Residual by Predicted for " _DEPLABEL is center-aligned by default.

n The prefix options HALIGN= and TEXTATTRS= override the center alignment
and font properties for the text _MODELLABEL.

This could have been coded as follows:

entrytitle halign=left
textattrs=GraphTitleText _MODELLABEL
halign=center textattrs=()
"Residual by Predicted for " _DEPLABEL;
 Using Text Commands 1341

In this example, the second HALIGN= and TEXTATTRS=() are necessary to reset
alignment and font properties to their defaults.

The string length of the resolved dynamic _MODELLABEL does not affect the
placement of the center-aligned text.

Using Text Commands

Available Text Commands

Text commands on ENTRY, ENTRYTITLE, and ENTRYFOOTNOTE statements are
special in-line instructions that either modify the appearance of the text or script
special characters. The following text commands are available:
{SUB}
specifies that the string or dynamic appears as a subscript.
{SUP}
specifies that the string or dynamic appears as a superscript.
{UNICODE}
specifies a glyph (graphical character) to be displayed using its Unicode
specification or keyword equivalent.

The general form of a text command is

{command argument(s)}

The opening and closing braces are required to denote the scope of the command.

Subscripting and Superscripting Text

The {SUB} and {SUP} text commands are used to subscript and superscript text.
Each of these text commands accepts a string or a dynamic for its argument(s).

In the following example, if _RSQUARE is an existing dynamic that resolves to the

value 0.7434, then the following ENTRY specification superscripts the string value
"2" to generate this text:

entry textattrs=(weight=bold) "R" {sup "2"}

textattrs=() "=" _RSQUARE;

The textattrs=() option cancels all style overrides and reverts to the default text
properties.
 1342 Chapter 10 / Managing Text Items

Using UNICODE Text

The {UNICODE} text command places special characters into the text and accepts
any of the following for its argument(s):
n a hexadecimal Unicode Code Point for a character (for example, "03B1"x)

n a reserved keyword for a commonly used code point (for example, BETA)

n a dynamic that resolves to a hexadecimal value or keyword.

The tables in “Reserved Keywords and Unicode Values” on page 1343 provide a list
of the commonly used reserved keywords and Unicode values (the tables are not
complete, but they provide an idea about what is possible).

Multiple arguments can be used within the scope of a single UNICODE text
command. For example, the following two specifications are equivalent:

{unicode "03b1"x beta}

{unicode "03b1"x} {unicode beta}

The default formatting for the UNICODE text is derived from the
GraphUnicodeText style element.

In the following example, if _ALPHAVAL is an existing dynamic that resolves to the

value 0.05, then the following ENTRY specification generates this text:

entry {unicode alpha} " = " _ALPHAVAL;

By combining the TEXTATTRS= prefix option with the {SUB} and {UNICODE} text
commands, you can generate the following text:

Entry textattrs=(style=italic) "E(Y)" textattrs=() " = "

{unicode beta} {sub "0"} " + "
{unicode beta} {sub "1"} "x" {sub "1"}
" + " {unicode beta} {sub "2"} "x" {sub "2"};

Rules for Unicode and Special Character

Specifications
The following rules apply to Unicode and special character specifications in ODS
graphics:
 Reserved Keywords and Unicode Values 1343

n Each character can be specified by looking up its code and specifying it as a

hexadecimal constant. Example: {unicode ’221e’x}.
n Lowercase Greek letters can be specified by using names instead of
hexadecimal constants. Example: {unicode alpha}.
n Uppercase Greek letters can be specified by using names followed by _u instead
of a hexadecimal constant. Example: {unicode alpha_u}.
n Superscript and subscript have special abbreviations. Examples: {sup 2} and
{sub 2}.
n The {SUP} and {SUB} specifications must not appear escaped and in quotation
marks in GTL. They must appear outside of quotation marks.
n Some characters overprint the character that comes before. Example: ’El nin’
{tilde} ’o’, which is equivalent to ’El nin’ {unicode ’0303’x} ’o’
creates ‘El niño’.
n Specifications inside quotation marks are escaped. Example: "(*ESC*){unicode
beta}".
n Specifications outside quotation marks are not escaped. Example: {unicode
beta}.

For more information about using text throughout GTL (for example, using Unicode
values in labels), see SAS Graph Template Language: User’s Guide.

Reserved Keywords and Unicode Values

Overview
The tables in this section show some of the reserved keywords and Unicode values
that can be used with the UNICODE text command. For information about
rendering Unicode characters, see “Managing the String on Text Statements” in SAS
Graph Template Language: User’s Guide.

Note the following:

n Keywords and Unicode values are not case-sensitive: "03B1"x is the same code
point as "03b1"x.
n The word blank is the keyword for a blank space.
 1344 Chapter 10 / Managing Text Items

Lowercase Greek Letters

Keyword Glyph Unicode Description

alpha α 03B1 lowercase alpha

beta β 03B2 lowercase beta

gamma γ 03B3 lowercase gamma

delta δ 03B4 lowercase delta

epsilon ε 03B5 lowercase epsilon

zeta ζ 03B6 lowercase zeta

eta η 03B7 lowercase eta

theta θ 03B8 lowercase theta

iota ι 03B9 lowercase iota

kappa κ 03BA lowercase kappa

lambda λ 03BB lowercase lambda

mu μ 03BC lowercase mu

nu ν 03BD lowercase nu

xi ξ 03BE lowercase xi

omicron ο 03BF lowercase omicron

pi π 03C0 lowercase pi

rho ρ 03C1 lowercase rho

sigma σ 03C3 lowercase sigma

tau τ 03C4 lowercase tau

upsilon υ 03C5 lowercase upsilon

phi φ 03C6 lowercase phi

 Reserved Keywords and Unicode Values 1345

Keyword Glyph Unicode Description

chi χ 03C7 lowercase chi

psi ψ 03C8 lowercase psi

omega ω 03C9 lowercase omega

Uppercase Greek Letters

Keyword Glyph Unicode Description

alpha_u Α 0391 uppercase alpha

beta_u Β 0392 uppercase beta

gamma_u Γ 0393 uppercase gamma

delta_u Δ 0394 uppercase delta

epsilon_u Ε 0395 uppercase epsilon

zeta_u Ζ 0396 uppercase zeta

eta_u Η 0397 uppercase eta

theta_u Θ 0398 uppercase theta

iota_u Ι 0399 uppercase iota

kappa_u Κ 039A uppercase kappa

lambda_u Λ 039B uppercase lambda

mu_u Μ 039C uppercase mu

nu_u Ν 039D uppercase nu

xi_u Ξ 039E uppercase xi

omicron_u Ο 039F uppercase omicron

pi_u Π 03A0 uppercase pi

 1346 Chapter 10 / Managing Text Items

Keyword Glyph Unicode Description

rho_u Ρ 03A1 uppercase rho

sigma_u Σ 03A3 uppercase sigma

tau_u Τ 03A4 uppercase theta

upsilon_u Υ 03A5 uppercase upsilon

phi_u Φ 03A6 uppercase phi

chi_u Χ 03A7 uppercase chi

psi_u Ψ 03A8 uppercase psi

omega_u Ω 03A9 uppercase omega

Special Characters
Keyword Glyph Unicode Description

prime ´ 00B4 single prime sign

bar ̅ 0305 combining overline1

bar2 ̿ 033F combining double overline1

tilde ̃ 0303 combining tilde1

hat ̂ 0302 combining circumflex accent1

1 This is an overstriking character that requires a Unicode font to render properly.

 1347

11
Text Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347
ENTRY Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347
ENTRYFOOTNOTE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1356
ENTRYTITLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1364

Dictionary

ENTRY Statement
Displays a line of text in the graphical area.

Requirement: An ENTRY statement must be specified within a LAYOUT, HEADER, SIDEBAR, or

CELL statement block.

Syntax
ENTRY text-item <text-item …> </options>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
1348 Chapter 11 / Text Statements

specifies whether a border is displayed around the text.

BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry
border.
ROTATE=0 | 90 | 180 | 270
specifies the angle of text rotation measured in degrees.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-
option, specifies the properties of individual text-items.

Location options
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the text is automatically aligned within its parent when
nested within an overlay-type layout.
VALIGN=CENTER | TOP | BOTTOM
specifies the vertical alignment of the text.

Text options
HALIGN=CENTER | LEFT | RIGHT
as a statement option, specifies the horizontal alignment of the text. As a
prefix-option, specifies the horizontal alignment of individual text-items.

Required Argument
text-item <…<text-item>>
specifies one or more pieces of text for the entry. Each text-item has the
following form:

<prefix-option <… prefix-option>>"string" | dynamic | character-expression |

{text-command}

Each piece of text can have multiple prefix settings that precede the piece of
text. A piece of text is either a string literal, a dynamic, or a text command. All
text-items with the same HALIGN= setting are concatenated into one string. Up
to three strings can result for each ENTRY statement. Leading and trailing
blanks in the concatenated string are always used.

When a prefix option is used, it applies not only to the text that immediately
follows the prefix option, but also to all subsequent text strings and text
commands. If the same prefix option appears more than once, then that prefix
overrides the last used prefix of the same name for the subsequent text strings
and text commands.
 ENTRY Statement 1349

Restriction The maximum length for the entry text is 512 characters in SAS
9.4 and earlier releases. Starting with SAS 9.4M1, this restriction is
removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1337

Optional Arguments
AUTOALIGN=NONE | AUTO | (location-list)
specifies whether the text is automatically aligned within its parent when
nested within an overlay-type layout. For more information about how child
positions are determined in an overlay-type layout, see the “LAYOUT OVERLAY
Statement” on page 159.

NONE
does not automatically align the text within the area.

Alignment is set with HALIGN= and VALIGN= options.

AUTO
within the parent layout, attempts to center the text in the area that is
farthest from any surrounding data point markers.

Restriction This option is available only if the parent layout contains a

scatter plot. Otherwise, it is ignored.

(location-list)
within the available area, restricts the text’s possible locations to those
locations in the specified location-list, and uses the location-list position
that least collides with the other graphics features in the area.

The location-list is space-separated and can contain any of these locations:

TOPLEFT, TOP, TOPRIGHT, LEFT, CENTER, RIGHT, BOTTOMLEFT,
BOTTOM, and BOTTOMRIGHT.

Example: (TOPRIGHT TOPLEFT)

Default NONE

Requirement For this option to take effect, the ENTRY statement must be in a
LAYOUT OVERLAY or LAYOUTOVERLAYEQUATED block.

Interaction When AUTOALIGN= is not NONE and the parent layout is an

overlay-type layout, the HALIGN= and the VALIGN= options are
ignored.

BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
1350 Chapter 11 / Text Statements

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect in order for the color to be seen.
By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

HALIGN=CENTER | LEFT | RIGHT

as a statement option, specifies the horizontal alignment of the text. As a prefix-
option, specifies the horizontal alignment of individual text-items. Each text-
item has a horizontal alignment, and text-items with the same alignment are
always grouped together.

Default CENTER

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.

AUTO
specifies that the default outside padding for this component is used.
 ENTRY Statement 1351

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.
1352 Chapter 11 / Text Statements

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Defaults Prior to SAS 9.4M6, padding is LEFT=3, RIGHT=3, TOP=0, and

BOTTOM=0.

Starting with SAS 9.4M6, if ROTATE=0 on page 1352, the default

padding is LEFT=3, RIGHT=3, TOP=0, and BOTTOM=0. If ROTATE= is
other than 0, the default padding is LEFT=0, RIGHT=0, TOP=3, and
BOTTOM=3.

Note The default units for dimension are pixels.

See “dimension” on page xiii

ROTATE=0 | 90 | 180 | 270

specifies the angle of text rotation measured in degrees. The angle is measured
from a horizontal line passing through the middle of the bounding box of the
text, counter-clockwise starting at the center of the bounding box.

Default 0. No rotation is performed.

Restriction Only angles of 0, 90, 180, or 270 degrees are allowed.

Interaction The bounding box is the determined by the size of the text in the
current font plus any horizontal and vertical padding. See
TEXTATTRS= and PAD= .

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Default The GraphValueText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1)
the TEXTATTRS= statement option or 2) the default style element for
the statement (GraphValueText) if no TEXTATTRS= statement option
is used.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

VALIGN=CENTER | TOP | BOTTOM

specifies the vertical alignment of the text.

Default CENTER
 ENTRY Statement 1353

Interaction This option is ignored when AUTOALIGN= is not NONE and the
parent layout is an overlay-type layout.

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
specifies the horizontal alignment of the text that follows.

See HALIGN= on page 1350

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the properties of the text that follows.

See TEXTATTRS= on page 1352

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entry "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as
superscript text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entry "R" {sup "2"} " = " {format (6.4) RSQUARED} ;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its
Unicode specification or keyword equivalent.

"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in
the current font.

For a complete listing, see http://unicode.org/charts/charindex.html.

keyword
a SAS keyword for a UNICODE character.

For a listing of keywords supplied by SAS, see “Reserved Keywords and

Unicode Values” on page 1343.

dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through
1354 Chapter 11 / Text Statements

their UNICODE value. Some fonts support only a limited set of

UNICODE values. If the UNICODE value is not accessible, then the
command might be ignored or a nonprintable character might be
substituted.

See “Using UNICODE Text” on page 1342

“Rules for Unicode and Special Character Specifications” on page 1342

Example entry {unicode alpha} "=" CONF;

entry {unicode "03B1"x} "=" CONF;

Details
An ENTRY statement creates one line of text in the plot area. The statement must
be specified within a LAYOUT, HEADER, SIDEBAR, or CELL statement block. It
cannot be specified outside of one of these blocks, where global statements like
ENTRYTITLE and ENTRYFOOTNOTE are used.

The text line specified in an ENTRY statement can be made of several pieces of the
text called text-items. Statement options that are used establish properties for the
entire line of text (that is, all text-items). These properties can be overridden with
prefix-options that are specific to one or more text-items. See “Required Argument”
on page 1348 for more information.

You can specify an ENTRY statement inside or outside of a nested statement block:
n When you specify an ENTRY statement inside a nested LAYOUT, HEADER,
SIDEBAR, or CELL statement block, then, by default, the text is placed inside
the bounding area of the results that the nested statement block creates.
n When you specify an ENTRY statement outside of a nested LAYOUT, HEADER,
SIDEBAR, or CELL statement block, then the text is placed outside of the
bounding area of the results that the nested statement block creates.

Example: ENTRY Statement

The following graph was generated by the “Example Program” on page 1355:
 ENTRY Statement 1355

Example Program
proc template;
define statgraph entry;
begingraph;
layout overlay;

entry halign=right "First entry statement" /

valign=top;

histogram weight;

entry halign=right "Second entry statement";

entry halign=right "Third entry statement" /

valign=bottom pad=(bottom=40px);

endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entry;

run;
 1356 Chapter 11 / Text Statements

ENTRYFOOTNOTE Statement
Displays a footnote.

Requirement: You must place the ENTRYFOOTNOTE statement directly inside the BEGINGRAPH
block. Do not embed it in any other GTL statement block.

Syntax
ENTRYFOOTNOTE text-item <text-item …> </options>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry
border.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-
option, specifies the properties of individual text-items.

Text options
HALIGN=CENTER | LEFT | RIGHT
as a statement option, specifies the horizontal alignment of the text. As a
prefix-option, specifies the horizontal alignment of individual text-items.
HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is
always centered in the graph area.
SHORTTEXT=(text-item <...text-item>)
specifies alternate text to use if the specified text is too long for the
output width.
 ENTRYFOOTNOTE Statement 1357

TEXTFITPOLICY=WRAP | SHORT | TRUNCATE

specifies how to handle text that is too long to fit in the output width.

Required Argument
text-item <…text-item>
specifies one or more pieces of text for the footnote. Each text-item has the
following form:

<prefix-option …<prefix-option>>"string" | dynamic | character-expression |

{text-command}

Each piece of text can have multiple prefix options that precede the piece of
text. A piece of text is either a string literal, a dynamic, or a text command. All
text-items with the same HALIGN= setting are concatenated into one string. Up
to three strings can result for each ENTRY statement. Leading and trailing
blanks in the concatenated string are always used.

When a prefix option is used, it applies not only to the text that immediately
follows the prefix option, but also to all subsequent text strings and text
commands. If the same prefix option appears more than once, then that prefix
overrides the last used prefix of the same name for the subsequent text strings
and text commands.

Restriction The maximum length for the footnote text is 512 characters in
SAS 9.4 and earlier releases. Starting with SAS 9.4M1, this
restriction is removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1337

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect in order for the color to be seen.
By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
1358 Chapter 11 / Text Statements

Default FALSE

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

HALIGN=CENTER | LEFT | RIGHT

as a statement option, specifies the horizontal alignment of the text. As a prefix-
option, specifies the horizontal alignment of individual text-items. Each text-
item has a horizontal alignment, and text-items with the same alignment are
always grouped together.

Default CENTER

HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is always
centered in the graph area.

AUTO
specifies that the system determines how the text is centered.

The text is centered as follows:

n For LAYOUT GRIDDED, LAYOUT OVERLAY3D, and LAYOUT REGION
layouts, and for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have more than one column, center the text
in the graph area.
n For LAYOUT OVERLAY and LAYOUT OVERLAYEQUATED layouts, and
for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have only one column, center the text in the
graph wall area. If the length of the text exceeds the width of the graph
wall area, then center the text in the graph area instead.

GRAPH
specifies that the text is always centered in the graph area.

Default AUTO

Interaction The prefix option HALIGN= must specify CENTER for this option to
have any effect.

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
 ENTRYFOOTNOTE Statement 1359

Default FALSE

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:
1360 Chapter 11 / Text Statements

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default LEFT=3, RIGHT=3, TOP=0, and BOTTOM=0

Note The default units for dimension are pixels.

See “dimension” on page xiii

SHORTTEXT=(text-item <...text-item>)
specifies alternate text to use if the specified text is too long for the output
width. If the shortened text is itself too long, then it is truncated.

Interactions This option has no effect unless TEXTFITPOLICY= SHORT.

This option is ignored if any text-items include an HALIGN= prefix

option.

The horizontal alignment of the shortened text is derived from the

horizontal alignment of the text to be shortened.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Default The GraphFootnoteText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1)
the TEXTATTRS= statement option or 2) the default style element for
the statement (GraphValueText) if no TEXTATTRS= statement option
is used.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

 ENTRYFOOTNOTE Statement 1361

TEXTFITPOLICY=WRAP | SHORT | TRUNCATE

specifies how to handle text that is too long to fit in the output width.

WRAP
specifies that the text wraps to the next line(s).

Restriction Text wrapping is available only for ENTRYFOOTNOTE

statements that appear in the outermost layout.

SHORT
specifies that the text indicated by the SHORTTEXT= option be substituted
when the text does not fit.

Requirement You must specify the SHORTTEXT= option for this option to
take effect.

Note If the short text is also too long, then it is truncated.

See “SHORTTEXT=(text-item <...text-item>) ” on page 1360

TRUNCATE
specifies that the text is truncated to make it fit.

Default WRAP

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
specifies the horizontal alignment of the text that follows.

See HALIGN= on page 1358

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the properties of the text that follows.

See TEXTATTRS= on page 1360

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entryfootnote "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as
superscript text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entryfootnote "R" {sup "2"} " = " RSQUARED;

1362 Chapter 11 / Text Statements

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its
Unicode specification or keyword equivalent.

"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in
the current font.

For a complete listing, see http://unicode.org/charts/charindex.html.

keyword
a SAS keyword for a UNICODE character.

For a listing of keywords supplied by SAS, see “Reserved Keywords and

Unicode Values” on page 1343.

dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through
their UNICODE value. Some fonts support only a limited set of
UNICODE values. If the UNICODE value is not accessible, then the
command might be ignored or a nonprintable character might be
substituted.

See “Using UNICODE Text” on page 1342

“Rules for Unicode and Special Character Specifications” on page 1342

Example entryfootnote {unicode alpha} "=" CONF;

entryfootnote {unicode "03B1"x} "=" CONF;

Details
The ENTRYFOOTNOTE statement places footnote text below the graphical area.
More than one ENTRYFOOTNOTE statement can be used. Footnotes appear in the
order of the ENTRYFOOTNOTE statements within the template.

When you add an ENTRYFOOTNOTE statement to a template definition, place the

statement in the BEGINGRAPH block, but do not embed it in another GTL
statement block. The ENTRYFOOTNOTE statement must be placed directly inside
the BEGINGRAPH block. The following conditions apply to ENTRYFOOTNOTE
statements.
 ENTRYFOOTNOTE Statement 1363

n All ENTRYFOOTNOTE statements that appear in the template are gathered and
placed in the ENTRYFOOTNOTE area.
n The placement of an ENTRYFOOTNOTE statement is relevant only in relation
to other ENTRYFOOTNOTE statements.
n As the number of ENTRYFOOTNOTE statements increases, the size of the
graphical area is reduced.

Footnotes always span the entire width of the output. By default, footnotes are
“center-aligned,” based on the type of the outermost layout. The meaning of
“center-aligned” varies by layout type and the number of columns in the layout:

Layout Type Default horizontal centering of footnotes

GRIDDED Centered on width of entire graph

OVERLAY3D
LATTICE (COLUMNS > 1)
DATAPANEL (COLUMNS > 1)
DATALATTICE (COLUMNS > 1)

OVERLAY Centered on the plot area

OVERLAYEQUATED
LATTICE (COLUMNS=1)
DATAPANEL (COLUMNS= 1)
DATALATTICE (COLUMNS = 1)

Example: ENTRYFOOTNOTE Statement

The following graph was generated by the “Example Program” on page 1364:
 1364 Chapter 11 / Text Statements

Example Program
proc template;
define statgraph entryfootnote;
begingraph;

entryfootnote "First entryfootnote statement" ;

layout overlay;
histogram weight;
endlayout;

entryfootnote "Second entryfootnote statement" ;

entryfootnote "Third entryfootnote statement" ;

endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entryfootnote;

run;

ENTRYTITLE Statement
Displays a title.
 ENTRYTITLE Statement 1365

Requirement: You must place the ENTRYTITLE statement directly inside the BEGINGRAPH
block. Do not embed it in any other GTL statement block.

Syntax
ENTRYTITLE text-item <text-item …> </options>;

Summary of Optional Arguments

Appearance options
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.
BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the attributes of the border line around the text.
OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).
OUTERPAD=AUTO | dimension | (pad-options)
specifies the amount of extra space to add outside the entry border.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry
border.
TEXTATTRS=style-element | style-element (text-options) | (text-options)
as a statement option, specifies the properties of the text. As a prefix-
option, specifies the properties of individual text-items.

Text options
HALIGN=CENTER | LEFT | RIGHT
as a statement option, specifies the horizontal alignment of the text. As a
prefix-option, specifies the horizontal alignment of individual text-items.
HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is
always centered in the graph area.
SHORTTEXT=(text-item <…text-item>)
specifies alternate text to use if the specified text is too long for the
output width. If the shortened text is itself too long, then it is truncated.
TEXTFITPOLICY=WRAP | SHORT | TRUNCATE
specifies how to handle text that is too long to fit in the output width.
1366 Chapter 11 / Text Statements

Required Argument
text-item <text-item …>
specifies one or more pieces of text for the entry. Each text-item has the
following form:

<prefix-option <…prefix-option>>"string" | dynamic | character-expression |

{text-command}

Each piece of text can have multiple prefix options that precede the piece of
text. A piece of text is either a string literal, a dynamic, or a text command. All
text-items with the same HALIGN= setting are concatenated into one string. Up
to three strings can result for each ENTRY statement. Leading and trailing
blanks in the concatenated string are always used.

When a prefix option is used, it applies not only to the text that immediately
follows the prefix option, but also to all subsequent text strings and text
commands. If the same prefix option appears more than once, then that prefix
overrides the last used prefix of the same name for the subsequent text strings
and text commands.

Restriction The maximum length for the title text is 512 characters in SAS 9.4
and earlier releases. Starting with SAS 9.4M1, this restriction is
removed.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

Note Leading spaces are preserved, and trailing spaces are stripped.

See Chapter 10, “Managing Text Items,” on page 1337 for more
information and several examples.

Optional Arguments
BACKGROUNDCOLOR=style-reference | color
specifies the color of the text background.

style-reference
specifies a style reference in the form style-element:style-attribute.

Only the style-attribute named COLOR or CONTRASTCOLOR is used.

Default The GraphBackground:Color style reference.

Interaction OPAQUE= TRUE must be in effect in order for the color to be seen.
By default, OPAQUE=FALSE.

BORDER=TRUE | FALSE
specifies whether a border is displayed around the text.

Default FALSE
 ENTRYTITLE Statement 1367

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the attributes of the border line around the text.

Default The GraphBorderLines style element.

Interaction BORDER= TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

HALIGN=CENTER | LEFT | RIGHT

as a statement option, specifies the horizontal alignment of the text. As a prefix-
option, specifies the horizontal alignment of individual text-items. Each text-
item has a horizontal alignment, and text-items with the same alignment are
always grouped together.

Default CENTER

HALIGNCENTER=AUTO | GRAPH
specifies whether the text is centered automatically by the system or is always
centered in the graph area.

AUTO
specifies that the system determines how the text is centered.

The text is centered as follows:

n For LAYOUT GRIDDED, LAYOUT OVERLAY3D, and LAYOUT REGION
layouts, and for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have more than one column, center the text
in the graph area.
n For LAYOUT OVERLAY and LAYOUT OVERLAYEQUATED layouts, and
for LAYOUT LATTICE, LAYOUT DATAPANEL, and LAYOUT
DATALATTICE layouts that have only one column, center the text in the
graph wall area. If the length of the text exceeds the width of the graph
wall area, then center the text in the graph area instead.

GRAPH
specifies that the text is always centered in the graph area.

Default AUTO

Interaction The prefix option HALIGN= must specify CENTER for this option to
have any effect.

OPAQUE=TRUE | FALSE
specifies whether the text background is opaque (TRUE) or transparent
(FALSE).

Default FALSE
1368 Chapter 11 / Text Statements

Interaction When this option is set to FALSE, the background color is not used.

See “boolean ” on page xii for other Boolean values that you can use.

OUTERPAD=AUTO | dimension | (pad-options)

specifies the amount of extra space to add outside the entry border.

AUTO
specifies that the default outside padding for this component is used.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default AUTO

Note The default units for dimension are pixels.

See “dimension” on page xiii

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the entry border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the entry border.

(pad-options)
a space-separated list of one or more name-value pair pad options, enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space to add to the left side.
 ENTRYTITLE Statement 1369

RIGHT=dimension
specifies the amount of extra space to add to the right side.

TOP=dimension
specifies the amount of extra space to add to the top.

BOTTOM=dimension
specifies the amount of extra space to add to the bottom.

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default LEFT=3, RIGHT=3, TOP=0, and BOTTOM=0

Note The default units for dimension are pixels.

See “dimension” on page xiii

SHORTTEXT=(text-item <…text-item>)
specifies alternate text to use if the specified text is too long for the output
width. If the shortened text is itself too long, then it is truncated.

Interactions This option has no effect unless TEXTFITPOLICY= SHORT.

This option is ignored if any text-items include an HALIGN= prefix

option.

The horizontal alignment of the shortened text is derived from the

horizontal alignment of the title to be shortened.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

as a statement option, specifies the properties of the text. As a prefix-option,
specifies the properties of individual text-items.

Default The GraphTitleText style element.

Notes This option can be used as both a prefix option and a statement option.
When used as a prefix option, it overrides the statement option.

When used as a prefix option, TEXTATTRS= cancels the last used

TEXTATTRS= prefix option. It resets all text options to those set by 1)
the TEXTATTRS= statement option or 2) the default style element for
the statement (GraphValueText) if no TEXTATTRS= statement option
is used.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Text Options” on page 1606 for available text-options.

TEXTFITPOLICY=WRAP | SHORT | TRUNCATE

specifies how to handle text that is too long to fit in the output width.
1370 Chapter 11 / Text Statements

WRAP
specifies that the text wraps to the next line(s).

Restriction Text wrapping is available only for ENTRYTITLE statements that

appear in the outermost layout.

SHORT
specifies that the text indicated by the SHORTTEXT= option be substituted
when the text does not fit.

Requirement You must specify the SHORTTEXT= option for this option to
take effect.

Note If the short text is also too long, then it is truncated.

See “SHORTTEXT=(text-item <...text-item>) ” on page 1360

TRUNCATE
specifies that the text is truncated to make it fit.

Default WRAP

Prefix Options
HALIGN=CENTER | LEFT | RIGHT
specifies the horizontal alignment of the text that follows.

See HALIGN= on page 1367

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the properties of the text that follows.

See TEXTATTRS= on page 1369

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entrytitle "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as
superscript text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example entrytitle "R" {sup "2"} " = " RSQUARED;

 ENTRYTITLE Statement 1371

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its
Unicode specification or keyword equivalent.

"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in
the current font.

For a complete listing, see http://unicode.org/charts/charindex.html.

keyword
a SAS keyword for a UNICODE character.

For a listing of keywords supplied by SAS, see “Reserved Keywords and

Unicode Values” on page 1343.

dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through
their UNICODE value. Some fonts support only a limited set of
UNICODE values. If the UNICODE value is not accessible, then the
command might be ignored or a nonprintable character might be
substituted.

See “Using UNICODE Text” on page 1342

“Rules for Unicode and Special Character Specifications” on page 1342

Example entrytitle {unicode alpha} "=" CONF;

entrytitle {unicode "03B1"x} "=" CONF;

Details
The ENTRYTITLE statement places title text above the graphical area. More than
one ENTRYTITLE statement can be used. Titles appear in the order of the
ENTRYTITLE statements within the template.

When you add an ENTRYTITLE statement to a template definition, place the

statement in the BEGINGRAPH block, but do not embed it in another GTL
statement block. The ENTRYTITLE statement must be placed directly inside the
BEGINGRAPH block. The following conditions apply to ENTRYTITLE statements.
1372 Chapter 11 / Text Statements

n All ENTRYTITLE statements that appear in the template are gathered and
placed in the ENTRYTITLE area.
n The placement of an ENTRYTITLE statement is relevant only in relation to
other ENTRYTITLE statements.
n As the number of ENTRYTITLE statements increases, the size of the graphical
area is reduced.

Titles always span the entire width of the output. By default, titles are “center-
aligned,” based on the type of the outermost layout. The meaning of “center-
aligned” varies by layout type and the number of columns in the layout:

Layout Type Default horizontal centering of titles

GRIDDED Centered on width of entire graph

OVERLAY3D
LATTICE (COLUMNS > 1)
DATAPANEL (COLUMNS > 1)
DATALATTICE (COLUMNS > 1)

OVERLAY Centered on the plot area

OVERLAYEQUATED
LATTICE (COLUMNS=1)
DATAPANEL (COLUMNS= 1)
DATALATTICE (COLUMNS = 1)

Example: ENTRYTITLE Statement

The following graph was generated by the “Example Program” on page 1373:
 ENTRYTITLE Statement 1373

Example Program
proc template;
define statgraph entrytitle;
begingraph;

entrytitle "First entrytitle statement" ;

layout overlay;
histogram weight;
endlayout;

entrytitle "Second entrytitle statement" ;

entrytitle "Third entrytitle statement" ;

endgraph;
end;
run;

proc sgrender data=sashelp.cars template=entrytitle;

run;
1374 Chapter 11 / Text Statements
 1375

PART 8

Custom Marker Definition Statements

Chapter 12
Custom Marker Definition Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1377
1376
 1377

12
Custom Marker Definition
Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1377
SYMBOLCHAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1377
SYMBOLIMAGE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1384

Dictionary

SYMBOLCHAR Statement
Defines a marker symbol using a Unicode character so that the symbol can be referenced in other
statements.

Requirement: The SYMBOLCHAR statement must be a child of the BEGINGRAPH block. It cannot
be embedded in any other GTL statement block.
Note: This statement is valid starting in SAS 9.4M1.

Syntax
SYMBOLCHAR NAME=marker-name CHAR="""hex-stringx | keyword | dynamic
</ options>;
1378 Chapter 12 / Custom Marker Definition Statements

Required Arguments
NAME=marker-name
specifies a name for the marker symbol. The name can be used in statements
that support marker symbols.

Interaction If the name matches one of the system-provided symbol names such
as CIRCLE, then the system symbol is replaced by the user-defined
symbol. See “Marker Options” on page 1604 for a list of the system-
provided symbols.

Note Do not enclose the name in quotation marks.

See “Details” on page 1380

CHAR="""hex-stringx | keyword | dynamic

specifies a glyph (character) to be used as the marker symbol. You specify the
character by using its Unicode specification or its keyword equivalent.

"hex-string"x
specifies a four-byte hexadecimal constant that represents a Unicode
character in the current font.

You can find a complete listing of the Unicode hexadecimal constants at the
following URL:http://www.unicode.org/charts/charindex.html

Requirement You must specify a Unicode character value.

keyword
specifies a SAS keyword for a Unicode character.

See Appendix 2, “Reserved Keywords and Unicode Values,” on page 1595

dynamic
specifies a reference to a dynamic variable that resolves to either a "hex-
string"x constant or a Unicode character keyword.

Tip This statement attempts to access the specified Unicode value in the
current font. Some fonts do not support accessing characters by using their
Unicode value. Other fonts support only a limited set of Unicode values. If
the Unicode value is not accessible, then this statement might be ignored
or a nonprintable character might be substituted.

Optional Arguments
HOFFSET=number
specifies a horizontal offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Interaction Starting with SAS 9.4M3, the specified offset is also applied to the
marker symbol that is displayed in the legend.
 SYMBOLCHAR Statement 1379

Notes Prior to SAS 9.4M3, a positive offset moves the marker symbol
bounding box to the right. A negative offset moves it to the left.

Starting with SAS 9.4M3, a positive offset moves the marker symbol
to the right within its bounding box, and a negative offset moves it
to the left. The bounding box remains centered on the data point.
After the offset, size, and rotation are applied to the marker symbol,
any portion of the symbol that falls outside of the marker bounding
box is clipped.

Tip If clipping occurs, use this option, the VOFFSET= and SCALE=
options, and the suboption SIZE= in the MARKERATTRS= option to
remove the clipping.

ROTATE=number
specifies the angle of rotation for the marker symbol in degrees. Positive angles
are measured in the counter-clockwise direction, and negative angles are
measured in the clockwise direction.

Default 0. No rotation is performed.

Note An angle that exceeds 360 degrees in absolute value can be specified.

SCALE=number
specifies a scale factor for the marker symbol as a percentage. The scale factor
is applied to the character's height and width.

Default 1.0 (100%)

Note The outer edges of the image might be clipped by the bounding box
when a large scale factor is specified.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text attributes for the character symbol.

Default The GraphUnicodeText style element.

Restriction Only the text attributes FAMILY=, STYLE=, and WEIGHT= are
used. The COLOR= and SIZE= text attributes are derived from the
plot statement's MARKERATTRS= option.

Requirement The font family specified in FAMILY= must support Unicode

characters.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

VOFFSET=number
specifies a vertical offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.
1380 Chapter 12 / Custom Marker Definition Statements

Interaction Starting with SAS 9.4M3, the specified offset is also applied to the
marker symbol that is displayed in the legend.

Notes Prior to SAS 9.4M3, a positive offset moves the marker symbol
bounding box up. A negative offset moves it down.

Starting with SAS 9.4M3, a positive offset moves the marker symbol
up within its bounding box, and a negative offset moves it down. The
bounding box remains centered on the data point. After the offset,
size, and rotation are applied to the marker symbol, any portion of
the symbol that falls outside of the marker bounding box is clipped.

Tip If clipping occurs, use this option, the HOFFSET= and SCALE=
options, and the suboption SIZE= in the MARKERATTRS= option to
remove the clipping.

Details
The SYMBOLCHAR statement defines a custom marker symbol from a Unicode
character. You can use the marker symbol that is created in the following options:
n the DATASYMBOLS= option in the BEGINGRAPH statement

n the SYMBOL= suboption of the MARKERATTRS= option, which is supported by

the following statements:

DISCRETELEGEND (VALUE statement) SCATTERPLOT

LEGENDITEM SCATTERPLOTMATRIX
LINECHART SERIESPLOT
NEEDLEPLOT STEPPLOT

Symbol specifications are not checked for uniqueness. More than one
SYMBOLCHAR statement can define the same character. Therefore, you can use
SYMBOLCHAR statements in IF/ELSE statements. Symbol specifications also are
not validated at compile time. An invalid specification might not generate a warning
when the output is rendered and might create unexpected results.

You can use the COLOR=, SIZE=, and TRANSPARENCY= suboptions of the
MARKERATTRS= option to modify the appearance of markers that are created by
the SYMBOCHAR statement. The WEIGHT= suboption has no effect on these
markers.

The markers are clipped to the original marker size after rotation, scaling, and
offsets are applied. If clipping occurs, then you can use the SIZE= suboptions of the
MARKERATTRS= option in the plot statement and the SCALE= option in the
SYMBOLCHAR statement to adjust the size and scaling to eliminate the clipping.
 SYMBOLCHAR Statement 1381

Example: SYMBOLCHAR Statement

This example shows how to create marker symbols from the Unicode ballot X,
heavy character ('2718'x) and check mark, heavy character ('2718'x). It also shows
you how to use the symbols in a scatter plot and how to add them to a legend. In
this example, a grouped bar chart shows the daily failure count for two processes,
Process A and Process B. A scatter plot is overlaid on the bar chart in order to show
the total failure count above each bar. Acceptable counts are indicated by a green
Unicode check mark. High counts are indicated by a red Unicode ballot X. Finally, a
legend is used to describe the custom symbols. The following figure shows the
output.

Example Program
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
1382 Chapter 12 / Custom Marker Definition Statements

data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;
/* Define the template */
proc template;
define statgraph failures;
begingraph;
/* Define the norm/high marker symbols */
symbolchar name=norm char='2714'x;
symbolchar name=high char='2718'x;
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=15pt color=green)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=15pt color=red)
label="High Failure Count";

/* Define an attribute map for the status markers */

discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=red);
value "B-HIGH" / markerattrs=(symbol=high color=red);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap" ;
/* Define the graph */
entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=eval(count+2) / datalabel=count
group=status
groupdisplay=cluster markerattrs=(size=15pt);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot)
curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;
proc sgrender data=failure template=failures;
run;
 SYMBOLCHAR Statement 1383

Program Description
Here is the SAS code for this example.

Create the graph data. The data in Sashelp.Failure is the source data. To get the
daily counts, the data is summarized by weekday and process. The Stat column is
then added to indicate whether the count is acceptable or high. Acceptable values
are values that do not exceed 120% of the target value. Values over 120% are
considered high. Acceptable values are A-NORM and B-NORM, and high values are
A-HIGH and B-HIGH, where A and B identify the process.
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;

Open the template definition.

/* Define the template */
proc template;
define statgraph failures;
begingraph;

Define the custom marker symbols. The NAME= option specifies an identifier for
the markers, and the CHAR= option specifies the Unicode value. Unicode value
'2714'x specifies a heavy check mark, and value '2718'x specifies a heavy ballot X.
/* Define the norm/high marker symbols */
symbolchar name=norm char='2714'x;
symbolchar name=high char='2718'x;

Create legend entries for the custom markers. Suboption SYMBOL= in the
MARKERATTRS= option specifies the identifier of the custom markers.
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=15pt color=green)
label="Normal Failure Count";
 1384 Chapter 12 / Custom Marker Definition Statements

legenditem type=marker name="high" /

markerattrs=(symbol=high size=15pt color=red)
label="High Failure Count";

Create a discrete attribute map to map the custom markers to the Stat column
values. Suboption SYMBOL= in the MARKERATTRS= option specifies the identifier
of the custom markers. Suboption COLOR= sets the color of the Unicode marker
character and the associated data label. The DISCRETEATTRVAR statement
creates attribute map variable STATUS.
/* Define an attribute map for the status markers */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=red);
value "B-HIGH" / markerattrs=(symbol=high color=red);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap" ;

Define the graph and close the template definition. The SCATTERPLOT statement
GROUP= option references the STATUS attribute map variable. The EVAL
statement in the Y= option positions each character just above the top of its bar.
The SIZE= suboption of the MARKERATTRS= option sets the marker size. The
custom markers are included in the DISCRETELEGEND statement.
/* Define the graph */
entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=eval(count+2) / datalabel=count
group=status
groupdisplay=cluster markerattrs=(size=15pt);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot)
curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;

Generate the graph.

proc sgrender data=failure template=failures;
run;

SYMBOLIMAGE Statement
Defines a marker symbol using an image file so that the image can be referenced in other statements.
 SYMBOLIMAGE Statement 1385

Requirement: The SYMBOLIMAGE statement must be a child of the BEGINGRAPH block. It

cannot be embedded in any other GTL statement block.
Note: This statement is valid starting in SAS 9.4M1.

Syntax
SYMBOLIMAGE NAME=marker-name IMAGE="image-file-spec" < / options >;

Required Arguments
NAME=marker-name
specifies a name for the marker symbol. The name can be used in statements
that support marker symbols.

Interaction If the name matches one of the system-provided symbol names such
as CIRCLE, then the system symbol is replaced by the user-defined
symbol. See “Marker Options” on page 1604 for a list of the system-
provided symbols.

Note Do not enclose the name in quotation marks.

See “Details” on page 1386

IMAGE="image-file-spec"
specifies the name and location of the image file. The supported image types
are GIF, JPEG, and PNG.

Requirements The image file specification must be enclosed in quotation marks.

The image file must be located on the local file system. URL
access is not supported.

Example image="c:\temp\saslogo.gif"

Optional Arguments
HOFFSET=number
specifies a horizontal offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Notes A positive offset moves the marker symbol bounding box to the right. A
negative offset moves it to the left.

The specified offset is not applied to the image symbol that is

displayed in the legend.
1386 Chapter 12 / Custom Marker Definition Statements

ROTATE=number
specifies the angle of rotation for the marker symbol in degrees. Positive angles
are measured in the counter-clockwise direction, and negative angles are
measured in the clockwise direction.

Default 0. No rotation is performed.

Note An angle that exceeds 360 degrees in absolute value can be specified.

SCALE=number
specifies a scale factor for the marker symbol as a percentage. The scale factor
is applied to the character's height and width.

Default 1.0 (100%)

Note The outer edges of the image might be clipped by the bounding box
when a large scale factor is specified.

VOFFSET=number
specifies a vertical offset for the marker symbol.

Default 0. The marker symbol is centered on its data point.

Range -0.5–+0.5, where 0.5 represents one-half of the original marker size.

Notes A positive offset moves the marker symbol bounding box up. A
negative offset moves it down.

The specified offset is not applied to the image symbol that is

displayed in the legend.

Details
The SYMBOLIMAGE statement defines a custom marker symbol from an image
that is stored in an image file. The image file must exist on the local file system.
URL access is not supported. The supported image formats are GIF, JPG, and PNG.
The marker symbol that is created can be used in the following options:
n the DATASYMBOLS= option in the BEGINGRAPH statement

n the SYMBOL= suboption of the MARKERATTRS= option, which is supported by

the following statements:

DISCRETELEGEND (VALUE statement) SCATTERPLOT

LEGENDITEM SCATTERPLOTMATRIX
LINECHART SERIESPLOT
NEEDLEPLOT STEPPLOT

Symbol specifications are not checked for uniqueness. More than one
SYMBOLIMAGE statement can define the same character. Therefore, you can use
SYMBOLIMAGE statements in IF/ELSE statements.
 SYMBOLIMAGE Statement 1387

The following options normally affect the appearance of markers. However, they
have no effect on image marker symbols that are created by the SYMBOLIMAGE
statement.
n the BEGINGRAPH statement DATACONTRASTCOLORS= option

n the COLOR= and WEIGHT= suboptions of the MARKERATTRS= option

n the FILLEDOUTLINEDMARKERS= option

The SIZE= and TRANSPARENCY= suboptions of the MARKERATTRS= option do

affect the appearance of markers that are created by the SYMBOLIMAGE
statement.

The markers are clipped to the original marker size after rotation, scaling, and
offsets are applied. If clipping occurs, then you can use the SIZE= suboptions of the
MARKERATTRS= option in the plot statement and the SCALE= option in the
SYMBOLIMAGE statement to adjust the size and scaling to eliminate the clipping.

Example: SYMBOLIMAGE Statement

This example shows how to create marker symbols from GIF icon image files. It also
shows you how to use the symbols in a scatter plot and how to add them to a
legend. In this example, a grouped bar chart shows the daily failure count for two
processes, Process A and Process B. A scatter plot is overlaid on the bar chart to
show the total failure count above each bar. Acceptable counts are indicated by a
green check mark icon. High counts are indicated by a caution icon. Finally, a legend
is used to describe the custom symbols. The following figure shows the output.
 1388 Chapter 12 / Custom Marker Definition Statements

Example Program
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;
/* Define the template */
proc template;
define statgraph failures;
begingraph;
/* Define the norm/high marker symbols */
symbolimage name=norm image="C:\temp\check_green.gif" /
voffset=0.5;
symbolimage name=high image="C:\temp\alert_orange.gif" /
voffset=0.5;
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=16px)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=16px)
label="High Failure Count";
/* Define an attribute map for the status markers
Note: COLOR= affects marker labels only in this case. */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=orange);
value "B-HIGH" / markerattrs=(symbol=high color=orange);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap";
/* Define the graph */
 SYMBOLIMAGE Statement 1389

entrytitle "Daily Failure Report By Process";

entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=count / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=16px);
referenceline y=&target /
lineattrs=(color=lightred pattern=dot)
curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;
proc sgrender data=failure template=failures;
run;

Program Description
Here is the SAS code for this example.

Create the graph data. The data in Sashelp.Failure is the source data. To get the
daily counts, the data is summarized by weekday and process. The Stat column is
then added to indicate whether the count is acceptable or high. Acceptable values
are values that do not exceed 120% of the target value. Values over 120% are
considered high. Acceptable values are A-NORM and B-NORM, and high values are
A-HIGH and B-HIGH, where A and B identify the process.
/* Summarize the data in SASHELP.FAILURE to compute the
failure count by day and process */
proc sort data=sashelp.failure out=temp;
by day process;
run;

proc summary data=temp;

by day process;
var count;
output out=failure sum=count;
run;

/* Determine the daily status for each process:

count at or under 120% of target is NORM, over is HIGH */
%let target=25;
data failure;
length stat $20;
set failure;
if (count <= (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-NORM"));
if (count > (&target*1.2)) then
stat=strip(cat(substr(process,index(process,' '),2),"-HIGH"));
label day="Weekday" count="Failure Count" stat="Status";
run;
1390 Chapter 12 / Custom Marker Definition Statements

Open the template definition.

/* Define the template */
proc template;
define statgraph failures;
begingraph;

Define the custom marker symbols. The NAME= option specifies an identifier for
the markers, and the IMAGE= option specifies the image file. The VOFFSET= option
raises the markers 50% from their data points in order to prevent them from
overlapping the bars in the bar chart.
/* Define the norm/high marker symbols */
symbolimage name=norm image="C:\temp\check_green.gif" /
voffset=0.5;
symbolimage name=high image="C:\temp\alert_orange.gif" /
voffset=0.5;

Create legend entries for the custom markers. Suboption SYMBOL= in the
MARKERATTRS= option specifies the identifier of the custom markers.
/* Create legend entries for the norm/high markers */
legenditem type=marker name="norm" /
markerattrs=(symbol=norm size=16px)
label="Normal Failure Count";
legenditem type=marker name="high" /
markerattrs=(symbol=high size=16px)
label="High Failure Count";

Create a discrete attribute map to map the custom markers to the Stat column
values. Suboption SYMBOL= in the MARKERATTRS= option specifies the identifier
of the custom markers. Suboption COLOR= sets the color of the associated data
label. It does not affect the color of the image. The DISCRETEATTRVAR statement
creates attribute map variable STATUS.
/* Define an attribute map for the status markers
Note: COLOR= affects marker labels only in this case. */
discreteattrmap name="statsymmap";
value "A-NORM" / markerattrs=(symbol=norm color=green);
value "B-NORM" / markerattrs=(symbol=norm color=green);
value "A-HIGH" / markerattrs=(symbol=high color=orange);
value "B-HIGH" / markerattrs=(symbol=high color=orange);
enddiscreteattrmap;
discreteattrvar attrvar=status var=stat attrmap="statsymmap";

Define the graph and close the template definition. The SCATTERPLOT statement
GROUP= option references the STATUS attribute map variable. The SIZE=
suboption of the MARKERATTRS= option sets the marker size. The custom markers
are included in the DISCRETELEGEND statement.
/* Define the graph */
entrytitle "Daily Failure Report By Process";
entryfootnote "A failure count that is more than 120% of the
target is considered high.";
layout overlay / yaxisopts=(offsetmax=0.1);
barchartparm category=day response=count / name="barchart"
group=process groupdisplay=cluster;
scatterplot x=day y=count / datalabel=count group=status
groupdisplay=cluster markerattrs=(size=16px);
 SYMBOLIMAGE Statement 1391

referenceline y=&target /
lineattrs=(color=lightred pattern=dot)
curvelabel="Target";
discretelegend "barchart" "norm" "high" / down=2
order=columnmajor;
endlayout;
endgraph;
end;
run;

Generate the graph.

proc sgrender data=failure template=failures;
run;
1392 Chapter 12 / Custom Marker Definition Statements
 1393

PART 9

Draw Statements

Chapter 13
Key Concepts for Using Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395

Chapter 14
Draw Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405
1394
 1395

13
Key Concepts for Using Draw
Statements

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395
Types of Elements That Can Be Drawn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1398
About the Drawing Space and Drawing Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1398
How the Drawn Elements Are Anchored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1401
About Drawing Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1402

Introduction
The draw statements enable you to customize a graph by drawing visual elements
anywhere within the graph. The following sections provide a comprehensive
example and a brief overview to the major drawing concepts. For more details
about the individual draw statements and a simple example of each, see the
discussion for each individual statement. For complete usage information for the
draw statements, consult the SAS Graph Template Language: User’s Guide.

Example
The following example shows a bar chart of the top global automobile makers in
2008. To focus the graph on the 2008 merger between Chrysler and Fiat, the
example uses DRAWRECTANGLE to highlight the three bars in the chart that
correspond to the unit sales for the two auto makers. The example also uses a
DRAWARROW statement and two BEGINPOLYLINE blocks to point to the sales
1396 Chapter 13 / Key Concepts for Using Draw Statements

figures, and it uses a DRAWTEXT statement to explain the implications of the

merger.

All of the draw statements in this example use the data space (see “About the
Drawing Space and Drawing Units” on page 1398 ) to integrate the drawn elements
into the graph’s data area.

/* Create the data and the macro variables for Chrysler, Fiat, and
Chrysler+Fiat widths */
data mydata;
length automaker $30;
input automaker $ 1-30 million_units;
FORMAT million_units 3.1;

select (automaker);
when ("Fiat") do;
colorvar=1;
call symput( "fwidth", million_units);
end;
when ("Chrysler") do;
colorvar=1;
call symput( "cwidth", million_units);
end;
when ("Fiat + Chrysler") do;
colorvar=2;
 Example 1397

call symput( "cfwidth", million_units);

end;
otherwise colorvar=3;
end;

datalines;
Toyota 8.7
GM 7.7
Volkswagen 6.0
Renault-Nissan 5.8
Ford 5.4
Fiat + Chrysler 4.5
Hyundai 4.2
Honda 3.8
PSA 3.2
Fiat 2.5
Suzuki 2.4
Chrysler 2.0
Daimler 1.9
BMW 1.4
Mazda 1.4
Mitsubishi 1.1
;

/* Create template definition */

proc template;
define statgraph automerger;
begingraph / drawspace=datavalue;
entrytitle halign=center
'Top Global Automakers (2008 Annual Unit Sales)';
layout lattice / rowdatarange=data columndatarange=data
rowgutter=10 columngutter=10;
layout overlay / xaxisopts=(label=('Units (millions)'))
yaxisopts=(reverse=true display=(ticks tickvalues line));
barchart category=automaker response=million_units /
group=colorvar
name='bar(h)' barlabel=true dataskin=pressed
orient=horizontal;
drawrectangle x=eval(&cwidth/2.0) y="Chrysler"
width=&cwidth height=0.95 / widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
drawrectangle x=eval(&fwidth/2.0) y="Fiat"
width=&fwidth height=0.95 /
widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
drawrectangle x=eval(&cfwidth/2.0) y="Fiat + Chrysler"
width=&cfwidth height=0.95 / widthunit=data heightunit=data
display=(outline) outlineattrs=(color=yellow);
beginpolyline x=eval(&cwidth + 0.5) y="Chrysler";
draw x=eval(&cwidth + 1.5) y="Chrysler";
draw x=eval(&cwidth + 1.5) y="Fiat";
draw x=eval(&fwidth + 0.5) y="Fiat";
endpolyline;
beginpolyline x=eval(&cwidth + 1.5) y="Suzuki";
draw x=eval(&cfwidth + 1.5) y="Suzuki";
draw x=eval(&cfwidth + 1.5) y="Fiat + Chrysler";
 1398 Chapter 13 / Key Concepts for Using Draw Statements

endpolyline;
drawarrow x1=eval(&cfwidth + 1.5) x2=eval(&cfwidth + 0.5)
y1="Fiat + Chrysler" y2="Fiat + Chrysler"
/ arrowheadscale=0.5 arrowheadshape=barbed;
drawtext "Alliance creates the #6 Global Automaker by volume" /
y="Honda" x=eval(&cfwidth+2.5) width=2 widthunit=data;
endlayout;
endlayout;
endgraph;
end;

proc sgrender data=mydata template=automerger;

run;

Types of Elements That Can Be Drawn

The GTL draw statements enable you to draw the following types of elements in a
graph:
n text

n arrows and lines

n geometric shapes like ovals, rectangles, polygons, and polylines (a set of

connecting lines)
n images

Using these elements individually and especially by combining them, you can
describe the non-data aspects of your graph. For example, you can display a
company logo in a specified location within the graph. Or you can create custom
features that are difficult to create by other means—for example, you could draw a
broken axis. Using the draw statements creatively, you can direct viewer attention
to features of interest in the graph by drawing elements that highlight those
features.

About the Drawing Space and Drawing

Units
Each draw statement positions a drawn element using Cartesian coordinates that
you specify as X and Y values in the statement. The coordinates that you specify
are positioned relative to the drawing space that is in effect for the statement. The
available drawing spaces are the data area, the wall area, the layout area, and the
graph area. The coordinates can be specified in pixel units, percentage units, or as
values that are in the units of the data (available only in the data area).
 About the Drawing Space and Drawing Units 1399

The images that accompany the following descriptions all use pixel data units.
Each, however, was drawn in a different drawing space using this same LAYOUT
OVERLAY definition block:
layout overlay /
xaxisopts=(display=(line ticks tickvalues))
yaxisopts=(display=(line ticks tickvalues));
scatterplot x=height y=weight;
drawrectangle x=0 y=0 width=50 height=50 /
anchor=bottomleft display=(fill) fillattrs=(color=green)
transparency=0.75 widthunit=pixel heightunit=pixel ;
drawline x1=0 y1=0 x2=18 y2=18 / lineattrs=(color=red) ;
endlayout;

Table 13.1 Drawing Spaces

Drawin Drawing
g Space Units Description Example

Data VALUE, The area where data is displayed in the

Area PERCENT graph, honoring the offsets that are set
, PIXEL for the axes. For a discussion on axis
offsets, see “Adjusting Axis Offsets” on
page 1053.
The data area does not apply to graphs
that do not have axes, such as pie
charts, which must be drawn in a
REGION layout.
For graphs produced within
LAYOUTDATAPANEL and
LAYOUTDATALATTICE layouts, drawn
elements are clipped if they extend
outside of the wall boundaries.

Wall PERCENT The area bounded by orthogonal axis

Area , PIXEL pairs, ignoring the offsets that are set
for the axes. In two-dimensional graphs,
there is one wall bounded by the XY
axes. including the secondary axes, if
used. In three-dimensional graphs, there
are three walls, bounded by the XY, YZ,
and XZ axes.
The wall area does not apply to graphs
that do not have axes, such as pie
charts, which must be drawn in a
REGION layout.
For graphs produced within
LAYOUTDATAPANEL and
LAYOUTDATALATTICE layouts, drawn
elements that extend outside of the wall
boundaries are clipped.
 1400 Chapter 13 / Key Concepts for Using Draw Statements

Drawin Drawing
g Space Units Description Example

Layout PERCENT The entire area of the layout container

Area , PIXEL that is the immediate parent container
of the draw statement. The figure to the
right shows the case where a LAYOUT
OVERLAY is the draw statement’s
layout container.
Titles and footnotes are always
displayed outside of the outermost
layout, so those areas are never part of
the layout drawing space.

Graph PERCENT The entire area that is available to the

Area , PIXEL graph display, whether a single-cell or
multi-cell graph.
Because the graph drawing space spans
the entire graph, the location of the
drawn element in the graph is
independent of the draw statement’s
placement within the template
definition, even if the draw statement is
specified within a nested layout.

To specify the drawing space for both the X and the Y dimension, you use the
DRAWSPACE= option. To specify the drawing space individually for either the X or
the Y dimension, you use the options XSPACE=, X1SPACE=, X2SPACE=, YSPACE=,
Y1SPACE=, or Y2SPACE=, depending on the draw statement. The value that you set
on any of these options is a single composite value that specifies both the drawing
space and the drawing units in the following format:
<DrawingSpace><Units>

For example, DRAWSPACE=GRAPHPIXEL specifies the GRAPH drawing space with

PIXEL drawing units, indicating that the statement’s X,Y coordinates are expressed
in pixels. Similarly, DRAWSPACE=LAYOUTPERCENT specifies the LAYOUT
drawing space with PERCENT drawing units, indicating that the statement’s X,Y
coordinates are expressed as percentages.

The global DRAWSPACE is LAYOUTPERCENT. The global DRAWSPACE= setting

for all of the draw statements is LAYOUTPERCENT. The draw statements inherit
the global setting from the DRAWSPACE= setting in the BEGINGRAPH statement.
n To change the global drawing space and drawing units for all of the draw
statements within the template definition, use the DRAWSPACE= option in the
BEGINGRAPH statement.
n To change the default for an individual draw statement, use that statement’s
DRAWSPACE= option. If needed, you can specify different settings for the
different draw statements within the template definition.
 How the Drawn Elements Are Anchored 1401

For the DATA drawing space, the VALUE drawing units specify that the coordinates
are expressed as values along the axis. When you specify the DATA drawing space,
you can use the draw statement’s XAXIS= and YAXIS= options to specify which
axis scale to use for the coordinates.
n If the specified axis does not exist in the plot or is not valid for the draw
statement's layout container, then the draw statement is ignored.
n For a discrete axis, if the statement’s specified X or Y value does not exist in the
data, then the draw statement is ignored.
n For a continuous axis, if the statement’s specified X or Y value does not exist in
the data, then the value is extrapolated.

When specifying the drawing space and drawing units, you can set a common
setting for all of the X and Y coordinates. Or you can specify different settings for
each individual coordinate. The DRAWSPACE= setting in the BEGINGRAPH
statement applies the global space and unit settings to all of the draw statements
within the BEGINGRAPH/ENDGRAPH block. The DRAWSPACE= setting in an
individual draw statement applies the space and unit setting only to the
coordinate(s) for that statement. Thus, for lines and arrows, the setting applies to
both the X1, Y1 coordinate and the X2, Y2 coordinate.

To specify the drawing space and drawing units separately for the X coordinate and
for the Y coordinate, use the XSPACE=, YSPACE=, X1SPACE=, Y1SPACE=,
X2SPACE=, and Y2SPACE= options, as applicable, in each draw statement. These
options override the DRAWSPACE= option.

How the Drawn Elements Are Anchored

When you specify the X and Y coordinates in a draw statement, the element is
drawn from an anchor point that is placed in the drawing area at the X,Y coordinate
point.
n For lines and arrows, the anchor point is the starting point of the line or arrow,
which is specified with the draw statement’s X1= and Y1= values.
n For elements that have height and width, the anchor point can be one of the
following points:

The default anchor position is CENTER. To change the anchor position, use the
draw statement’s ANCHOR= option.
 1402 Chapter 13 / Key Concepts for Using Draw Statements

About Drawing Layers

A draw statement can draw its element in either of two “layers” in the graph: the
front layer or the back layer. By default, the statement draws in the front layer,
which places the element in front of all other graphics elements, including data
points, data labels, axis labels, and so on. In some cases, this might cause the drawn
element to block the view of other graphics elements in the graph.

To prevent a drawn element for blocking the view of other graphics elements in the
graph, you can use the draw statement’s TRANSPARENCY= option to add
transparency to the drawn element. With an appropriate transparency setting, you
should be able to see any graphics elements that are behind the drawn element.

Alternatively, you can use the draw statement’s LAYER= option to draw the
element in the back layer, which places the drawn element behind all other graphics
elements in the graph. For example, the following figure shows two different
versions of a graph that uses DRAWLINE to draw a diagonal line across the axis
area. The version to the left draws the line in the front layer, which causes the line
to cover some of the data labels in the graph and portions of the series line. The
version to the right uses LAYER=BACK to draw the line in the back layer. This
prevents the line from covering the data labels and the portions of the series line.

The following code fragment shows the code that positions the line in the back
layer:
layout overlay / walldisplay=(outline)
 About Drawing Layers 1403

xaxisopts=(griddisplay=on display=(line ticks tickvalues))

yaxisopts=(griddisplay=on display=(line ticks tickvalues));
seriesplot x=open y=close / datalabel=date;
drawline x1=0 y1=0 x2=100 y2=100 /
x1space=wallpercent y1space=wallpercent
x2space=wallpercent y2space=wallpercent
lineattrs=(color=cyan thickness=6) layer=back ;
endlayout;

The Back Layer is Behind the Background. Although drawing elements in the back
layer prevents them from obstructing other data elements in the graph, it is not
always the right solution to the problem.

If a draw statement uses LAYER=BACK, then it draws the element behind all other
graphics elements, such as the layout background or a discrete legend’s
background. To ensure that the element is visible in the graph, you might have to do
one or more of the following:
n In overlay-type layouts or in a SCATTERPLOTMATRIX, you can use the parent
layout’s WALLDISPLAY= option to turn off the display of the plot wall. In the
example code just shown, WALLDISPLAY=(OUTLINE) displays an outline
around the graph wall but does not display the wall fill. Suppressing the fill
ensures that the drawn line is visible behind the plot wall.
n If a layout container uses OPAQUE=TRUE so that it can set visual attributes for
the background, then the opaque background covers and therefore hides any
drawn element that is behind the background. When assigning visual attributes
to a graph background, therefore, it might be better to use TRANSPARENCY=
rather than LAYER= to prevent drawn elements from covering other graphics
elements in the graph.
n If the results of a plot statement or other GTL-statement covers the drawn
element, then you can use transparency to reveal the drawn element. For
example, you could use a plot statement’s DATATRANSPARENCY= option to
set an appropriate transparency level for the plot.
1404 Chapter 13 / Key Concepts for Using Draw Statements
 1405

14
Draw Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405
BEGINPOLYGON Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405
BEGINPOLYLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1413
DRAWARROW Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1420
DRAWIMAGE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1427
DRAWLINE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1437
DRAWOVAL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1444
DRAWRECTANGLE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1452
DRAWTEXT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1461

Dictionary

BEGINPOLYGON Statement
Defines the starting point for drawing a polygon in the graph.

Syntax
BEGINPOLYGON X=constant | scalar-expression
Y=constant | scalar-expression </options>;
DRAW X=constant | scalar-expression
Y=constant | scalar-expression </options>;
<… more DRAW statements …>
ENDPOLYGON;
1406 Chapter 14 / Draw Statements

Summary of Optional Arguments

Appearance options
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the polygon.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this BEGINPOLYGON block.
LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYGON block’s output is
drawn.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the outline attributes for this BEGINPOLYGON block.
TRANSPARENCY=number
specifies the degree of the transparency of this BEGINPOLYGON block’s
output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this
BEGINPOLYGON block.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.

ODS options
URL="string" | string-expression
 BEGINPOLYGON Statement 1407

specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
The following options are required in the BEGINPOLYGON statement and the DRAW
statement.
X=constant | scalar-expression
specifies the X value for a point in the polygon. When used in the
BEGINPOLYGON statement, it specifies the X value of the starting point of the
polygon.

Interaction The value that is set for this argument is interpreted using the
XSPACE= option. When XSPACE=DATAVALUE, the value is
interpreted using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value for a point in the polygon. When used in the
BEGINPOLYGON statement, it specifies the Y value of the starting point of the
polygon.

Interaction The value that is set for this argument is interpreted using the
YSPACE= option. When YSPACE=DATAVALUE, the value is
interpreted using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
The following options can be used in the BEGINPOLYGON statement.
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.
1408 Chapter 14 / Draw Statements

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the polygon.

STANDARD
displays an outlined polygon.

ALL
displays an outlined, filled polygon.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or both of the following options:

OUTLINE
displays an outlined polygon

FILL
displays a filled polygon

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the polygon.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this BEGINPOLYGON
block.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this BEGINPOLYGON statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the XSPACE= and
YSPACE= options.

See “About the Drawing Space and Drawing Units” on page 1398

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this BEGINPOLYGON block.

Default The GraphAnnoShape style element

 BEGINPOLYGON Statement 1409

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outlines but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYGON block’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this BEGINPOLYGON block.

Default The ContrastColor, LineStyle, and LineThickness attributes of the

GraphAnnoLine style element.

Interaction For this option to have any effect, the outline must be enabled by
the DISPLAY= option.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element .

“Line Options” on page 1602 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this BEGINPOLYGON block’s output.

Default 0

Range 0 (opaque) to 1 (entirely transparent)

Tip The FILLATTRS option can be used to set transparency for just the
polygon’s filled area. You can combine this option with FILLATTRS= to
set one transparency for the outlines but a different transparency for
the fill. Example:
transparency=0.2 fillattrs=(transparency=0.6)
1410 Chapter 14 / Draw Statements

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

 BEGINPOLYGON Statement 1411

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

DRAW Statement Optional Arguments

The following options can be used in the DRAW statement.

Statement Option Description

DRAWSPACE Specifies a default drawing space and drawing units for the
drawn lines.

XSPACE Specifies the drawing space and drawing units for interpreting
the X value.

YSPACE Specifies the drawing space and drawing units for interpreting
the Y value.

Details

Statement Description
A polygon is built by using the BEGINPOLYGON statement to specify the polygon’s
starting X,Y coordinate, and then specifying the remaining points by nesting a series
of DRAW statements (see “DRAW Statement” on page 1411 ) within the
BEGINPOLYGON block. The block is closed with an ENDPOLYGON statement. To
manage the location and drawing units for the polygon, you can use the XAXIS=,
YAXIS=, XSPACE=, and YSPACE= options.

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

DRAW Statement
The nested DRAW statements within a BEGINPOLYGON block specify a series of
points for a polygon. Each DRAW statement draws a straight line from the previous
point to the endpoint that is specified in the DRAW statement's X and Y arguments.
The first DRAW statement starts its line from the X,Y point that is specified in the
BEGINPOLYGON statement.

You can specify as many DRAW statements as needed to complete the polygon. If
the last DRAW statement does not end at the first point in the polygon (specified
 1412 Chapter 14 / Draw Statements

in BEGINPOLYGON), then a line is automatically generated to close the polygon

shape. For a specific example, see the “Example Program” on page 1412.

Example: BEGINPOLYGON Statement

The following graph was generated by the “Example Program” on page 1412. The
example uses a BEGINPOLYGON block to draw a polygon around the Setosa
species of Iris in a plot that shows petal sizes for three Iris species. The
DRAWSPACE= option in the BEGINGRAPH statements specifies that the polygon
be drawn in the data space. The BEGINPOLYGON statement specifies the starting
X,Y point. For the BEGINPOLYGON options, DISPLAY= displays only the fill for the
polygon. TRANSPARENCY= adds a degree of transparency to the fill, and
FILLATTRS= sets the fill color to yellow. The example also uses DRAWARROW and
DRAWTEXT statements to draw an annotation for the polygon.

Example Program
proc template;
define statgraph discretelegend;
begingraph / drawspace=datavalue;
entrytitle "Iris Petal Dimensions";
layout overlayequated / equatetype=equate;
scatterplot x=petallength y=petalwidth / group=species
name="s";
ellipse x=petallength y=petalwidth / type=predicted alpha=.2
name="p80" legendlabel="80%" outlineattrs=graphconfidence;
 BEGINPOLYLINE Statement 1413

ellipse x=petallength y=petalwidth / type=predicted alpha=.05

name="p95" legendlabel="95%" outlineattrs=graphconfidence2;
beginpolygon x=9 y=2 /display=(fill)
fillAttrs=(color=yellow transparency=0.75);
draw x=13 y=5;
draw x=16 y=7;
draw x=17 y=6;
draw x=20 y=5;
draw x=20 y=1;
draw x=17 y=1;
draw x=15 y=0;
draw x=14 y=0;
draw x=11 y=0;
endpolygon;
drawtext textattrs=(size=11pt) "Setosa" /
x=33 y=0 width=15 anchor=top;
drawarrow x1=28 y1=-1 x2=19 y2=2 /
lineattrs=(thickness=1px)
arrowheadshape=barbed arrowheadscale=0.1;
discretelegend "s" / title="Species:";
discretelegend "p80" "p95" /across=1 autoalign=(topleft)
location=inside;
endlayout;
entryfootnote halign=left "Fisher's Iris Data" ;
endgraph;
end;
run;

proc sgrender data=sashelp.iris template=discretelegend;

run;

BEGINPOLYLINE Statement
Defines the starting point for drawing a polyline in the graph.

Syntax
BEGINPOLYLINE X=constant | scalar-expression
Y=constant | scalar-expression </options>;
DRAW X=constant | scalar-expression
Y=constant | scalar-expression </options>;
<… more DRAW statements …>
ENDPOLYLINE;

Summary of Optional Arguments

Appearance options
1414 Chapter 14 / Draw Statements

LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYLINE block’s output is
drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this BEGINPOLYLINE block.
TRANSPARENCY=number
specifies the degree of the transparency of the BEGINPOLYLINE block’s
output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this
BEGINPOLYLINE block.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
The following options are required in the BEGINPOLYLINE statement and the DRAW
statement.
X=constant | scalar-expression
specifies the X value for the starting point in the polyline.
 BEGINPOLYLINE Statement 1415

Interaction The value that is set for this argument is interpreted using the
XSPACE= option. When XSPACE=DATAVALUE, the value is
interpreted using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value for the starting point in the polyline.

Interaction The value that is set for this argument is interpreted using the
YSPACE= option. When YSPACE=DATAVALUE, the value is
interpreted using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
The following options can be used in the BEGINPOLYLINE statement.
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this BEGINPOLYLINE
block.
1416 Chapter 14 / Draw Statements

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this BEGINPOLYLINE statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the XSPACE= and
YSPACE= options.

See “About the Drawing Space and Drawing Units” on page 1398

LAYER=FRONT | BACK
specifies the layer on which this BEGINPOLYLINE block’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this BEGINPOLYLINE block.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element .

“Line Options” on page 1602 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of the BEGINPOLYLINE block’s output.

Default 0

Range 0 (opaque) to 1 (entirely transparent)

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

 BEGINPOLYLINE Statement 1417

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

DRAW Statement Optional Arguments

The following options can be used in the DRAW statement.
1418 Chapter 14 / Draw Statements

Statement Option Description

DRAWSPACE Specifies a default drawing space and drawing units for the
drawn lines.

XSPACE Specifies the drawing space and drawing units for interpreting
the X value.

YSPACE Specifies the drawing space and drawing units for interpreting
the Y value.

Details

Statement Description
A polyline is built by using the BEGINPOLYLINE statement to specify the polyline’s
starting X,Y coordinate, and then specifying the remaining points by nesting a series
of DRAW statements (see “DRAW Statement” on page 1418 ) within the
BEGINPOLYLINE block. The block is closed with an ENDPOLYLINE statement. To
manage the location and drawing units for the polyline, you can use the XAXIS=,
YAXIS=, XSPACE=, and YSPACE= options.

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

DRAW Statement
The nested DRAW statements within a BEGINPOLYLINE block specify a series of
points for a polyline. Each DRAW statement draws a straight line from the previous
point to the endpoint that is specified in the DRAW statement's X and Y arguments.
The first DRAW statement starts its line from the X,Y point that is specified in the
BEGINPOLYLINE statement. You can specify as many DRAW statements as needed
to complete the polyline.

For a specific example, see the “Example Program” on page 1419.

Example: BEGINPOLYLINE Statement

The following graph was generated by the “Example Program” on page 1419. The
example uses two BEGINPOLYLINE blocks to highlight two student data points.
The DRAWSPACE= option in the BEGINGRAPH statements specifies that the
polyline be drawn in the data space. The BEGINPOLYLINE statements specify the
starting X,Y points for two polyline, and the DRAW statements complete the lines.
 BEGINPOLYLINE Statement 1419

The example also uses two DRAWTEXT statements to label the data points of
interest.

Example Program
proc template;
define statgraph drawoval;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawtext "Alfred" / x=69 y=112 anchor=top;
drawtext "Barbara" / x=65.4 y=97 anchor=top width=15;
beginpolyline x=69 y=105 ;
draw x=69 y=85 ;
draw x=65.3 y=85 ;
draw x=65.3 y=90 ;
endpolyline ;
beginpolyline x=67 y=85 ;
draw x=67 y=78 ;
endpolyline ;
drawtext "New in Class" / x=67 y=77 anchor=top width=15;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=drawoval;

run;
 1420 Chapter 14 / Draw Statements

DRAWARROW Statement
Draws an arrow (a directed line segment) from one point to another point.

Syntax
DRAWARROW X1=constant | scalar-expression
Y1=constant | scalar-expression
X2=constant | scalar-expression
Y2=constant | scalar-expression < /options >;

Summary of Optional Arguments

Appearance options
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this draw statement.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Arrow options
ARROWHEADDIRECTION=OUT | IN | BOTH
specifies the direction of the arrowhead(s) at the end(s) of the arrow
shaft.
ARROWHEADSCALE=positive-number
specifies an arrowhead scale factor based on the thickness of the arrow
shaft.
ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED
specifies the shape of the arrowhead(s).

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
 DRAWARROW Statement 1421

specifies whether the data value for the arguments Y1 and Y2 is

interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the X1
value.
X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the X2
value
Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1
value.
Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2
value

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
X1=constant | scalar-expression
specifies the X value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
X1SPACE= option.

When X1SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
1422 Chapter 14 / Draw Statements

Y1=constant | scalar-expression
specifies the Y value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
Y1SPACE= option.

When Y1SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

X2=constant | scalar-expression
specifies the X value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
X2SPACE= option.

When X2SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y2=constant | scalar-expression
specifies the Y value of one arrow-shaft endpoint.

Interactions This value that is set for this option is interpreted using the
Y2SPACE= option.

When Y2SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
ARROWHEADDIRECTION=OUT | IN | BOTH
specifies the direction of the arrowhead(s) at the end(s) of the arrow shaft.

OUT
specifies a single arrowhead drawn at (X2,Y2) and pointing away from
(X1,Y1)

IN
specifies a single arrowhead drawn at (X1,Y1) and pointing away from
(X2,Y2)

BOTH
specifies two arrowheads, one at the IN position and one at the OUT
position
 DRAWARROW Statement 1423

Default OUT

Tip Use the ARROWHEADSHAPE= option to control the arrowhead

appearance.

ARROWHEADSHAPE=OPEN | CLOSED | FILLED | BARBED

specifies the shape of the arrowhead(s). The following figure shows the
arrowhead shapes.

Default OPEN

Tip Use the ARROWHEADDIRECTION= option to control the arrow

direction.

ARROWHEADSCALE=positive-number
specifies an arrowhead scale factor based on the thickness of the arrow shaft.

Default 1.0

Restriction The minimum size for arrowheads is 8 pixels. If you specify a value
for ARROWHEADSCALE= that scales the arrowhead below 8 pixels,
an 8-pixel arrowhead is used instead. There is no restriction on the
maximum size.

Tip Use a factor greater than 1.0 to make a larger arrowhead.

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options X1SPACE=, X2SPACE=,
Y1SPACE=, or Y2SPACE= use DATAVALUE, and when X1, X2, Y1, or
Y2 are values on a discrete axis. For nondiscrete axes, this option is
ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
1424 Chapter 14 / Draw Statements

the OFFSETMIN= and OFFSETMAX= axis options to increase the

axis minimum and maximum offsets to accommodate the discrete
offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options
X1SPACE=, Y1SPACE=, X2SPACE=, or Y2SPACE=.

See “About the Drawing Space and Drawing Units” on page 1398

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this draw statement.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

 DRAWARROW Statement 1425

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if X1SPACE=DATAVALUE.

X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X1
value.

X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X2
value.
1426 Chapter 14 / Draw Statements

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if Y1SPACE=DATAVALUE.

Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y1
value.

Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y2
value.

Details
A DRAWARROW statement draws a line (arrow shaft) from a specified starting
point (X1,Y1) to a specified ending point (X2,Y2). It also displays an arrowhead at
either or both ends of the line. DRAWARROW is similar to a DRAWLINE statement,
using many of the same options, but it has additional options for controlling the
arrowhead(s).

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWARROW Statement

The following graph was generated by the “Example Program” on page 1427. The
example shows a common application of a DRAWARROW and DRAWTEXT
statements to identify a specific part of the graph and add explanatory text.
 DRAWIMAGE Statement 1427

Example Program
proc template;
define statgraph arrow;
begingraph;
entrytitle "Micosoft Stock Prices between 2000 and 2002";
layout overlay;
seriesplot x=date y=close;
drawarrow x1="01NOV2001"d y1=75 x2="01NOV2001"d y2=64.21 /
x1space=datavalue y1space=wallpercent
x2space=datavalue y2space=datavalue
arrowheadshape=filled lineattrs=(color=red) ;
drawtext "Introduction of Windows XP" / width=25 anchor=bottom
border=true borderattrs=(color=red)
x="01NOV2001"d y=75 xspace=datavalue yspace=wallpercent;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=arrow;

where stock="Microsoft" and Year(date) between 2000 and 2002;
run;

DRAWIMAGE Statement
Draws an image in the graph.
1428 Chapter 14 / Draw Statements

Syntax
DRAWIMAGE "image-file-spec" < /options >;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the image.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the border line attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
ROTATE=number
specifies the angle of rotation for the image, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.
 DRAWIMAGE Statement 1429

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the image.
X=constant | scalar-expression
specifies the anchor point’s X coordinate.
Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHT=positive-number
specifies the height of the image’s bounding box.
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.
SCALE=FIT | FITHEIGHT | FITWIDTH | TILE
specifies how the image is scaled within the bounding box.
SIZEUNIT=PERCENT | PIXEL | DATA
specifies whether the default units for the size of the image’s bounding
box are percentage values, or pixel values, or values that are in the unit of
the data.
WIDTH=positive-number
specifies the width of the image’s bounding box.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.

Required Argument
image-file-spec
specifies the name, image type, and location of the image. The image-file-spec
value must be enclosed in double quotation marks and must be specified as a
local, physical file path (for example, "c:\temp\saslogo.gif"). The supported
image types are GIF, JPEG, and PNG raster or bitmap format.

Restrictions URL access to image files is not supported. The image file must
exist on the file system.

In SAS 9.4 and earlier releases, you cannot use a dynamic variable
for the image-file-spec value. This restriction is removed starting
with SAS 9.4M1.
1430 Chapter 14 / Draw Statements

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT | BOTTOMLEFT
| BOTTOM | BOTTOMRIGHT
specifies an anchor point for the image. The anchor point is relative to the
unrotated image. The anchor point can be at the center of the image or at eight
points on the border of the image bounding box. The following figure shows the
anchor points for TOPLEFT and LEFT.

The coordinates of the anchor point are set by the X= and Y= options, and by
the XSPACE= and YSPACE= options. The XAXIS= and YAXIS= option might
affect positioning when the XSPACE= or YSPACE= options are set to
DATAPIXEL, DATAPERCENT, or DATAVALUE.

The image has a fixed height and a fixed width, determined by the HEIGHT=,
HEIGHTUNIT=, WIDTH= and WIDTHUNIT= options. The height of the text
grows in a direction that is related to the anchor point. For example, if
ANCHOR=TOPLEFT, then the image height extends downward from the anchor
point and its width extends to the right. If ANCHOR=CENTER, then half the
image width and half the image height extend equally left and right, as well as
top to bottom from the anchor point. If ANCHOR=BOTTOM, then the image
height extends upward from the anchor point and the image width is centered at
the anchor point.

When the image is rotated, the anchor point remains relative to the unrotated
image while the image is rotated on its anchor point. See ROTATE= on page
1432.

Default CENTER

BORDER=TRUE | FALSE
specifies whether a border is drawn around the image.

Default FALSE

Tip Use the BORDERATTRS= option to control the appearance of the

border.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the border line attributes for this draw statement.

Default The GraphBorderLines style element.

Interaction BORDER=TRUE must be set for this option to have any effect.
 DRAWIMAGE Statement 1431

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1398

HEIGHT=positive-number
specifies the height of the image’s bounding box.

Default The height of the image.

Interaction The interpretation of this height setting is determined by the

combined settings of the HEIGHTUNIT= and YSPACE= options.
1432 Chapter 14 / Draw Statements

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the
height that is set in the HEIGHT= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

ROTATE=number
specifies the angle of rotation for the image, measured in degrees. The image is
rotated around its anchor point. The angle is measured from a horizontal line
passing through the anchor point of the image to the right. The following figure
shows the rotation of an image around a top anchor point.

Positive angles rotate the image counter clockwise, and negative angles rotate
the image clockwise. The angle specification can exceed 360 degrees in
absolute value.

Default 0. No rotation is performed.

Note If you specify the SVG format for your graph and your graph includes a
rotated image, then the graph is rendered as a bitmapped image rather
 DRAWIMAGE Statement 1433

than an SVG image. In that case, a note is written to the SAS log
indicating the change in format.

See ANCHOR= on page 1430

SCALE=FIT | FITHEIGHT | FITWIDTH | TILE

specifies how the image is scaled within the bounding box.

FIT
scales the image to fit the bounding box.

Aspect ratio is not maintained.

FITHEIGHT
scales the image to fit the height of the bounding box.

The width is computed from the height and the image's aspect ratio.

FITWIDTH
scales the image to fit the width of the bounding box.

The height is computed from the width and the image's aspect ratio.

TILE
tiles the image as needed to fit the bounding image.

The last tile in a row or column might be clipped by the bounding box.

Default FIT

SIZEUNIT=PERCENT | PIXEL | DATA

specifies whether the default units for the size of the image’s bounding box are
percentage values, or pixel values, or values that are in the unit of the data.

Default PERCENT

Interaction If the HEIGHTUNIT= or WIDTHUNIT= option is also used, then it

overrides this option for that dimension.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Note This option also affects the image border.

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

1434 Chapter 14 / Draw Statements

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

WIDTH=positive-number
specifies the width of the image’s bounding box.

Default The width of the image.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the
width that is set in the WIDTH= option.

X=constant | scalar-expression
specifies the anchor point’s X coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
XSPACE= option.

If XSPACE=DATAVALUE, then this option's value is interpreted

using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.
 DRAWIMAGE Statement 1435

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
YSPACE= option.

If YSPACE=DATAVALUE, then this option's value is interpreted

using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWIMAGE statement draws an image in a graph. By default, the image is
drawn in the center of the graph. You can change the default position with the
options ANCHOR=, X=, Y=, XSPACE=, and YSPACE=. By default, the image is drawn
in the actual image size. You can change the default size with the WIDTH= and
HEIGHT= options.

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
 1436 Chapter 14 / Draw Statements

drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWIMAGE Statement

Example Graph
The following graph was generated by the “Example Program” on page 1436. The
example shows how to display an image in the bottom right corner of the graph
wall.

Example Program
proc template;
define statgraph image;
begingraph;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight / primary=true;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawimage "c:\temp\saslogo.gif" /
 DRAWLINE Statement 1437

anchor=bottomright x=98 y=2

drawspace=wallpercent ;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=image;

run;

DRAWLINE Statement
Draws a line from one point to another point.

Syntax
DRAWLINE X1=constant | scalar-expression
Y1=constant | scalar-expression
X2=constant | scalar-expression
Y2=constant | scalar-expression < /options >;

Summary of Optional Arguments

Appearance options
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
LINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the line attributes for this draw statement.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

1438 Chapter 14 / Draw Statements

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the X1
value.
X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the X2
value
Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1
value.
Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2
value

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Required Arguments
X1=constant | scalar-expression
specifies the X value of the starting point for the line.

Interactions This value that is set for this option is interpreted using the
X1SPACE= option.

When X1SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y1=constant | scalar-expression
specifies the Y value of the starting point for the line.

Interactions This value that is set for this option is interpreted using the
Y1SPACE= option.
 DRAWLINE Statement 1439

When Y1SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y1SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

X2=constant | scalar-expression
specifies the X value of the endpoint for the line.

Interactions This value that is set for this option is interpreted using the
X2SPACE= option.

When X2SPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When X2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y2=constant | scalar-expression
specifies the Y value of the endpoint for the line.

Interactions This value that is set for this option is interpreted using the
Y2SPACE= option.

When Y2SPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When Y2SPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Optional Arguments
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options X1SPACE=, X2SPACE=,
Y1SPACE=, or Y2SPACE= use DATAVALUE, and when X1, X2, Y1, or
Y2 are values on a discrete axis. For nondiscrete axes, this option is
ignored.
1440 Chapter 14 / Draw Statements

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options
X1SPACE=, Y1SPACE=, X2SPACE=, or Y2SPACE=.

See “About the Drawing Space and Drawing Units” on page 1398

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes for this draw statement.

Default The GraphAnnoLine style element.

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Line Options” on page 1602 for available line-options.

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.
 DRAWLINE Statement 1441

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if X1SPACE=DATAVALUE.

X1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the X1
value.

X2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the X2 value

Default The setting that is in effect for the DRAWSPACE= option.

1442 Chapter 14 / Draw Statements

Interaction This option overrides the DRAWSPACE= setting only for the X2
value.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if Y1SPACE=DATAVALUE.

Y1SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y1 value.

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y1
value.

Y2SPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL

| WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the Y2 value

Default The setting that is in effect for the DRAWSPACE= option.

Interaction This option overrides the DRAWSPACE= setting only for the Y2
value.

Details
A DRAWLINE statement draws a line from a starting point that is specified with the
X1 and Y1 arguments, to an ending point that is specified with the X2 and Y2
arguments. DRAWLINE is similar to a DRAWARROW statement, using many of the
same options, but without the options for controlling the arrowhead(s).

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWLINE Statement

The following graph was generated by the “Example Program” on page 1443. The
example shows how to draw a diagonal reference line. One endpoint is point 0,0
and the other is point 100,100. Both points are specified in the WALL area with
PERCENT units, making it easy to position the line without regard to the axis data
ranges or the axis offsets. To draw the line behind the series line and grid lines, you
 DRAWLINE Statement 1443

can set LAYER= BACK and use the parent layout statement to turn off the display
of the wall.

Example Program
proc template;
define statgraph diagonal;
begingraph;
entrytitle "Open vs. Close Price for Intel Stock 2003";
layout overlay / walldisplay=(outline)
xaxisopts=(griddisplay=on)
yaxisopts=(griddisplay=on);
seriesplot x=open y=close / datalabel=date;
drawline x1=0 y1=0 x2=100 y2=100 /
x1space=wallpercent y1space=wallpercent
x2space=wallpercent y2space=wallpercent
lineattrs=GraphReference layer=back ;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.stocks template=diagonal;

where stock="Intel" and year(date)=2003;
format date monname3.;
run;
 1444 Chapter 14 / Draw Statements

DRAWOVAL Statement
Draws an oval in the graph.

Syntax
DRAWOVAL X=constant | scalar-expression
Y=constant | scalar-expression
WIDTH=constant | scalar-expression
HEIGHT=constant | scalar-expression < /options >;

Summary of Optional Arguments

Appearance options
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the oval.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the outline attributes for this draw statement.
ROTATE=number
specifies the angle of rotation for the oval, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

 DRAWOVAL Statement 1445

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the oval.

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.

Required Arguments
X=constant | scalar-expression
specifies the X value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
XSPACE= option.

When XSPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
1446 Chapter 14 / Draw Statements

Y=constant | scalar-expression
specifies the Y value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
YSPACE= option.

When YSPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

WIDTH=constant | scalar-expression
specifies the width of the oval.

Interactions The value that is set for this argument is interpreted using the
WIDTHUNIT= and XSPACE= options.

When WIDTHUNIT=DATA, the value is interpreted using the

XAXIS= option.

HEIGHT=constant | scalar-expression
specifies the height of the oval.

Interactions The value that is set for this argument is interpreted using the
HEIGHTUNIT= and YSPACE= options.

When HEIGHTUNIT=DATA, the value is interpreted using the

YAXIS= option.

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT | BOTTOMLEFT
| BOTTOM | BOTTOMRIGHT
specifies an anchor point for the oval. The anchor point is relative to the
unrotated oval. The anchor point can be at the center of the oval or at eight
points on the bounding box of the rectangle. The following figure shows the
anchor points for TOPLEFT and CENTER.

The coordinates of the anchor point are set by the statement’s X and Y values,
and by the settings for the XSPACE= and YSPACE= options. The XAXIS= and
YAXIS= option might affect positioning when the XSPACE= or YSPACE= options
are set to DATAPIXEL, DATAPERCENT, or DATAVALUE.

When the oval is rotated, the anchor point remains relative to the unrotated
oval while the oval is rotated on its anchor point. See ROTATE= on page 1449.
 DRAWOVAL Statement 1447

Default CENTER

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the oval.

STANDARD
displays an outlined oval.

ALL
displays an outlined, filled oval.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or both of the following options:

OUTLINE displays an outlined oval

FILL displays a filled oval

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the oval.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT
1448 Chapter 14 / Draw Statements

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1398

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this draw statement.

Default The GraphAnnoShape style element Color attribute

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outline but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the
height that is set in the HEIGHT= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .
 DRAWOVAL Statement 1449

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this draw statement.

Default The GraphAnnoLine style element.

Interaction For this option to have any effect, the outline must be enabled by
the DISPLAY= option or by the ODS style that is in effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ROTATE=number
specifies the angle of rotation for the oval, measured in degrees. The angle is
measured as if a horizontal line extended to the right through the oval anchor
point as shown in the following figure.

Positive angles rotate the oval counter clockwise, and negative angles rotate
the oval clockwise. The angle specification can exceed 360 degrees in absolute
value.

Default 0. No rotation is performed

See ANCHOR= on page 1446

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
filled area. You can combine this option with FILLATTRS= to set one
transparency for the outline but a different transparency for the fill.
Here is an example:
transparency=0.2 fillattrs=(transparency=0.6)

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.
1450 Chapter 14 / Draw Statements

Note: This option is valid starting in SAS 9.4M1.

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the
width that is set in the WIDTH= option.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y
 DRAWOVAL Statement 1451

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWOVAL statement draws an oval in a graph. The oval position is determined
by the X and Y anchor points, and the size is determined by the HEIGHT and
WIDTH settings. You can manage the oval position with the options ANCHOR= ,
XSPACE= , and YSPACE= . You can manage the oval size with the HEIGHTUNIT=
and WIDTHUNIT= options.

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWOVAL Statement

The following graph was generated by the “Example Program” on page 1452. The
example uses DRAWOVAL to highlight a student’s data point. It draws an oval
around the marker symbol that represents the student’s height and weight, and it
displays the student’s name inside the oval. In the BEGINGRAPH statement, the
setting for the DRAWSPACE= option sets the drawing space and drawing units for
the DRAWOVAL and DRAWTEXT statements. In the DRAWOVAL statement,
setting TRANSPARENCY=0.75 ensures that the marker for Alfred is visible behind
the oval. The DRAWTEXT statement draws the text that identifies the student's
name, using the ANCHOR=, X=, and Y= options to position the text within the oval.
 1452 Chapter 14 / Draw Statements

Example Program
proc template;
define statgraph drawoval;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawoval x=69 y=112.5 width=15 height=20 /
display=all fillattrs=(color=green)
transparency=0.75 ;
drawtext "Alfred" / x=69 y=112 anchor=top;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=drawoval;

run;

DRAWRECTANGLE Statement
Draws a rectangle in the graph.
 DRAWRECTANGLE Statement 1453

Syntax
DRAWRECTANGLE X=constant | scalar-expression
Y=constant | scalar-expression
WIDTH=constant | scalar-expression
HEIGHT=constant | scalar-expression < /options >;

Summary of Optional Arguments

Appearance options
CORNERRADIUS=number
specifies the radius of the rectangle corners.
DISPLAY=STANDARD | ALL | (display-options)
specifies the features to display for the rectangle.
FILLATTRS=style-element | style-element (fill-options) | (fill-options)
specifies the fill attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
OUTLINEATTRS=style-element | style-element (line-options) | (line-options)
specifies the outline attributes for this draw statement.
ROTATE=number
specifies the angle of rotation for the rectangle, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
1454 Chapter 14 / Draw Statements

specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the rectangle.

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
HEIGHTUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the HEIGHT=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.

Required Arguments
X=constant | scalar-expression
specifies the X value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
XSPACE= option.

When XSPACE=DATAVALUE, the value is interpreted using the

XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

Y=constant | scalar-expression
specifies the Y value of the anchor point.

Interactions The value that is set for this argument is interpreted using the
YSPACE= option.

When YSPACE=DATAVALUE, the value is interpreted using the

YAXIS= option.
 DRAWRECTANGLE Statement 1455

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

WIDTH=constant | scalar-expression
specifies the width of the rectangle.

Interactions The value that is set for this argument is interpreted using the
WIDTHUNIT= and XSPACE= options.

When WIDTHUNIT=DATA, the value is interpreted using the

XAXIS= option.

HEIGHT=constant | scalar-expression
specifies the height of the rectangle.

Interactions The value that is set for this argument is interpreted using the
HEIGHTUNIT= and YSPACE= options.

When HEIGHTUNIT=DATA, the value is interpreted using the

YAXIS= option.

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT | BOTTOMLEFT
| BOTTOM | BOTTOMRIGHT
specifies an anchor point for the rectangle. The anchor point is relative to the
unrotated rectangle. The anchor point can be at the center of the rectangle or at
eight points on the rectangle border. The following figure shows the anchor
points for TOPLEFT and CENTER.

The coordinates of the anchor point are set by the statement’s X and Y values,
and by the settings for the XSPACE= and YSPACE= options. The XAXIS= and
YAXIS= option might affect positioning when the XSPACE= or YSPACE= options
are set to DATAPIXEL, DATAPERCENT, or DATAVALUE.

When the rectangle is rotated, the anchor point remains relative to the
unrotated rectangle while the rectangle is rotated on its anchor point. See
ROTATE= on page 1458.

Default CENTER

CORNERRADIUS=number
specifies the radius of the rectangle corners.

Default 0
1456 Chapter 14 / Draw Statements

Range 0–1, where 0 specifies square corners and 1 specifies the most rounded
corners

Example cornerradius=0.2

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DISPLAY=STANDARD | ALL | (display-options)

specifies the features to display for the rectangle.

STANDARD
displays an outlined rectangle.

ALL
displays an outlined, filled rectangle.

(display-options)
specifies a space-separated list of one or more display options, enclosed in
parentheses.

The list can contain one or bothj of the following options:

OUTLINE displays an outlined rectangle

FILL displays a filled rectangle

Default STANDARD

Tip Use the OUTLINEATTRS= and FILLATTRS= options to control the

appearance of the rectangle.
 DRAWRECTANGLE Statement 1457

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1398

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

specifies the fill attributes for this draw statement.

Default The GraphAnnoShape style element Color attribute

Tip The TRANSPARENCY= option sets the transparency for the fill and the
outline. You can combine this option with TRANSPARENCY= to set one
transparency for the outline but a different transparency for the fill.
Example:
transparency=0.2 fillattrs=(transparency=0.6)

See “General Syntax for Attribute Options” on page 1599 for the syntax on
using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

HEIGHTUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the HEIGHT= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the YSPACE= setting to interpret the
height that is set in the HEIGHT= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT
1458 Chapter 14 / Draw Statements

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

OUTLINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the outline attributes for this draw statement.

Default The GraphAnnoLine style element.

Interaction For this option to have any effect, the outline must be enabled by
the DISPLAY= option or by the ODS style that is in effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

ROTATE=number
specifies the angle of rotation for the rectangle, measured in degrees. The angle
is measured as if a horizontal line extended to the right through the rectangle
anchor point as shown in the following figure.

Positive angles rotate the rectangle counter clockwise, and negative angles
rotate the rectangle clockwise. The angle specification can exceed 360 degrees
in absolute value.

Default 0. No rotation is performed

See ANCHOR= on page 1455

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

Tip The FILLATTRS= option can be used to set transparency for just the
filled area. You can combine this option with FILLATTRS= to set one
transparency for the outline but a different transparency for the fill.
Here is an example:
 DRAWRECTANGLE Statement 1459

transparency=0.2 fillattrs=(transparency=0.6)

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the
width that is set in the WIDTH= option.

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

1460 Chapter 14 / Draw Statements

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Details
A DRAWRECTANGLE statement draws a rectangle in a graph. The rectangle
position is determined by the X and Y anchor points, and the size is determined by
the HEIGHT and WIDTH settings. You can manage the rectangle position with the
options ANCHOR= , XSPACE= , and YSPACE= . You can manage the rectangle size
with the HEIGHTUNIT= and WIDTHUNIT= options.

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWRECTANGLE Statement

The following graph was generated by the “Example Program” on page 1461. The
example uses DRAWRECTANGLE to highlight a student’s data point. It draws a
rectangle around the marker symbol that represents the student’s height and
weight, and it displays the student’s name inside the rectangle. In the BEGINGRAPH
statement, the setting for the DRAWSPACE= option sets the drawing space and
drawing units for the DRAWRECTANGLE and DRAWTEXT statements. In the
DRAWRECTANGLE statement, setting TRANSPARENCY=0.75 ensures that the
marker for Alfred is visible behind the rectangle. The DRAWTEXT statement draws
the text that identifies the student's name, using the ANCHOR=, X=, and Y= options
to position the text within the rectangle.
 DRAWTEXT Statement 1461

Example Program
proc template;
define statgraph drawrectangle;
begingraph / drawspace=datavalue;
entrytitle "Regression Fit Plot";
layout overlay;
modelband "myclm";
scatterplot x=height y=weight;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
drawrectangle x=69 y=112.5 width=10 height=15 /
display=all fillattrs=(color=green)
transparency=0.75 ;
drawtext "Alfred" / x=69 y=112 anchor=top;
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=drawrectangle;

run;

DRAWTEXT Statement
Draws and anchors in a graph a text box that contains one or more lines of formatted text.
1462 Chapter 14 / Draw Statements

Syntax
DRAWTEXT text-item <text-item …> < /options >;

Summary of Optional Arguments

Appearance options
BORDER=TRUE | FALSE
specifies whether a border is drawn around the text box.
BORDERATTRS=style-element | style-element (line-options) | (line-options)
specifies the border line attributes for this draw statement.
LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.
PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the text box’s
border.
ROTATE=number
specifies the angle of rotation of the text box, measured in degrees.
TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Axes options
DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.
XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is
interpreted using the primary X-axis scale or to the secondary X (X2) axis
scale.
YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is
interpreted using the primary Y-axis scale or to the secondary Y (Y2) axis
scale.

Drawing space options

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw
statement.
XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |
LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the X= option.
 DRAWTEXT Statement 1463

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies the drawing space and drawing units for interpreting the value
that is specified in the Y= option.

Location options
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT |
BOTTOMLEFT | BOTTOM | BOTTOMRIGHT
specifies an anchor point for the text box.
JUSTIFY=LEFT | CENTER | RIGHT
specifies the alignment of text that wraps within the text box.
X=constant | scalar-expression
specifies the anchor point’s X coordinate.
Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

ODS options
URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Size options
WIDTH=positive-number
specifies the width of the text box.
WIDTHUNIT=PERCENT | PIXEL | DATA
specifies whether the positive-number that is specified in the WIDTH=
option is interpreted as a percentage value, a pixel value, or a value that
is in the units of the data.

Required Argument
text-item <…text-item>
specifies one or more pieces of text for the text box. Each text-item has the
following form:

<prefix-option> "string" | dynamic | character-expression | {text-command}

Each piece of text can have a prefix setting that precedes the piece of text. A
piece of text is either a string literal, a dynamic, or a text command. All text-
items are concatenated into one string, which might be wrapped, based on the
settings for the WIDTH= and WIDTHUNIT= settings. Leading and trailing blanks
in the concatenated string are always used.

When used, a prefix option applies to the immediately following piece of text
and also to all subsequent text strings and text-commands until another prefix
option is specified.

Requirements string must be enclosed in quotation marks.

character-expression must be enclosed in an EVAL function.

text-command must be enclosed in braces.

1464 Chapter 14 / Draw Statements

Note Leading and trailing spaces are preserved.

See Chapter 10, “Managing Text Items,” on page 1337

Optional Arguments
ANCHOR=CENTER | TOPLEFT | TOP | TOPRIGHT | LEFT | RIGHT | BOTTOMLEFT
| BOTTOM | BOTTOMRIGHT
specifies an anchor point for the text box. The anchor point is relative to the
unrotated text. It can be at the center of the text box or at eight points on the
border of the text box bounding box. The following figure shows the anchor
points for TOPLEFT and LEFT.

The coordinates of the anchor point are set by the X= and Y= options, and by
the XSPACE= and YSPACE= options. The XAXIS= and YAXIS= option might
affect positioning when the XSPACE= or YSPACE= options are set to
DATAPIXEL, DATAPERCENT, or DATAVALUE.

The text box has a fixed width, determined by the WIDTH= and WIDTHUNIT=
options. The height of the text box is based on the amount of text specified and
the font size. The height of the text grows in a direction that is related to the
anchor point. For example, if ANCHOR=TOPLEFT, then the text box height
extends downward from the anchor point and its width extends to the right. If
ANCHOR=CENTER, then half the text box width and half the text box height
extend equally left and right, as well as top to bottom from the anchor point. If
ANCHOR=BOTTOM, the text box height extends upward from the anchor point
and the text box width is centered at the anchor point.

When the text is rotated, the anchor point remains relative to the unrotated text
box while the text box is rotated on its anchor point. See ROTATE= on page
1467.

Default CENTER

BORDER=TRUE | FALSE
specifies whether a border is drawn around the text box.

Default FALSE

Tip Use the BORDERATTRS= option to control the appearance of the

border.

See “boolean ” on page xii for other Boolean values that you can use.

BORDERATTRS=style-element | style-element (line-options) | (line-options)

specifies the border line attributes for this draw statement.
 DRAWTEXT Statement 1465

Default The GraphBorderLines style element.

Interaction BORDER=TRUE must be set for this option to have any effect.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

DISCRETEOFFSET=number
specifies an amount to offset from the discrete X values, or the discrete Y
values, or both.

Default 0 (no offset, output is centered on the discrete X values, or the

discrete Y values, or both)

Range –0.5 to +0.5, where 0.5 represents half the distance between
discrete ticks. If the X axis is discrete, then a positive offset is to the
right. If the Y axis is discrete, then a positive offset is up. If
REVERSE=TRUE on the X or Y axis, then the offset direction is also
reversed.

Restriction This option applies only when the options XSPACE= or YSPACE=
use DATAVALUE, and when X or Y are values on a discrete axis. For
nondiscrete axes, this option is ignored.

Tip Setting the discrete offset for the plots does not affect the axis
minimum and maximum offsets. In some cases, setting a discrete
offset can cause clipping at each end of the axis. In those cases, use
the OFFSETMIN= and OFFSETMAX= axis options to increase the
axis minimum and maximum offsets to accommodate the discrete
offset.

DRAWSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT |

LAYOUTPIXEL | WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL |
DATAVALUE
specifies a default drawing space and drawing units for this draw statement.

Default LAYOUTPERCENT

Interactions This statement and all of the draw statements inherit the global
DRAWPSACE= setting from the DRAWSPACE= option in the
BEGINGRAPH statement. Setting this option changes the setting
for only this draw statement.

This option sets the default drawing space, but individual settings
in the X or Y dimension can be overridden by the options XSPACE=,
YSPACE=, HEIGHTUNIT=, or WIDTHUNIT=.

See “About the Drawing Space and Drawing Units” on page 1398

JUSTIFY=LEFT | CENTER | RIGHT

specifies the alignment of text that wraps within the text box.
1466 Chapter 14 / Draw Statements

LEFT
forces the first character of each line to appear at the left margin (distance
from the left border plus the left pad amount).

CENTER
forces each line to be centered in the text box between the left and right pad
amounts.

RIGHT
forces the last character of each line to appear at the right margin (distance
from the right border minus the right pad amount).

Default LEFT

Interaction Text is wrapped based on the width of the specified text, the font
size, and the setting in the PAD= option.

LAYER=FRONT | BACK
specifies the layer on which this draw statement’s output is drawn.

FRONT
draws the output on top of the graph.

BACK
draws the output behind the background areas, such as a layout or legend
background.

Default FRONT

Tip For elements that are obstructed because they are in the back layer,
you can suppress the display of filled areas in the graph. You can also
use transparency to manage the element visibility. For more
information, see “About Drawing Layers” on page 1402 .

PAD=dimension | (pad-options)
specifies the amount of extra space that is reserved inside the text box’s border.

dimension
specifies a dimension to use for the extra space at the left, right, top, and
bottom of the text box border.

(pad-options)
a space-separated list of one or more name-value-pair pad options enclosed
in parentheses.

The list can include one or more of the following pad options:

LEFT=dimension
specifies the amount of extra space added to the left side.

RIGHT=dimension
specifies the amount of extra space added to the right side.

TOP=dimension
specifies the amount of extra space added to the top.

BOTTOM=dimension
specifies the amount of extra space added to the bottom.
 DRAWTEXT Statement 1467

Note Sides that are not assigned padding are padded with the default
amount.

Tip Use pad-options to create non-uniform padding.

Default (LEFT=3 RIGHT=3 TOP=0 BOTTOM=0)

Note The default units for dimension are pixels.

See “dimension” on page xiii

ROTATE=number
specifies the angle of rotation of the text box, measured in degrees. The text
box is rotated around its anchor point. The angle is measured from a horizontal
line passing through the anchor point of the text box to the right. The following
figure shows the rotation of text around a top anchor point.

Positive angles rotate the text box counter clockwise, and negative angles
rotate the text box clockwise. The angle specification can exceed 360 degrees
in absolute value.

Default 0. No rotation is performed

See ANCHOR= on page 1464

TRANSPARENCY=number
specifies the degree of the transparency of this draw statement’s output.

Default 0

Range 0–1, where 0 is opaque and 1 is entirely transparent

URL="string" | string-expression
specifies an HTML page that is displayed when the output of this draw
statement is selected.

Note: This option is valid starting in SAS 9.4M1.

1468 Chapter 14 / Draw Statements

"string"
string-expression
specifies a valid HTML page reference (HREF) for the graphical element that
is drawn by this draw statement.

Example http://www.sas.com

Restriction Links specified with URL= are not supported in SVG output in
HTML. In that case, use a raster image format such as PNG.

Requirement To generate selectable graphical elements, you must include an

ODS GRAPHICS ON statement that specifies the IMAGEMAP
option, and you must write the output to the ODS HTML
destination.

Tip By default, drill-down links open in a new browser window (link

target _blank). To specify a different target for your drill-down
links, such as _self or _parent, add the DRILLTARGET="target"
option to your ODS GRAPHICS statement. See DRILLTARGET= on
page 1568.

WIDTH=positive-number
specifies the width of the text box.

Default 10

Interaction This option's value is interpreted using the WIDTHUNIT= option.

WIDTHUNIT=PERCENT | PIXEL | DATA

specifies whether the positive-number that is specified in the WIDTH= option is
interpreted as a percentage value, a pixel value, or a value that is in the units of
the data.

Default PERCENT

Interaction This setting combines with the XSPACE= setting to interpret the
width that is set in the WIDTH= option.

X=constant | scalar-expression
specifies the anchor point’s X coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
XSPACE= option.

If XSPACE=DATAVALUE, then this option's value is interpreted

using the XAXIS= option.

Note When XSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.
 DRAWTEXT Statement 1469

XAXIS=X | X2
specifies whether the data value for the arguments X1 and X2 is interpreted
using the primary X-axis scale or to the secondary X (X2) axis scale.

Default X

Interaction This option has effect only if XSPACE=DATAVALUE.

XSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the X= option.

Default The setting that is in effect for the DRAWSPACE= option.

Y=constant | scalar-expression
specifies the anchor point’s Y coordinate.

Default 50

Interactions The DRAWSPACE= option determines the default interpretation of

the units for this setting. You can override the default with the
YSPACE= option.

If YSPACE=DATAVALUE, then this option's value is interpreted

using the YAXIS= option.

Note When YSPACE=DATAVALUE and a character value is specified,

leading blanks are honored and trailing blanks are ignored when the
specified value is compared with the data values.

YAXIS=Y | Y2
specifies whether the data value for the arguments Y1 and Y2 is interpreted
using the primary Y-axis scale or to the secondary Y (Y2) axis scale.

Default Y

Interaction This option has effect only if YSPACE=DATAVALUE.

YSPACE=GRAPHPERCENT | GRAPHPIXEL | LAYOUTPERCENT | LAYOUTPIXEL |

WALLPERCENT | WALLPIXEL | DATAPERCENT | DATAPIXEL | DATAVALUE
specifies the drawing space and drawing units for interpreting the value that is
specified in the Y= option.

Default The setting that is in effect for the DRAWSPACE= option.

Prefix Option
TEXTATTRS=style-element | style-element (text-options) | (text-options)
specifies the color and font properties of the entire text string or individual text-
items.

Default The GraphAnnoText style element.

1470 Chapter 14 / Draw Statements

Interaction When multiple TEXTATTRS= prefix options are used, each one
cancels the last, resetting all text properties to the default set by
the GraphAnnoText style element. Subsequent text-items to the
right are then assigned the text properties specified in the closest
TEXTATTRS= setting to their left. Thus, to vary the text properties
across text-items, you do not have to override settings from a
previous TEXTATTRS= setting. Each TEXTATTRS= specification
resets all text properties to the default so that only the new settings
are applied to subsequent text-items.

Tip To ensure that all text has the same text properties, use this prefix
option once only and place it before the first text-item.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Text Options” on page 1606 for available text-options.

Text Commands
{ SUB "string" | dynamic }
text-command that specifies that the string or dynamic is to appear as subscript
text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example drawtext "y = " b{sub "0"} " + b" {sub "1"} "x";

{ SUP "string" | dynamic }

text-command that specifies that the string or dynamic is to appear as
superscript text.

See “Rules for Unicode and Special Character Specifications” on page 1342

Example drawtext "R" {sup "2"} " = " {format (6.4) RSQUARED} ;

{ UNICODE "hex-string"x | keyword | dynamic }

text-command that specifies a glyph (character) to be displayed using its
Unicode specification or keyword equivalent.

"hex-string"x
a four-byte hexadecimal constant that represents a UNICODE character in
the current font. For a complete listing, see http://unicode.org/charts/
charindex.html.

keyword
a SAS keyword for a UNICODE character.

For a listing of SAS supplied keywords, see Appendix 2, “Reserved Keywords

and Unicode Values,” on page 1595.

dynamic
a dynamic variable that resolves to either "hex-string"x or a keyword for a
UNICODE character.
 DRAWTEXT Statement 1471

Note The UNICODE text command attempts to access a UNICODE value in

the current font. Not all fonts support accessing characters through
their UNICODE value. Some fonts support only a limited set of
UNICODE values. If the UNICODE value is not accessible, then the
command might be ignored or a nonprintable character might be
substituted.

See “Using UNICODE Text” on page 1342

“Rules for Unicode and Special Character Specifications” on page 1342

Example The following statements show how to use the {UNICODE} text
command:
drawtext {unicode alpha} "=" CONF;

drawtext {unicode "03B1"x} "=" CONF;

Details
A DRAWTEXT statement draws a text box that contains one or more lines of text.
The text can be formatted, using the TEXTATTRS= prefix option. By default, the
text box is drawn in the center of the graph. You can change the default position
with the options ANCHOR= , X= , Y= , XSPACE= , and YSPACE= .

For general information about the types of elements that can be drawn with the
draw statements, the drawing space and drawing units that they use, and how the
drawn elements are anchored, see Chapter 13, “Key Concepts for Using Draw
Statements,” on page 1395 . For detailed usage information, consult the SAS Graph
Template Language: User’s Guide.

Example: DRAWTEXT Statement

The following graph was generated by the “Example Program” on page 1472. The
first DRAWTEXT statement shows how to draw multiple lines of text in a specific
position within the graph. The second DRAWTEXT statement shows how to create
a watermark, which is achieved by applying transparency to text that is rotated
within the graph.
 1472 Chapter 14 / Draw Statements

Example Program
proc template;
define statgraph modelfit;
begingraph;
entrytitle "Regression Fit Plot";
layout lattice;
layout overlay / xaxisopts=(offsetmin=0.1);
drawtext textattrs=(style=italic size=8pt)
"Band shows 99% Confidence Limit of Mean" /
anchor=bottomleft width=15 widthunit=percent
xspace=wallpercent yspace=wallpercent
x=0 y=10 justify=center ;
modelband "myclm";
scatterplot x=height y=weight / primary=true;
regressionplot x=height y=weight / alpha=0.01 clm="myclm";
endlayout;
endlayout;
drawtext textattrs=(color=gray size=52pt) "CONFIDENTIAL" /
transparency=0.75 rotate=-35
width=110 widthunit=percent justify=center ;
endgraph;
end;

proc sgrender data=sashelp.class template=modelfit;

run;
 1473

PART 10

GTL Annotation Facility

Chapter 15
About the GTL Annotation Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475

Chapter 16
The ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483
1474
 1475

15
About the GTL Annotation
Facility

Data-Set-Driven Annotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475

Annotation Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1476
About the Annotation Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1476
Annotation Data Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1477
About the Coordinate Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1478
About the Drawing Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1479
Using the SGANNO Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1480
See Also . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481
ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481
SGRENDER Statement SGANNO Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481

Data-Set-Driven Annotations
GTL supports data-set-driven annotations, which enable you to add the following
graphics elements to your graphs:
n text

n lines and arrows

n circles and ovals

n squares and rectangles

n polygons and polylines

n images

Unlike graphics elements that are drawn using GTL draw statements in a GTL
template, data-set-driven annotations are drawn from graphics instructions that
are stored in a SAS data set. The GTL annotation facility is similar to the
SAS/GRAPH Annotate facility. The GTL annotation facility enables you to separate
your annotation instructions from your template statements. To change your
 1476 Chapter 15 / About the GTL Annotation Facility

annotations, you can specify a different annotation data set or modify the
instructions in the original data set. You do not have to modify your template code.

The following items are required in order to use the GTL annotation facility:
n a SAS data set that contains the annotation instructions. The annotation
instructions must be stored in a SAS data set. Annotation instructions stored in
a Cloud Analytic Services (CAS) in-memory table are not supported.
n at least one ANNOTATE statement in your GTL template.

n an SGRENDER statement that includes the SGANNO=data-set-name option,

which specifies the name of the data set that contains the annotation
instructions.

This section describes the requirements for using the GTL annotation facility. For
examples of how to use the facility to annotate your graphs, see “Adding Data-
Driven Annotations to Your Graph” in SAS Graph Template Language: User’s Guide.

Annotation Data Set

About the Annotation Data Set

The annotation data set contains the instructions for drawing annotations. Each
observation in the data set contains columns for an annotation instruction. The
following information is typically stored in these columns:
n the annotation function name

n the coordinates of the annotation location

n the drawing space

n the attributes of the annotation, such as color, font, and so on

The information that is required for each instruction is function-dependent. See

“Annotation Data Requirements” on page 1477.

To create the annotation data set, use the same methods that you would use to
create any SAS data set. You must store the annotation instructions in a SAS data
set. Annotation instructions stored in a CAS in-memory table are not supported.
The most common method for creating a data set is to use a DATA step. For more
information about the DATA step, see SAS Programmer’s Guide: Essentials. Macros
are available that you can run in your DATA step to create the observations for your
annotations. See “Using the SGANNO Macros” on page 1480.
 Annotation Data Set 1477

Annotation Data Requirements

Each observation in the annotation data set must provide sufficient information to
complete an annotation instruction. To provide this information, certain columns
must be included in each observation. All instructions require a Function column,
which is a character column that stores the name of the function that is to be
performed. An annotation function draws an annotation, such as a line, arrow, oval,
and so on. The Function column can specify one of the following function names:

ARROW POLYGON TEXT

IMAGE POLYLINE TEXTCONT
LINE POLYCONT
OVAL RECTANGLE

For more information about these functions, see “SG Annotation Function
Dictionary” in SAS ODS Graphics: Procedures Guide.

Each function requires one or more additional columns that provide required
information such as the location coordinates, dimensions, and so on. The following
table lists the additional columns that are required for each function. At a minimum,
you must include the Function column and the columns listed in the following table
in each observation in your annotation data set.

Additional Columns
FUNCTION Required1 Description

ARROW X1 or XC1 These columns store the starting

coordinates (numeric or character).
Y1 or YC1

X2 or XC2 These columns store the ending

coordinates (numeric or character).
Y2 or YC2

IMAGE Image This column stores the path to the image

file (character).

LINE X1 or XC1 These columns store the starting

coordinates (numeric or character).
Y1 or YC1

X2 or XC2 These columns store the ending

coordinates (numeric or character).
Y2 or YC2

OVAL Height These columns store the oval dimensions

(numeric).
Width

X1 or XC1 These columns store the oval anchor

coordinates (numeric or character).
 1478 Chapter 15 / About the GTL Annotation Facility

Additional Columns
FUNCTION Required1 Description

Y1 or YC1

POLYGON X1 or XC1 These columns store the polygon starting

coordinates (numeric or character).
Y1 or YC1

POLYLINE X1 or XC1 These columns store the polyline starting

coordinates (numeric or character).
Y1 or YC1

POLYCONT X1 or XC1 These columns store the ending

coordinates of a polygon or polyline
Y1 or YC1
segment (numeric or character).

RECTANGLE Height These columns store the rectangle

dimensions (numeric).
Width

X1 or XC1 These columns store the rectangle

anchor coordinates (numeric).
Y1 or YC1

TEXT and Label This column stores the annotation text

TEXTCONT (character).

1 See “About the Coordinate Columns” on page 1478.

For the remaining information such as LINECOLOR, TEXTFONT, and so on, default
values are used. To change the default values, you can add the necessary columns
to your instruction observations. For information about other columns that you can
add for each function, see “SG Annotation Function Dictionary” in SAS ODS
Graphics: Procedures Guide

Note: For polylines and polygons, the initial POLYLINE or POLYGON observation
and all of the subsequent POLYCONT observations for the annotation must be
contiguous. For text that is continued, the initial TEXT observation and all of the
subsequent TEXTCONT observations for the annotation must also be contiguous.

About the Coordinate Columns

The annotation coordinates specify the location of the annotation as X and Y
values. In some cases, such as for a line or arrow, two sets of coordinates are
required in order to specify the beginning and ending locations of the annotation.
You can specify numeric or character columns for each coordinate value. For
numeric coordinate values, the location or starting location coordinates are stored
in the X1 and Y1 columns in the annotation data set. When ending coordinates are
 Annotation Data Set 1479

required, the numeric coordinates are stored in the X2 and Y2 columns. For
character values, the coordinates are stored in the XC1 and YC1 columns, and when
required, the XC2 and YC2 columns.

Note: If both the numeric and character columns are specified for a coordinate
value, then the numeric column takes precedence and the character column is
ignored.

About the Drawing Space

When you specify X and Y coordinates for an annotation, the coordinate values
refer to a drawing space. By default, this space is a percentage of the graph area.
Specifying X=50 and Y=50, for example, refers to a point in the center of the graph
area. Here is a complete list of the drawing spaces that you can use.

DATAPERCENT GRAPHPERCENT LAYOUTPIXEL

DATAPIXEL GRAPHPIXEL WALLPERCENT
DATAVALUE LAYOUTPERCENT WALLPIXEL

For more information about these drawing spaces, see “About the Drawing Space
and Drawing Units” on page 1398.

To change the default drawing space for your annotations, you can add the columns
shown in the following table to your annotation observations.

Column Name Description

DrawSpace Specifies the drawing space for all coordinates. This value cannot
be used with the POLYCONT function.

X1Space Specifies the drawing space for the X1 coordinate. This value cannot
be used with the TEXTCONT function.

X2Space Specifies the drawing space for the ending X2 coordinate. This
value can be used with the ARROW and LINE functions only.

Y1Space Specifies the drawing space for the Y1 coordinate. This value cannot
be used with the TEXTCONT function.

Y2Space Specifies the drawing space for the ending Y2 coordinate. This value
can be used with the ARROW and LINE functions only.

For information about these columns, see “SG Annotation Function Dictionary” in
SAS ODS Graphics: Procedures Guide
 1480 Chapter 15 / About the GTL Annotation Facility

Using the SGANNO Macros

Table 15.1 SG Annotation Macros

Macro Name1 Description

%SGANNO Compiles the available macros and makes them

available for you to use.
Note: You must run the %SGANNO macro in your
SAS session before you can use any of the other
annotation macros that are listed in this table.

%SGANNO_HELP Displays help information for a specified annotation

macro.

%SGARROW Creates an observation that draws an arrow.

%SGIMAGE Creates an observation that displays an image.

%SGLINE Creates an observation that draws a line.

%SGOVAL Creates an observation that draws an oval.

%SGPOLYCONT Creates an observation that draws a polygon or

polyline segment.

%SGPOLYGON Creates an observation that specifies the starting

point of a polygon.
Note: Use the %SGPOLYCONT macro to draw the
segments.

%SGPOLYLINE Creates an observation that specifies the starting

point of a polyline.
Note: Use the %SGPOLYCONT macro to draw the
segments.

%SGRECTANGLE Creates an observation that draws a rectangle.

%SGTEXT Creates an observation that draws a single line of

text or the first line of text in a multiline annotation.
Note: For multiple lines of text, use the
%SGTEXTCONT macro for the subsequent lines.
 SGRENDER Statement SGANNO Option 1481

Macro Name1 Description

%SGTEXTCONT Creates an observation that draws a continuing line

of text in a multiline annotation.

1 The macros listed in this table are valid starting with SAS 9.4M1. You can run these macros in a
DATA step to simplify the process of creating your SG annotation data sets.

For more information about these macros, see SAS ODS Graphics: Procedures
Guide.

See Also
“Adding Observations for Polylines and Polygons” in SAS Graph Template Language:
User’s Guide

ANNOTATE Statement
To render the annotations in your annotation data set, you must include at least
one ANNOTATE statement in the GTL template for your graph. You can place the
ANNOTATE statement anywhere in the template to render your annotations. The
annotations are drawn using the context in which the ANNOTATE statement is
encountered. You can use the ANNOTATE statement with no options to render all
of the annotations in your annotation data set. You can use the ID= option in the
ANNOTATE statement to render only a subset of the annotations.

For more information about the ANNOTATE statement, see “ANNOTATE

Statement” on page 1483.

SGRENDER Statement SGANNO Option

The SGRENDER statement SGANNO= option specifies the name of a SAS data set
that contains annotation instructions. To render your graph with annotations, you
must use the SGANNO= option in the SGRENDER statement to specify the name of
your annotation data set. When the ANNOTATE statement is encountered in the
template, the annotation instructions are read from the annotation data set and are
rendered in the current context in the graph. If the template does not include the
ANNOTATE statement, the annotations in the annotation data set are ignored. See
“SGANNO=annotation-data-set” in SAS ODS Graphics: Procedures Guide.
1482 Chapter 15 / About the GTL Annotation Facility
 1483

16
The ANNOTATE Statement

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483
ANNOTATE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483

Dictionary

ANNOTATE Statement
Draws annotations from annotation instructions that are stored in a SAS data set.

Requirements: The annotation instructions must be stored in a SAS data set.

The annotation data set must be specified by the SGANNO= option in the
SGRENDER statement.
Tips: When the ANNOTATE statement is placed inside a LAYOUT container, the layout
container is used as the context when DRAWSPACE for an annotation is specified
as LAYOUT, WALL, or DATA.
When the DRAWSPACE for an annotation object is GRAPH, the context for drawing
the object is the graph area regardless of where the ANNOTATE statement is
located.
The LAYER option determines whether the annotations are drawn in front of the
graph elements or behind them.
See: “Annotation Data Set” on page 1476
“SG Annotation Function Dictionary” in SAS ODS Graphics: Procedures Guide
SGRENDER statement SGANNO= option
“Adding Data-Driven Annotations to Your Graph” in SAS Graph Template Language:
User’s Guide
1484 Chapter 16 / The ANNOTATE Statement

Syntax
ANNOTATE </ID="annotation-ID">

Optional Argument
ID="annotation-identifier"
specifies the ID column value of the annotations that are to be drawn. The ID
column of the annotation data set contains a unique character value that
identifies the subset to which each annotation belongs. All annotations in the
annotation data set with an ID column value that matches the specified
annotation identifier are drawn. If the annotation data set does not contain an
ID column or if no ID column value matches the specified identifier, then no
annotations are drawn.

Default All of the annotations in the annotation data set are drawn.

Requirement The annotation data set must contain an ID column, and at least
one ID column value must match the specified identifier value.

Tips The ID= option can be used with the annotation data set ID column
to subset the annotations that are stored in an annotation data set.

The identifier value matching is case sensitive.

See “Subsetting Annotations” in SAS Graph Template Language: User’s

Guide

Details
You can use the ANNOTATE statement anywhere in a GTL template to render
annotation objects in the current context. The annotation objects are read from an
annotation data set that is specified by the SGANNO= option in the SGRENDER
statement. You must use at least one ANNOTATE statement in the template to
draw the annotations. By default, all of the annotations in the annotation data set
are drawn. You can use the ID column in the annotation data set with the ID= option
in the ANNOTATE statement to draw only a subset of the annotations in the data
set, if you want to. See “Subsetting Annotations” in SAS Graph Template Language:
User’s Guide.

Examples

Example 1: Adding Simple Text Annotations

Here is an example that uses the GTL annotation facility to place custom Y-axis
labels inside the plot wall for each of two side-by-side plots. Each label is rotated
90 degrees and is placed on the inside edge of the plot wall near the plot’s Y axis.
The default Y-axis labels are suppressed.
 ANNOTATE Statement 1485

Here is the output for this example.

Example Program
/* Create the annotation data set */
data anno;
length label $30 id $5 anchor $12;
drawspace="wallpercent";

/* Create the bar chart Y-axis label. */

id='BAR'; function='text'; x1=0;
textweight='bold'; anchor='top'; rotate=90;
width=1000; widthunit="pixel";
label="MPG (City)";
output;

/* Create the histogram Y-axis label. */

id='HIST'; anchor='bottom'; x1=100;
label="Distribution of MPG (Percent)";
output;
run;
/* Define the template */
proc template;
define statgraph anno;
begingraph;
entrytitle "Vehicle Statistics";
layout lattice / columns=2 columngutter=10;
layout overlay /
xaxisopts=(offsetmin=0.2 offsetmax=0.2)
yaxisopts=(display=(ticks tickvalues));

/* Draw the barchart of origin and MPG city */

barchart x=origin y=mpg_city / name="bar" stat=mean
1486 Chapter 16 / The ANNOTATE Statement

group=type groupdisplay=cluster clusterwidth=0.7

dataskin=sheen;
discretelegend "bar";
annotate / id="BAR"; /* Draw the barchart label. */
endlayout;
layout overlay /
y2axisopts=(display=(ticks tickvalues));

/* Draw the histogram of MPG city */

histogram mpg_city / dataskin=sheen yaxis=y2;
densityplot mpg_city / yaxis=y2;
annotate / id="HIST"; /* Draw the histogram label. */
endlayout;
endlayout;
endgraph;
end;
run;
/* Render the graph with the annotation */
proc sgrender data=sashelp.cars template=anno sganno=anno;
where type in ("Sedan" "Sports" "SUV");
run;

Program Description
Create the annotation data set Anno. The annotation data set contains two
observations, one for each label. The DRAWSPACE is set to WALLPERCENT for
both labels. The ID column specifies an identifier for each label. The bar chart label
is identified as BAR. It is rotated 90 degrees, anchored on TOP, and is placed at 0%
of the wall space along the X axis. The default value 50% percent is used along the
Y axis to center the label vertically. This places the label on inside left edge of the
bar chart wall. This histogram label is identified as HIST. It is rotated 90 degrees,
anchored on BOTTOM, and is placed at 100% of the wall space along the X axis.
Like the bar chart label, the default 50% percent is used along the Y axis. This
places the label on the right inside edge of the histogram plot wall.
/* Create the annotation data set */
data anno;
length label $30 id $5 anchor $12;
drawspace="wallpercent";

/* Create the bar chart Y-axis label. */

id='BAR'; function='text'; x1=0;
textweight='bold'; anchor='top'; rotate=90;
width=1000; widthunit="pixel";
label="MPG (City)";
output;

/* Create the histogram Y-axis label. */

id='HIST'; anchor='bottom'; x1=100;
label="Distribution of MPG (Percent)";
output;
run;
 ANNOTATE Statement 1487

Define the graph template. The template defines a two-column, one-row lattice for
the two plots. For both plots, the DISPLAY= Y-axis option suppresses the default
axis label. For the bar chart, the OFFSETMIN= and OFFSETMAX= X-axis options
reserve space for the Y-axis label. The ANNOTATE statement draws the axis label.
The ID="BAR" option in the ANNOTATE statement draws the bar chart Y-axis label
in the context of the bar chart’s OVERLAY layout. For the histogram, the
ANNOTATE statement draws the histogram Y-axis label (ID="HIST") in the context
of the histogram’s OVERLAY layout.
/* Define the template */
proc template;
define statgraph anno;
begingraph;
entrytitle "Vehicle Statistics";
layout lattice / columns=2 columngutter=10;
layout overlay /
xaxisopts=(offsetmin=0.2 offsetmax=0.2)
yaxisopts=(display=(ticks tickvalues));

/* Draw the barchart of origin and MPG city */

barchart x=origin y=mpg_city / name="bar" stat=mean
group=type groupdisplay=cluster clusterwidth=0.7
dataskin=sheen;
discretelegend "bar";
annotate / id="BAR"; /* Draw the barchart label. */
endlayout;
layout overlay /
y2axisopts=(display=(ticks tickvalues));

/* Draw the histogram of MPG city */

histogram mpg_city / dataskin=sheen yaxis=y2;
densityplot mpg_city / yaxis=y2;
annotate / id="HIST"; /* Draw the histogram label. */
endlayout;
endlayout;
endgraph;
end;
run;

Render the graph with the annotations. The SGANNO=ANNO option is included in
the SGRENDER statement to specify the name of the annotation data set. When
the graph is rendered, the ANNOTATE statements in the graph template cause the
annotations to be drawn.
/* Render the graph with the annotation */
proc sgrender data=sashelp.cars template=anno sganno=anno;
where type in ("Sedan" "Sports" "SUV");
run;

Example 2: Using the SG Annotation Macros to Generate Your

Annotation Data
In “Example 1: Adding Simple Text Annotations” on page 1484, standard SAS DATA
step statements are used to create the observations for the text annotations. This
example shows you how to use the SG annotation macros instead to create the
same data set.
 1488 Chapter 16 / The ANNOTATE Statement

Example Program
/* Create the annotation data set */
data anno;
/* Compile the annotation macros */
%sganno;
/* Create the bar chart Y-axis label. */
%sgtext(
reset=all,
id=BAR,
drawspace=wallpercent,
x1=0,
label="MPG (City)",
textweight=bold,
anchor=top,
rotate=90,
width=1000,
widthunit=pixel);
/* Create the histogram Y-axis label. */
%sgtext(
id=HIST,
x1=100,
label="Distribution of MPG (Percent)",
anchor=bottom);
run;

Program Description
Here is the SAS code.

Open the DATA step and run the %SGANNO macro. After you open the DATA step,
you must run the %SGANNO macro to compile the SG annotation macros and make
them available for your use. You need to compile the macros only once during a
SAS session.
/* Create the annotation data set */
data anno;
/* Compile the annotation macros */
%sganno;

Create the observation for the bar chart Y-axis label. Use the %SGTEXT macro to
generate the observation for the bar chart Y-axis label.
/* Create the bar chart Y-axis label. */
%sgtext(
reset=all,
id=BAR,
drawspace=wallpercent,
x1=0,
label="MPG (City)",
textweight=bold,
anchor=top,
rotate=90,
width=1000,
 ANNOTATE Statement 1489

widthunit=pixel);

Create the histogram Y-axis label, and close the DATA step. Specify the new
values for the ID, X1, Label, and Anchor columns. To carry over the remaining
column values from the previous macro call, do not specify RESET=ALL.
/* Create the histogram Y-axis label. */
%sgtext(
id=HIST,
x1=100,
label="Distribution of MPG (Percent)",
anchor=bottom);
run;

See Also
n Chapter 15, “About the GTL Annotation Facility,” on page 1475
n “Adding Data-Driven Annotations to Your Graph” in SAS Graph Template
Language: User’s Guide
1490 Chapter 16 / The ANNOTATE Statement
 1491

PART 11

Attribute Maps

Chapter 17
Key Concepts for Using Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1493

Chapter 18
Discrete Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1505

Chapter 19
Range Attribute Map Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1521
1492
 1493

17
Key Concepts for Using
Attribute Maps

About Attribute Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1493

Defining a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1494
How to Define a Discrete Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1494
Specifying the Attribute Mapping Information in a DISCRETEATTRMAP Block . 1494
Specifying the Attribute Mapping Information in a SAS Data Set . . . . . . . . . . . . . . . . . 1495
Defining a Range Attribute Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1498
Referencing an Attribute Map in Your Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1499
Referencing a Discrete Attribute Map in Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . 1499
Referencing a Range Attribute Map in Plot Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1503

About Attribute Maps

By default, many of the graphical attributes of a plot vary with the plot data. For
example, when plots display grouped values, by default, the graphical attributes for
each group value are selected from the GraphData1–GraphDataN style elements in
data order. Changes in the data order can significantly change the appearance of
the plot. When plots display a color gradient, by default, the colors assigned to the
classification variable are derived from a color ramp based on the actual range of
the data. The color assigned to each value can vary with the range of the
classification values. Attribute maps enable you to assign the same graphical
properties to specific values or ranges of values regardless of data order or the data
range. They are useful when you want your graphs to have consistent visual
properties when the data varies.

GTL supports two types of attribute maps: discrete attribute maps and range
attribute maps. A discrete attribute map maps discrete values to graphical
properties. A range attribute map maps numeric values or ranges of numeric values
to graphical properties.
 1494 Chapter 17 / Key Concepts for Using Attribute Maps

Defining a Discrete Attribute Map

How to Define a Discrete Attribute Map

You define a discrete attribute map in the following way:
n Specify the attribute mapping information in a SAS data set or in a
DISCRETEATTRMAP block in the template code. See “Specifying the Attribute
Mapping Information in a DISCRETEATTRMAP Block” on page 1494 and
“Specifying the Attribute Mapping Information in a SAS Data Set” on page 1495.

Note: Defining a discrete attribute map in a SAS data set is valid in SAS 9.4M1
and later releases.

n Reference the attribute map in plot statements by using one of the following
methods:
o Include a DISCRETEATTRVAR statement in the template code to create an
attribute map variable. The variable associates the attribute map with a
classification column in the plot data. Then use the attribute map variable to
reference the discrete attribute map in plot statements. This method applies
to an attribute map that is defined in a SAS data set or in a
DISCRETEATTRMAP block in the template code.
o Specify the classification column, which is in the plot data, in the plot
statements where needed. When the graph is rendered, the SGRENDER
DATTRVAR statement is used to associate the attribute map that is defined
in a SAS data set with the classification column in the plot data. This method
applies only to an attribute map that is defined in a SAS data set.

Note: The DATTRVAR statement is valid in SAS 9.4M1 and later releases.
For information about the DATTRVAR statement, see SAS ODS Graphics:
Procedures Guide.

Specifying the Attribute Mapping Information in a

DISCRETEATTRMAP Block
Use a DISCRETEATTRMAP block to specify the attribute mapping information in
the template code. Place the block in the global definition area of your template
between the BEGINGRAPH statement and the first layout statement. The block
 Defining a Discrete Attribute Map 1495

contains one or more VALUE statements that specify the attribute mapping
information. Each VALUE statement specifies a single discrete value and one or
more graphical properties that are assigned to that value. The NAME= option in the
DISCRETEATTRMAP statement specifies the name of the attribute map. The
DISCRETEATTRVAR statement creates the reference variable for the attribute
map. The following options are in the DISCRETEATTRVAR statement:
n The ATTRVAR= option specifies a unique name for the attribute-map-to-data-
set-column association.
n The ATTRMAP= option specifies the value of the NAME= option that is
specified in the DISCRETEATTRMAP statement.
n The VAR= option specifies the name of the numeric or character column in the
plot data set, an expression, or the name of a dynamic variable.

Use the attribute map variable to reference the attribute map in plot statements.

Note: Do not use the attribute variable in an expression. Doing so might produce
unexpected results.

Note: The values and graphical attributes that are defined in a discrete attribute
map cannot be displayed by a CONTINUOUSLEGEND statement.

For more information, see “Defining a Discrete Attribute Map in a

DISCRETEATTRMAP Block” on page 1510.

Specifying the Attribute Mapping Information in a

SAS Data Set
Starting with SAS 9.4M1, you can specify the attribute mapping information for a
discrete attribute map in a SAS data set instead of in a DISCRETEATTRMAP block
in the template. You must store the attribute mapping information in a SAS data
set. Attribute mapping information stored in a Cloud Analytic Services (CAS) in-
memory table is not supported. Each observation in the data set maps a discrete
value to one or more specific graphical properties. The following table lists the
columns that you must include in your data set for each observation.

Table 17.1 Columns That Are Required in Each Observation

Column Name Type Description

ID Character Specifies the name of the attribute map to which

this observation is associated. The ID value must
be a SAS name token and not a literal. A data set
can define multiple attribute maps. All
observations that are associated with a specific
attribute map must have the same ID value. When
1496 Chapter 17 / Key Concepts for Using Attribute Maps

Column Name Type Description

you are creating an attribute map variable for this

attribute map, specify the name that you specified
in the ID column in the DISCRETEATTRVAR
statement ATTRMAP= argument. The ID column
can also be specified in an SGRENDER
DATTRVAR statement when the graph is
rendered. The ID value is case sensitive.

Value Character Specifies the discrete classification value that is

to be mapped to the graphical properties that are
specified in this observation. The value is case
sensitive.

The following table lists the columns that you can use to specify the graphical
properties for each classification value. Properties that you do not specify default
to the properties that are normally used when an attribute map is not specified.

Table 17.2 Optional Columns That Specify Graphical Properties

Column Name Type Description

FillColor Character Specifies the color for the filled areas for this
classification value. When FillStyleElement is also
specified, this column overrides the specified
style element’s Color attribute.

FillStyleElement Character Specifies the name of a style element, such as

GraphData3, that is to provide the fill attributes.

FillTransparency Numeric Specifies the transparency of the fill color for this
classification value. Values are in the range 0
(opaque) to 1 (completely transparent).

LineColor Character Specifies the line color for this classification

value. When LineStyleElement is also specified,
this column overrides the specified style
element’s ContrastColor attribute.

LinePattern Character Specifies the line pattern for this classification

or value. You can specify the pattern name or
Numeric number. See “Available Line Patterns” on page
1607.

LineStyleElement Character Specifies the name of a style element, such as

GraphData3, that is to provide the line attributes.

LineThickness Numeric Specifies the line thickness, in pixels, for this

classification value. The values must be integers.
 Defining a Discrete Attribute Map 1497

Column Name Type Description

MarkerColor Character Specifies the color of the marker symbol for this
classification value. When MarkerStyleElement is
also specified, this column overrides the specified
style element’s ContrastColor attribute.

MarkerSize Numeric Specifies the size of the marker, in pixels, for this
classification value. The values must be integers.

MarkerSyleElemen Character Specifies the name of a style element, such as

t GraphData3, that is to provide the marker
attributes.

MarkerSymbol Character Specifies the marker symbol to use for this

classification value. See SYMBOL= in “Marker
Options” on page 1604. When
MarkerStyleElement is also specified, this column
overrides the specified style element’s
MarkerSymbol attribute.

MarkerTransparenc Numeric Specifies the transparency of the marker for this

y classification value. Values are in the range 0
(opaque) to 1 (completely transparent).

NoCase Character Specifies whether the attribute-map value

comparisons are case-insensitive. Valid values are
TRUE and FALSE. When TRUE, value comparisons
are case insensitive. The default is TRUE.
Note: This column is valid starting with SAS
9.4M3.

TextColor Character Specifies the text color for this classification

value.

TextFont Character Specifies the text font by name for this

classification value.

TextSize Numeric Specifies the text font size in points for this
classification value. The values must be integers.

TextStyle Character Specifies the text style for this classification

value. Valid values are NORMAL and ITALIC.

TextStyleElement Character Specifies the name of a style element, such as

GraphLabelText, that is to provide the text
attributes.
 1498 Chapter 17 / Key Concepts for Using Attribute Maps

Column Name Type Description

TextWeight Character Specifies the text weight for this classification

value. Valid values are NORMAL and BOLD.

Show Character Specifies whether the legend displays only the

attribute map values that appear in the data or
always displays all of the values in the attribute
map. Valid values are DATA and ATTRMAP. The
default is DATA.
Note: This column is valid starting with SAS
9.4M3.

The ID column provides the name of the attribute map, which is specified in the
ATTRMAP= option in the DISCRETEATTRVAR statement or in a DATTRVAR
statement. When the graph is rendered, the name of the attribute map data set
must be specified in the DATTRMAP= option in the SGRENDER statement. See
“DATTRVAR Statement” in SAS ODS Graphics: Procedures Guide. For information
about how to reference the attribute map in your plot statements, see “Referencing
a Discrete Attribute Map in Plot Statements” on page 1499.

Defining a Range Attribute Map

A range attribute map is defined as follows:
n The attribute mapping information is specified in a RANGEATTRMAP block in
the template code. It must be placed in the global definition area of your
template between the BEGINGRAPH statement and the first layout statement.
The NAME= option in the RANGEATTRMAP statement specifies the name of
the attribute map. The block contains one or more RANGE statements that
specify the attribute mapping information. Each RANGE statement specifies a
numeric value or numeric value range, and one or more graphical properties that
are assigned to that value or range. An ENDRANGEATTRMAP statement is used
to close the block.

Note: Range attribute maps do not support attribute map information that is
stored in a SAS data set.

n The RANGEATTRVAR statement is used to create an attribute map variable

that associates the attribute map with a column in the data that provides the
classification values. In the RANGEATTRVAR statement:
o The ATTRVAR= option specifies a unique name for range attribute map to
data set column association.
o The ATTRMAP= option specifies the value of the NAME= option that is
specified in the RANGEATTRMAP statement.
 Referencing an Attribute Map in Your Plot Statements 1499

o The VAR= option specifies the name of the numeric column in the plot data
set with which the range attribute map is to be associated.

Note: A RANGEATTRMAP can be used with a numeric column only.

n The attribute map variable is used to reference the discrete attribute map in plot
statements.

Note: The values and graphical attributes defined in a range attribute map
cannot be displayed by a DISCRETELEGEND statement.

For information about how to reference a range attribute map, see “Referencing
a Discrete Attribute Map in Plot Statements” on page 1499.

The RANGE statements in the RANGEATTRMAP block can associate a range of

values or a single value with a single color or a color ramp. The syntax of the
RANGE statement is as follows:

RANGE low-value< < > – < < >high-value / options

The optional exclusion operator (<) can be placed after the low value or before the
high value to exclude that value from the range endpoint. The low value and high
value can be an unformatted numeric value or a range keyword. For the low value,
keyword MIN, NEGMAX, or NEGMAXABS can be used instead of numeric value. For
the high value, keyword MAX or MAXABS can be used. For information about the
range keywords, see SAS Graph Template Language: Reference.

Note: If two ranges share a common endpoint, such as 10–20 and 20–30, and no
exclusion operator ( < ) is used, then the common endpoint belongs to the lower
range, which is 10–20 in this case.

For more information about how to create a range attribute map, see “Creating and
Using a Range Attribute Map” on page 1527.

Referencing an Attribute Map in Your

Plot Statements

Referencing a Discrete Attribute Map in Plot

Statements
If you use a DISCRETEATTRVAR statement to create a variable for an attribute
map, then reference the attribute map in a plot statement by specifying the name
1500 Chapter 17 / Key Concepts for Using Attribute Maps

that is assigned in the DISCRETEATTRVAR statement’s ATTRVAR= argument. Each

plot statement’s documentation in this reference indicates which options support a
reference to a discrete attribute map variable as the specified value.

Note: A reference to a discrete attribute variable must be a direct reference to the

attribute variable. It cannot be set by a dynamic variable.

When you use a DATTRVAR statement in the SGRENDER statement, the attribute
map is referenced implicitly when a column that is assigned to the attribute map ID
in the DATTRVAR statement is referenced in a plot statement. The following table
lists the statement options that support a reference to a discrete attribute map.

Table 17.3 Statement Options That Support a Discrete Attribute Map

Statement Options

AXISTABLE COLORGROUP= on page 222

TEXTGROUP= on page 231

BANDPLOT GROUP= on page 248

BARCHART GROUP= on page 276

BARCHARPARM GROUP= on page 323

BOXPLOT GROUP= on page 380

BOXPLOTPARM GROUP= on page 421

BUBBLEPLOT GROUP= on page 452

ELLIPSEPARM GROUP= on page 519

FRINGEPLOT GROUP= on page 528

HEATMAPPARM COLORGROUP= on page 550

HIGHLOWPLOT GROUP= on page 573

LINECHART GROUP= on page 573

LINEPARM GROUP= on page 654

LOESSPLOT GROUP= on page 666

MOSAICPLOTPARM COLORGROUP= on page 687

NEEDLEPLOT GROUP= on page 708

 Referencing an Attribute Map in Your Plot Statements 1501

Statement Options

PBSPLINEPLOT GROUP= on page 724

PIECHART CATEGORY= on page 734

POLYGONPLOT GROUP= on page 757

REGRESSIONPLOT GROUP= on page 800

SCATTERPLOT GROUP= on page 824

SCATTERPLOTMATRIX GROUP= on page 864

SERIESPLOT GROUP= on page 900

STEPPLOT GROUP= on page 944

VECTORPLOT GROUP= on page 1007

WATERFALLCHART COLORGROUP= on page 1022

In a DISCRETELEGEND statement, reference the plot statement that uses the

attribute map. The plot statement must have a NAME= option that assigns a name
to the plot, because the DISCRETELEGEND statement references that name.
Because the attribute map is referenced in the plot statement, the legend uses the
attribute map to represent the group values that exist in the data.

TIP If the discrete attribute map is defined in a DISCRETEATTRMAP block,

then you can use the DISCRETELEGENDENTRYPOLICY=ATTRMAP option
in the DISCRETEATTRMAP statement to display all of the items that are
defined in the attribute map regardless of whether the values appear in the
data. See DISCRETELEGENDENTRYPOLICY= on page 1506.

For an example of how to reference an attribute map variable that is created with
the DISCRETEATTRVAR statement, see “Example: DISCRETEATTRVAR Statement
with an Attribute Map Data Set” on page 1517.

The following code shows you how to reference an attribute map.

data attrds;
input ID $1-7 VALUE $8-14 FILLCOLOR $15-26 LINECOLOR $27-38;
datalines;
origin Asia DodgerBlue Black
origin Europe Chartreuse Black
origin USA SlateGray Black
;
run;

proc sort data=sashelp.cars out=cars;

1502 Chapter 17 / Key Concepts for Using Attribute Maps

by origin;
run;

proc summary data=cars;

class type;
ways 1;
by origin;
output out=counts(drop=_TYPE_ rename=(_FREQ_=Count));
run;

/* Define the template for this graph */

proc template;
define statgraph barchart;
begingraph / attrpriority=none;
entrytitle "Total Models Produced By Origin and Type";
layout overlay /
xaxisopts=(label="Vehicle Type")
yaxisopts=(label="Total Models Produced");
barchart category=type response=count / name="bar"
display=all barlabel=true
group=origin groupdisplay=cluster;
discretelegend "bar";
endlayout;
endgraph;
end;
run;

proc sgrender data=counts dattrmap=attrds template=barchart;

dattrvar origin="origin";
where origin ne "USA";
run;

SAS data set Attrds is created to define attribute map ORIGIN. For each value of
origin, attribute map ORIGIN specifies a fill color and a line color. The SUMMARY
procedure is used to create data set Counts, which contains the counts of vehicle
models by type and origin from table Sashelp.Cars. The TEMPLATE procedure
defines a bar chart of columns Type and Count grouped by column Origin in table
Counts. The SGRENDER procedure renders the graph.

In the SGRENDER procedure step, the DATTRMAP= option specifies the Attrds
attribute map data set. The DATTRVAR statement associates discrete attribute
map ORIGIN with classification column Origin in table Counts. A WHERE
statement eliminates cars that are manufactured in the USA. The WHERE
statement applies to the data in data set Counts only. It does not apply to the data
in attribute map data set Attrds.

TIP To subset data in your attribute map data set using a WHERE clause,
use a WHERE statement in a DATA step to create a new attribute map data
set or use the data set WHERE option in the SGRENDER procedure
DATTRMAP= option. Here is an example of the latter:
dattrmap=attrds(where=(value ne "USA"))

Here is the output for this example.

 Referencing an Attribute Map in Your Plot Statements 1503

Referencing a Range Attribute Map in Plot

Statements
In a plot statement, you reference a range attribute map by specifying the name
that is assigned in the RANGEATTRVAR statement’s ATTRVAR= argument. Each
plot statement’s documentation in this reference indicates which options support a
reference to a range attribute map variable as the specified value.

Note: A reference to a range attribute variable must be a direct reference to the

attribute variable. It cannot be set by a dynamic variable.

The following table lists the statement options that support a reference to a range
attribute map variable.

Table 17.4 Statement Options That Support a Range Attribute Map Variable
Reference

Statement Options

BUBBLEPLOT COLORRESPONSE=

HEATMAP COLORRESPONSE=

MOSAICPLOTPARM COLORRESPONSE=

POLYGONPLOT COLORRESPONSE=
1504 Chapter 17 / Key Concepts for Using Attribute Maps

Statement Options

SCATTERPLOT COLORRESPONSE= or MARKERCOLORGRADIENT=

SCATTERPLOTMATRIX COLORRESPONSE= or MARKERCOLORGRADIENT=

SURFACEPLOTPARM COLORRESPONSE or SURFACECOLORGRADIENT=

WATERFALLCHART COLORRESPONSE=

In a CONTINUOUSLEGEND statement, reference the plot statement that uses the

attribute map. The plot statement must have a NAME= option that assigns a name
to the plot, because the CONTINUOUSLEGEND statement references that name.
Because the attribute map is referenced in the plot statement, the legend uses that
attribute map to represent the numeric values that are present in the data.

For an example, see “Example 1: Using a Range Attribute Map to Control Marker
Colors” on page 1528.
 1505

18
Discrete Attribute Map
Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1505
DISCRETEATTRMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1505
DISCRETEATTRVAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1516

Dictionary

DISCRETEATTRMAP Statement
Defines a set of graphical properties that can be associated with user-defined sets of values.

Requirements: You must place the DISCRETEATTRMAP block directly inside the BEGINGRAPH
block between the BEGINGRAPH statement and the first LAYOUT statement. Do
not embed it in any other GTL statement block.
The DISCRETEATTRMAP block must contain at lease one VALUE statement.
The DISCRETEATTRVAR statement must be used to associate the discrete
attribute map with a data column.
Notes: The graphical properties for a discrete attribute map can also be defined in a SAS
data set. See “Defining a Discrete Attribute Map” on page 1494.
Prior to the third release for SAS 9.4, when a discrete attribute map is used for
group values in a plot that contributes to a discrete legend and attributes are
overridden in the plot statement, the attributes of some plot features and their
corresponding legend items might not match. Starting with SAS 9.4M3, the
attributes of the legend items always match the attributes of the corresponding
plot features.
 1506 Chapter 18 / Discrete Attribute Map Statements

See: “DISCRETEATTRVAR Statement” on page 1516

Syntax
DISCRETEATTRMAP NAME="string" </options>;
VALUE value-spec </options>;
<… more VALUE statements …>
ENDDISCRETEATTRMAP;

Required Argument
NAME="string "
assigns a name to the attribute definition. The name can be referenced in a
DISCRETEATTRVAR statement, which is used to associate the attribute map
with an input data column. The name can also be referenced in a
DISCRETELEGEND statement to map the specified graphical properties directly
to a discrete legend.

Restrictions The string is case sensitive, must be enclosed in quotation marks,

and must define a unique name within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

Optional Arguments
DISCRETELEGENDENTRYPOLICY=DATA | ATTRMAP
specifies whether the items that are contributed to a discrete legend by the plot
associated with this attribute map are only items that appear in the data or only
items that are defined in the attribute map.

Note: This option is valid starting in SAS 9.4M1.

DATA
the associated plot contributes to a discrete legend only items that appear in
the data.

ATTRMAP
the associated plot contributes to a discrete legend only items that are
defined in the discrete attribute map.

Interaction If this option is set to ATTRMAP, then data skins, overrides from
the DATATRANSPARENCY= option, and overrides from the
TRANSPARENCY= suboption in the FILLATTRS= and
MARKERATTRS= options are displayed in the discrete legend.
Overrides from other options such as the COLOR= suboption in
the FILLATTRS= option are not displayed.

Default DATA
 DISCRETEATTRMAP Statement 1507

IGNORECASE=TRUE | FALSE
specifies whether case is ignored when comparing the values that are specified
in the attribute map with values from an input data column.

Default FALSE. Value comparisons are case-sensitive.

Tip The effect of this option can be achieved by applying a function like
UPCASE to the data column and using only uppercase strings in each
VALUE statement.

See “boolean ” on page xii for other Boolean values that you can use.

TRIMLEADING=TRUE | FALSE
specifies whether leading blanks are trimmed (removed) from both the attribute
map values and the input data values before those values are compared. Trailing
blanks are always trimmed.

Default TRUE. Leading blanks are trimmed.

See “boolean ” on page xii for other Boolean values that you can use.

VALUE Statement Required Argument

value-spec
specifies one or more formatted strings or the keyword OTHER. Strings are
always quoted. Multiple strings must be separated by blanks, and each of the
strings must be enclosed in its own set of quotation marks. The formatted
strings must be equal to the formatted values of the classification column that
is used with the DISCRETEATTRVAR statement.

OTHER
creates a category for all other column values that are not explicitly
assigned with VALUE statements. This keyword is not quoted. The default
attributes for these values are derived from the GraphOther style element.

Note If OTHER is not specified, then data values that are not explicitly
assigned with VALUE statements are mapped to attributes as if a
discrete attribute map is not in effect.

The following examples elaborate on the value-spec strings:

"Hybrid" By default, all string comparisons are case-sensitive. By

default, the string Hybrid does not match the string
HYBRID.
"HYBRID" If IGNORECASE=TRUE in the DISCRETEATTRMAP
statement, then you can specify an upper-, lower-, or
mixed-cased string for the value-spec string. When
IGNORECASE=TRUE, the string Hybrid matches the string
HYBRID.
"15JAN2011" If a numeric column is being mapped with a VALUE
statement, then you must specify the formatted value of
the column. This example shows how to specify the value
1508 Chapter 18 / Discrete Attribute Map Statements

of a numeric SAS date column that has a DATE9. format

associated with it.
"." If a numeric column has a missing value, then you should
use the formatted value for missing, which is "." by default. If
the MISSING= system option is used to change the default
string, then you should match that value. For example, if
OPTIONS MISSING="M" is specified in the SAS program,
then you should use "M" in the VALUE statement to
represent missing values.
"" If a character column has a missing value, then you should
use the formatted value for missing, which is " " by default.
"Truck" Multiple strings can be specified to indicate that each of the
"SUV" specified values matches to the same graphical properties.
It does not mean that a single new category is formed. The
list of strings is separated by blanks, and each string is
enclosed in its own set of quotation marks.

Restrictions A dynamic variable reference or run-macro variable reference

cannot be used in value-spec.

If a user-defined format is associated with the classification

column, then you should specify the same formatted strings that
appear in the format definition.

Note When the specified value strings are compared with the data
values, leading blanks are honored and trailing blanks are ignored.

Tip In SAS 9.4M1 and earlier releases, if you create and use a format to
display a special value for missing character values, in some cases,
" " is returned instead of the formatted value. If your attribute map
assigns attributes to the formatted missing value in that case, the
attributes are not assigned to the missing values. To correct this
problem, specify both " " and your formatted missing value in the
VALUE statement for your missing-value attributes. This issue is
resolved in SAS 9.4M2.

VALUE Statement Optional Arguments

The following options can be used in the VALUE statement.

FILLATTRS=style-element | style-element (fill-options) | (fill-options)

the fill attributes to be used when an attribute map is applied to filled areas in a
graph.

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the
attribute map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP, unspecified

attributes derive attributes from the GraphDataDefault style
element.
 DISCRETEATTRMAP Statement 1509

Restriction In SAS 9.4 and earlier releases, the TRANSPARENCY= fill option is
ignored. Starting with SAS 9.4M2, the TRANSPARENCY= fill
option is supported.

Requirement When you specify TRANSPARENCY, you must also specify

COLOR. Otherwise, TRANSPARENCY is ignored.

See “General Syntax for Attribute Options” on page 1599 for the
syntax on using a style-element.

“Fill Color Options” on page 1600 for available fill-options.

LINEATTRS=style-element | style-element (line-options) | (line-options)

specifies the line attributes to be used when an attribute map is applied to lines
in a graph.

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the
attribute map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP, unspecified

attributes derive attributes from the GraphDataDefault style
element.

Restrictions In SAS 9.4 and earlier releases, the THICKNESS= line option is
ignored. Starting with SAS 9.4M1, the THICKNESS= line option is
supported.

Line option PATTERN= is ignored for the bar outlines in histograms.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Line Options” on page 1602 for available line-options.

MARKERATTRS=style-element | style-element (marker-options) | (marker-

options)
specifies the marker attributes to be used when an attribute map is applied to
marker symbols in a graph.

Defaults When DISCRETELEGENDENTRYPOLICY=DATA, unspecified

attributes receive the attributes that they would have if the
attribute map were not defined.

When DISCRETELEGENDENTRYPOLICY=ATTRMAP, unspecified

attributes derive attributes from the GraphDataDefault style
element.

Restrictions The WEIGHT= marker option is ignored.

In SAS 9.4 and earlier releases, the SIZE=, TRANSPARENCY=, and

WEIGHT= marker options are ignored. Starting with SAS 9.4M1, the
SIZE=, TRANSPARENCY=, and WEIGHT= marker options are
supported.
1510 Chapter 18 / Discrete Attribute Map Statements

For filled, outlined markers, COLOR and TRANSPARENCY are

ignored.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style-element.

“Marker Options” on page 1604 for available marker-options.

TEXTATTRS=style-element | style-element (text-options) | (text-options)

specifies the text attributes to use when an attribute map is applied to text in a
graph.

Default The GraphDataText style element.

Restriction This option is honored only by the TEXTGROUP= option in the

AXISTABLE statement and the GROUP= option in the TEXTPLOT
statement.

See “General Syntax for Attribute Options” on page 1599 for the syntax
on using a style element.

“Text Options” on page 1606 for available text options.

Details

About Discrete Attribute Maps

Attribute maps can be useful for ensuring that a particular value (a company name,
for example) is always represented by the same visual characteristics in your
graphs, regardless of the value’s order in the input data. When specified directly in a
discrete legend, an attribute map can be used to display legend entries for group
values that are not in the data. (See “Displaying Legend Entries for Group Values
That Are Not in the Data” on page 1512.)

If you do not want to manage the graphical properties that are associated with each
unique discrete value, then you can simply specify an input column in the grouping
option. In that case, each discrete value is represented by different graphical
properties in the graph. The default properties are derived from options that are set
in the plot statement or from the GraphData1–GraphDataN style elements that are
defined in the ODS style that is in effect. However, to ensure that specific graphical
properties are used to represent a discrete value in the graph, regardless of that
value’s order in the data, you can use an attribute map to create that association.
Any values in the data that are not accounted for in the attribute map are assigned
the graphical properties that they would receive if the attribute map is not defined.

Defining a Discrete Attribute Map in a DISCRETEATTRMAP

Block
A DISCRETEATTRMAP block creates an attribute map that matches graphical
properties to discrete values. The attribute map can be associated with a data input
 DISCRETEATTRMAP Statement 1511

column that is used as a classification variable in a graph. It can also be specified

directly in a discrete legend.

To define a discrete attribute map in a DISCRETEATTRMAP block, do the following:

n Use the DISCRETEATTRMAP statement to start the attribute definition and
assign a name to it. The DISCRETEATTRMAP statement must be located within
the BEGINGRAPH block but outside of the outermost layout block. It cannot be
nested in another DISCRETEATTRMAP statement or in a RANGEATTRMAP
statement. The DISCRETEATTRMAP statement determines whether the data
mapping is case-sensitive and whether leading blanks are trimmed from the
data values during the mapping.
n Nest within the DISCRETEATTRMAP block at least one VALUE statement that
specifies graphical properties to associate with a classification value. See
“About the VALUE Statement” on page 1511. Use one VALUE statement for each
classification value that you want to map. Values that are referenced in the
attribute map can be character or numeric. Discrete values that are not
accounted for in the attribute map are assigned the graphical properties that
they would receive if the attribute map is not defined.
n Use the ENDDISCRETEATTRMAP statement to close the block.

For information about how to use the attribute map, see “Using a Discrete Attribute
Map” on page 1512.

About the VALUE Statement

A VALUE statement within the DISCRETEATTRMAP block associates graphical
properties with a discrete value in the attribute map. To associate graphical
properties with multiple values, specify multiple VALUE statements in the attribute
map using the following general syntax for each of the statements:

VALUE value-spec </options>;

For a specific example, see the “Example Program” on page 1515.

If the discrete attribute map is referenced by a plot statement in the template, then
the graphical properties that are defined in the VALUE statements are used in the
plot. If a discrete legend is generated for the plot, then the graphical properties are
represented in that legend.

If a discrete attribute map is referenced directly in a DISCRETELEGEND statement,

then the graphical properties that are defined in the VALUE statement are mapped
directly to the legend and are independent of the values in the data. For more
information and an example, see the “Defining a Discrete Attribute Map in a
DISCRETEATTRMAP Block” on page 1510 .

If two or more VALUE statements define attributes to associate with the same
classification values, then the last VALUE statement's settings are used.

By default when comparing a column's value to a string that is specified for the
VALUE statement's value-spec,
n the column value is formatted to a string, using the format that is defined for
the column or the default format if no format is defined for the column
1512 Chapter 18 / Discrete Attribute Map Statements

n leading spaces are trimmed from the string that is specified in the VALUE
statement
n a case-sensitive comparison is performed between the column string and the
VALUE string.

To change the default behavior for the comparison, you can use the
DISCRETEATTRMAP statement’s TRIMLEADING= and IGNORECASE= options.

Using a Discrete Attribute Map

To use a discrete attribute map that is defined in a DISCRETEATTRMAP block, do
the following:
n Use the DISCRETEATTRVAR statement to create a named association between
the defined attribute map and the input column that contains the classification
values. The ATTRMAP= argument identifies the attribute map and the VAR=
argument identifies the input column. Use the ATTRVAR= argument to assign a
name that can be used to reference the named association in plot statements
within the template.
n Reference the attribute map variable in the plot statements where needed. See
“Referencing a Discrete Attribute Map in Plot Statements” on page 1499.

For an example, see “Example Program” on page 1515.

Displaying Legend Entries for Group Values That Are Not in the
Data
To display legend entries for the values in a grouped plot, you typically use the plot
statement’s NAME= option to assign a name to the plot, and then reference that
name in the DISCRETELEGEND statement. In this usage case, the legend displays
entries for the group values that exist in the data.

To represent all of the group values in the legend, regardless of whether they exist
in the data, you can specify DISCRETELEGENDENTRYPOLICY=ATTRMAP in the
DISCRETEATTRMAP statement for the attribute map. When the
DISCRETELEGENDENTRYPOLICY=ATTRMAP option is in effect, the associated
plot contributes all of the items in the attribute map to the discrete legend
regardless of whether they exist in the data.

Displaying all of the items in a discrete attribute map in the legend can be useful for
flagging data in the graph. For example, assume you have weight and height values
for all students in an analysis group and you want to create a scatter plot of the
data. However, some of the observations are incomplete and do not record the
student’s sex. You want to include the incomplete observations in your plot, but
you want to visually distinguish them from the others. In that case, you can do the
following:
n Represent the unknown values in your data by entering the value U for sex.

n Define the discrete attribute map for the plot in a DISCRETEATTRMAP block.
Include the DISCRETELEGENDENTRYPOLICY=ATTRMAP option in the
DISCRETEATTRMAP statement. Include VALUE statements that specify the
properties for the M, F, and U values.
 DISCRETEATTRMAP Statement 1513

n Use the DISCRETEATTRVAR statement to create a discrete attribute map

variable that associates the attribute map with column Sex in the input data.
n In the SCATTERPLOT statement, specify the discrete attribute map variable
name in the GROUP= option.
n In the DISCRETELEGEND statement, reference the scatter plot.

As a result, the legend displays the attribute-map definitions, and observations

with the value U in column Sex are displayed as incomplete observations.

Here is an example of a template that uses a discrete attribute map to uniquely

display observations with the value U in column Sex.
proc template;
define statgraph discreteattrmapdatapresent;
begingraph;
entrytitle "Height and Weight by Sex";
/* Define the attribute map and assign the name "symbols" */
discreteattrmap name="symbols" / trimleading=true ignorecase=true
discretelegendentrypolicy=attrmap;
value "M" / markerattrs=(color=blue symbol=diamondfilled);
value "F" / markerattrs=(color=green symbol=circlefilled);
value "U" / markerattrs=(color=red symbol=starfilled);
enddiscreteattrmap;

/* Create attribute map variable GROUPMARKERS to associate

attribute
map SYMBOLS with column Sex */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";

/* Use the attribute map by referencing GROUPMARKERS in a plot

statement */
layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers;
discretelegend "scatter";
endlayout;
endgraph;
end;

To test the template, you can generate test data set Testclass from Sashelp.Class
by changing the value of column Sex to U for John and Carol as shown in the
following code.
data testclass;
set sashelp.class;
if (name="John") then sex="U";
if (name="Carol") then sex="U";
run;

The following figure shows the output of this template when it is run with the test
data set Testclass.
1514 Chapter 18 / Discrete Attribute Map Statements

The two red stars in the plot indicate the observations with the value U in column
Sex. The next figure shows the output of this template when it is run with the
complete data in data set Sashelp.Class.

The absence of the red stars in the plot indicates that the value U is no longer
present in the data. All of the observations are now complete. When
DISCRETELEGENDENTRYPOLICY=ATTRMAP is specified for an attribute map, be
aware that the legend entries that are contributed by an associated plot are defined
entirely by the attribute map and are independent of the data.

Example: DISCRETEATTRMAP and

DISCRETEATTRVAR Statements
The following graph was generated by the “Example Program” on page 1515. The
example defines graphical properties to associate with classification values in an
input column that is used in a scatter plot. The DISCRETEATTRMAP statement
starts the attribute map definition, assigns a name to it, and ensures that the data
mapping is not case sensitive. The VALUE statements define the colors and marker
symbols to associate with the values M and F. The DISCRETEATTRVAR statement
associates the attribute map with the data column Sex and assigns the name
GROUPMARKERS to the association. The SCATTERPLOT statement references the
named association in its GROUP= option.
 DISCRETEATTRMAP Statement 1515

Example Program
proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";

/* define the attribute map and assign the name "symbols" */

discreteattrmap name="symbols" / ignorecase=true ;
value "m" / markerattrs=(color=blue symbol=diamondfilled) ;
value "f" / markerattrs=(color=red symbol=circlefilled) ;
enddiscreteattrmap ;

/* associate the attribute map with input data column Sex and
assign
* the name GROUPMARKERS to the named association */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols" ;

/* reference GROUPMARKERS in a plot statement */

layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers ;
discretelegend "scatter";
endlayout;
endgraph;
end;

proc sgrender data=sashelp.class template=scatterplot;

run;
 1516 Chapter 18 / Discrete Attribute Map Statements

DISCRETEATTRVAR Statement
Creates a named association between a user-defined discrete attribute map and an input data
column.

Requirements: You must place the DISCRETEATTRVAR statement directly inside the
BEGINGRAPH block between the BEGINGRAPH statement and the first LAYOUT
statement. Do not embed it in any other GTL statement block.
A discrete attribute map must be created using the DISCRETEATTRMAP
statement.
See: “Example: DISCRETEATTRMAP and DISCRETEATTRVAR Statements” for an
example.

Syntax
DISCRETEATTRVAR ATTRVAR=attrvar-name
VAR=data-column | expression | dynamic
ATTRMAP="attrmap-name";

Required Arguments
ATTRVAR=attrvar-name
specifies a SAS name for this association between the attribute map and the
input column. This name must be unique within the template and can be
referenced by other statements that can be associated with the attribute map.
The attribute map variable name should not be used in an expression. If it is,
then the results are unpredictable.

Restrictions The name that is assigned in this argument is used to associate an

attribute map with the discrete values in an input data column.
Thus, it is not the name to reference when you want to display
legend entries that are independent of the data. For that special
use, a DISCRETELEGEND statement can reference the attribute
map directly by the name that is assigned in the
DISCRETEATTRMAP statement. For more information, see the
DISCRETEATTRMAP statement’s “Defining a Discrete Attribute
Map in a DISCRETEATTRMAP Block” on page 1510 .

Using an attribute variable in an expression is not supported.

Note The assigned SAS name can be the same as the name of the data
input column, but it is not recommended. If an assigned attrvar-
name matches the name of an input data column, then the attrvar-
name takes precedence.
 DISCRETEATTRVAR Statement 1517

VAR=data-column | expression | dynamic

specifies an input data column to be associated with an attribute map at run
time. If an expression is used, a new column of transformed values is created
and then matched with the attribute map.

Interaction If the column is not found or the column is of the wrong type for the
attribute map, then the DISCRETEATTRVAR statement is ignored.

Tip The input data column can be character or numeric, but the values
must match the type of the values that are specified in the attribute
map. For numeric columns, all values are treated as discrete values.

ATTRMAP="attrmap-name "
specifies the name of an existing discrete attribute map.

Restriction The attrmap-name is case sensitive, must be enclosed in quotation

marks, and must be the name that was assigned to the attribute map
in the DISCRETEATTRMAP statement’s NAME= argument.

Details
The DISCRETEATTRVAR statement creates and names an association between
graphical properties that are specified in a DISCRETEATTRMAP block and a
classification column that is in the data. The name that is assigned to the
association in the DISCRETEATTRVAR statement is the name that plot statements
must reference to use the attribute map.

Defining and using a discrete attribute map requires you to coordinate settings on
several statements. For more information, see the DISCRETEATTRMAP
statement’s “Defining a Discrete Attribute Map in a DISCRETEATTRMAP Block” on
page 1510 .

The DISCRETEATTRVAR statement must be located within the BEGINGRAPH

block but outside of the outermost layout block. It cannot be nested in a
DISCRETEATTRMAP statement.

Example: DISCRETEATTRVAR Statement with an

Attribute Map Data Set
Starting with SAS 9.4M1, you can define an attribute map in a SAS data set. This
example shows you how to use the DISCRETEATTRVAR statement with an
attribute map that is defined in a SAS data set. It is the example in “Example:
DISCRETEATTRMAP and DISCRETEATTRVAR Statements” on page 1514 modified
to use a SAS data set instead of a DISCRETEATTRMAP block.

Here is the output for this example.

 1518 Chapter 18 / Discrete Attribute Map Statements

Example Program
Here is the SAS code for this example.
/* Create the attribute map data set */
data attrds;
input ID $1-7 VALUE $9 MARKERSYMBOL $11-23 MARKERCOLOR $25-30;
datalines;
symbols M diamondfilled blue
symbols F circlefilled red
;
run;

/* Define the template for this graph */

proc template;
define statgraph scatterplot;
begingraph;
entrytitle "Height and Weight by Sex";

/* Associate the attribute map with input data column Sex and
assign
* the name GROUPMARKERS to the named association */
discreteattrvar attrvar=groupmarkers var=sex attrmap="symbols";

/* Reference GROUPMARKERS in a plot statement */

layout overlay;
scatterplot x=height y=weight / name="scatter"
group=groupmarkers;
discretelegend "scatter";
endlayout;
 DISCRETEATTRVAR Statement 1519

endgraph;
end;

/* Generate the graph */

proc sgrender data=sashelp.class dattrmap=attrds template=scatterplot;
run;

Details
The attribute map is defined in data set Attrds. The data set includes the required
columns ID and VALUE, which specify a name for this attribute map and the value
of the classification variable that is to be mapped. The MarkerSymbol and
MarkerColor columns specify the graphical properties that are mapped to each
value. The values specified in the Value column are case-sensitive. Unlike the
DISCRETEATTRMAP statement, when you define your attribute map in a data set,
there is no option that you can specify to ignore case. You must ensure that the
case of the values in your attribute map matches the case in the actual data.

The DISCRETEATTRVAR statement associates the attribute map with the data
column Sex and assigns the name GROUPMARKERS to the attribute map variable.
The ATTRMAP= option specifies the name in the ID column of the attribute map
data set. The value is case-sensitive. The attribute map variable name is specified
in the GROUP= option in the SCATTERPLOT statement.

Finally, in the SGRENDER statement, the DATTRMAP= option specifies the name
of the attribute map data set. When your attribute map is defined in a SAS data set,
you must specify the name of the data set in the DATTRMAP= option in the
SGRENDER statement that renders the graph.

You can also use the DATTRVAR= statement with the SGRENDER statement to
associate the attribute map with the data column Sex. See “Referencing an
Attribute Map in Your Plot Statements” on page 1499.
1520 Chapter 18 / Discrete Attribute Map Statements
 1521

19
Range Attribute Map Statements

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1521
RANGEATTRMAP Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1521
RANGEATTRVAR Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1531

Dictionary

RANGEATTRMAP Statement
Creates an attribute map that matches colors to numeric values or numeric ranges so that the colors
can be associated with the values of an input data column.

Restriction: A RANGEATTRMAP cannot be directly referenced in a legend statement.

Requirements: You must place the RANGEATTRMAP block directly inside the BEGINGRAPH block
between the BEGINGRAPH statement and the first LAYOUT statement. Do not
embed it in any other GTL statement block.
The RANGEATTRMAP block must contain at lease one RANGE statement.
The RANGEATTRVAR statement must be used to associate the range attribute
map with a data column.
Note: The RANGEATTRMAP statement defines a set of graphical properties for ranges of
data values.
See: “RANGEATTRVAR Statement” on page 1531

Syntax
RANGEATTRMAP NAME="string";
1522 Chapter 19 / Range Attribute Map Statements

RANGE range-spec < /options >;

<… more-RANGE statements …>
ENDRANGEATTRMAP;

Required Argument
NAME="string "
assigns a name to the attribute definition for reference in a RANGEATTRVAR
statement.

Restrictions The string is case sensitive, must be enclosed in quotation marks,

and must define a name that is unique among RANGEATTRMAP
names within the template.

This option does not support variables that are created by the
DYNAMIC, MVAR, and NMVAR template statements.

RANGE Statement Required Argument

range-spec
specifies a range of numeric values or a keyword, such as OTHER or MISSING.

A range of numeric values is specified in the form low-value - high-value. Both

the low value and the high value can be specified as an unformatted numeric
value. A less-than symbol (<) can be placed after the low numeric value, before
the high numeric value, or in both positions to exclude that value from the range
endpoint (similar to the VALUE statement of PROC FORMAT). If you are
excluding the first value in a range, then put the < after the low value. If you are
excluding the last value in a range, then put the < before the high value. You can
also exclude both the low and the high value.

For example, the following range does not include 0:

0 < - 100

Likewise, the following range does not include 100:

0 - < 100

If a value at the high end of one range also appears at the low end of another
range and you do not use the < exclusion notation, then the value is assigned to
the first range.

If two or more RANGE statements define colors to associate with the same
numeric values or ranges, then the first RANGE statement's settings are used. If
any RANGE statement's range overlaps another RANGE statement's range (for
example, 10 - 20 and 15 - 25), then the entire attribute map is ignored and
default coloring is used.

If two ranges share a common endpoint (for example, 10 - 20 and 20 - 30) and
no exclusion operator is used, then the common endpoint belongs to the lower
encountered range (10 - 20 in this case). The order of the specification does not
matter.

To set a single numeric value, specify the same value for both the low value and
the high value.
 RANGEATTRMAP Statement 1523

If the low value is not less than or equal to the high value, then the range
specification is invalid, and the RANGE statement is ignored in the attribute
map definition.

Note: If a range is not defined for keyword OTHER, then gaps within the
attribute map ranges are assigned the default color that is defined by the
GraphOther:ContrastColor style reference.

Rather than using a numeric value, you can specify one of the following
keywords as the low value or the high value:

MIN indicates the minimum data value for column values.

MAX indicates the highest data value for the column values.
MAXABS indicates max(abs(MIN) , abs(MAX))
NEGMAX indicates -MAX
NEGMAXABS indicates -max(abs(MIN) , abs(MAX))

Rather than specifying a low-value-to-high-value range, you can use one of the
following keywords for the range specification:

MISSING
indicates a mapping for missing values. The visual attributes for this setting
are obtained from the GraphMissing style element. If one RANGE statement
specifies this value and another RANGE statement specifies keyword
OTHER, then the OTHER range does not include missing values.

UNDER
creates a range for all data values between the lowest mapped value and the
lowest actual data value. The visual attributes for this setting are obtained
from the GraphUnderflow style element. If one RANGE statement specifies
this value and another RANGE statement specifies keyword OTHER, then
the OTHER range does not include underflow values.

OVER
creates a range for all data between the highest mapped value and the
highest actual data value. The visual attributes for this setting are obtained
from the GraphOverflow style element. If one RANGE statement specifies
this value and another RANGE statement specifies keyword OTHER, then
the OTHER range does not include overflow values.

OTHER
creates a category for all other column values not explicitly assigned to a
range. The OTHER values can be composed of several non-contiguous
ranges. The visual attributes for this setting are obtained from the
GraphOther style element.

Restriction A dynamic variable reference or run-time macro variable reference

cannot be used in range-spec.
1524 Chapter 19 / Range Attribute Map Statements

RANGE Statement Optional Arguments

RANGEALTCOLOR=style-reference | color | GRADIENTSTEPPER(color1,color2,
num-steps, step)
specifies a contrast color for lines, markers, and text to represent the defined
value range.

GRADIENTSTEPPER (color1,color2, num-steps, step)

a gradient stepper that partitions a color range into equal-sized intervals and
returns the color that is in the specified step position. The start and end
colors for the range are specified in parameters color1 and color2. The
number of equal-sized intervals is specified in parameter num-steps, and the
step position for the color to return is specified in parameter step.

Example:
rangeattrmap name="incomemap";
range min - 13000 / rangealtcolor=gradientstepper(red,green,4,1);
range 13000 < - 25000 / rangealtcolor=gradientstepper(red,green,4,2);
range 25000 < - 50000 / rangealtcolor=gradientstepper(red,green,4,3);
range 50000 < - max / rangealtcolor=gradientstepper(red,green,4,4);
endrangeattrmap;

Default The GraphDataDefault:ContrastColor style reference.

Restriction A dynamic variable reference or run-time macro variable reference

can be used to specify individual token values, but it cannot be used
to replace the entire list of tokens.

Interaction If this option is specified, then the RANGEALTCOLORMODEL=

option is ignored

Tips Because lines and markers have less surface area, choose a
contrasting color in order to improve visibility.

Use RANGECOLOR= to control fill color.

RANGECOLOR=style-reference | color | GRADIENTSTEPPER(color1,color2, num-

steps, step )
specifies a color for area fill to represent the defined value range.

GRADIENTSTEPPER (color1,color2, num-steps, step )

a gradient stepper that partitions a color range into equal-sized intervals and
returns the color that is in the specified step position. The start and end
colors for the range are specified in parameters color1 and color2. The
number of equal-sized intervals is specified in parameter num-steps, and the
step position for the color to return is specified in parameter step. Example:
rangeattrmap name="incomemap";
range min - 13000 / rangecolor=gradientstepper(red,blue,4,1);
range 13000 < - 25000 / rangecolor=gradientstepper(red,blue,4,2);
range 25000 < - 50000 / rangecolor=gradientstepper(red,blue,4,3);
range 50000 < - max / rangecolor=gradientstepper(red,blue,4,4);
endrangeattrmap;

Default The GraphDataDefault:ContrastColor style reference.

 RANGEATTRMAP Statement 1525

Restriction A dynamic variable reference or run-time macro variable reference

can be used to specify individual token values, but it cannot be used
to replace the entire list of tokens.

Interaction If this option is specified, then the RANGECOLORMODEL= option is

ignored.

Tip Use RANGEALTCOLOR= to control line and marker color.

RANGEALTCOLORMODEL=style-element | (list-of-colors)
specifies either a style element or a list of one or more specific contrast colors
for lines, markers, and text to represent the defined value range in this
argument.

style-element
specifies the name of a style element. To display the range as a gradient
ramp, choose a style element such as TwoColorRamp, TwoColorAltRamp,
ThreeColorRamp, or ThreeColorAltRamp. The style element should contain
the following style attributes:

STARTCOLOR specifies a color for the smallest data value.

NEUTRALCOLOR specifies a color for the midpoint of the data range.
This attribute is not needed when defining a two-
color ramp.
ENDCOLOR specifies a color for the highest data value.

To display the range or single value contrast color as a color that is defined
in a style, use a style-reference (for example, GraphData1:color) to refer to
a Color attribute. The following style references correspond to the keywords
that are available in this statement’s range-spec argument:

range-spec Keyword Corresponding Style Reference

MISSING GraphMissing:ContrastColor

OTHER GraphOther:ContrastColor

UNDER GraphUnderflow:ContrastColor

OVER GraphOverflow:ContrastColor

(list-of-colors)
a space-separated list of two or more color keywords that is enclosed in
parentheses.

Two colors create the endpoints of a ramp. The first color is assigned to the
low value in the range specification, and the second color is assigned to the
high value. Three or more colors partition the range specification into n-1
equal-sized intervals where each adjacent color pair defines a two-color
ramp.
1526 Chapter 19 / Range Attribute Map Statements

Interaction This option is ignored if the RANGEALTCOLOR= option is specified.

Tips Because lines and markers have less surface area, choose
contrasting colors in order to improve visibility.

Use RANGECOLORMODEL= to control fill color.

RANGECOLORMODEL=style-element | (list-of-colors)
specifies either a style element or a list of one or more specific colors for area
fills to represent the defined value range in this range-spec argument.

style-element
specifies the name of a style element. To display the range as a gradient
ramp, choose a style element such as TwoColorRamp, TwoColorAltRamp,
ThreeColorRamp, or ThreeColorAltRamp. The style element should contain
the following style attributes:

STARTCOLOR specifies a color for the smallest data value.

NEUTRALCOLOR specifies a color for the midpoint of the data range.
This attribute is not needed when defining a two-
color ramp.
ENDCOLOR specifies a color for the highest data value.

To display the range or single value as a color that is defined in a style, use a
style-reference (for example, GraphData1:color) to refer to a color attribute.
The following style references correspond to the keywords that are available
in this statement’s range-spec argument:

range-spec Keyword Corresponding Style Reference

MISSING GraphMissing:ContrastColor

OTHER GraphOther:ContrastColor

UNDER GraphUnderflow:ContrastColor

OVER GraphOverflow:ContrastColor

(list-of-colors)
a space-separated list of two or more color keywords that is enclosed in
parentheses.

Two colors create the endpoints of a ramp. The first color is assigned to the
low value in the range specification, and the second color is assigned to the
high value. Three or more colors partition the range specification into n–1
equal-sized intervals, where each adjacent color pair defines a two-color
ramp.

Interaction This option is ignored if the RANGECOLOR option is specified.

Tip Use RANGEALTCOLORMODEL= to control line and marker color.

 RANGEATTRMAP Statement 1527

Details

Creating and Using a Range Attribute Map

The RANGEATTRMAP statement creates an attribute map that matches colors to
numeric values or value ranges. The attribute map can be associated with a data
input column that uses color to represent response values in a graph. Attribute
maps can be useful for controlling the application of gradient color in a graph. In
addition, they enable you to map color to data values, independent of the actual
data that is used in the graph. For example, for temperature data, you can set Blue
for 0 and Red for 100, even if the values 0 and 100 are not in the data.

Defining and using the attribute map requires you to coordinate settings on several
statements:
n Use the RANGEATTRMAP statement to start the attribute definition and assign
a name to it.
n Nest within the RANGEATTRMAP block at least one RANGE statement that
specifies a numeric value or numeric range and the color, contrast color, or both,
to associate with that value or range. Use one RANGE statement for each value
range that you want to map. You can use keywords like MIN and MAX in the
range specification. For complete details about the range specifications, see the
RANGE statement.
n Use the RANGEATTRVAR statement to create a named association between
the defined attribute map and the input column that contains the numeric
values. The ATTRMAP= argument identifies the attribute map and the VAR=
argument identifies the input column. Use the ATTRVAR= argument to assign a
name that can be used to reference the named association in plot statements
within the template.
n Reference the attribute map where needed. See “Referencing a Range Attribute
Map in Plot Statements” on page 1503.

The RANGEATTRMAP statement must be located within the BEGINGRAPH block

but outside of the outermost layout block. It cannot be nested in another
RANGEATTRMAP statement or in a RANGEATTRMAP statement. The
RANGEATTRMAP block must contain at least one RANGE statement. All values
that are referenced in the attribute map must be numeric. Data values that are not
accounted for in the attribute map receive the default color from the GraphOther,
GraphOverflow, or GraphUnderflow style element, depending on where the
unassigned values are relative to the specified data ranges.

Note: Unlike a DISCRETEATTRMAP statement, a RANGEATTRMAP statement

cannot be directly referenced in a legend. This is because the RANGE statements
can specify keywords like MIN and MAX that require a data association for
interpretation.

About the RANGE Statement

A RANGE statement within the RANGEATTRMAP block matches color to a
numeric value, or a color ramp to a numeric range. To match colors to multiple
1528 Chapter 19 / Range Attribute Map Statements

values or ranges, specify multiple RANGE statements using the following general
syntax for each of the statements:

RANGE range-spec < /options >;

For a specific example, see the “Example Program” on page 1529.

When defining multiple RANGE statements, be careful not to define conflicting

ranges. A RANGE statement might be syntactically correct by itself while
conflicting with settings in another RANGE statement.

Specifying Color in a RANGE Option

For specifying a color in one of the RANGE options, the color keywords can be any
of the following:
n a SAS color name (for example, blue)

n an RGB name (for example, CX0000FF or x0000FF)

n an HLS value (for example, H14E162D)

n a gray-scale color code (for example, GRAYBB)

n an HTML color name (for example, AZURE)

n a SAS session color (for example, DMSBLUE).

Examples

Example 1: Using a Range Attribute Map to Control Marker

Colors
The following graph was generated by the “Example Program” on page 1529. The
example defines colors to associate with data ranges in an input column that is
used in a grouped scatter plot. The RANGEATTRMAP statement starts the
attribute map definition and assigns a name to it. The RANGE statements define
the value ranges and the colors to associate with those ranges. Because the
SCATTERPLOT statement uses the ContrastColor style attribute for the marker
colors in a grouped plot, the RANGEALTCOLOR= and RANGEALTCOLORMODEL=
options are used in the RANGE statement to define the range colors. These options
can also be used to define range colors for lines and some text elements.

The highest range value specified, which is 0.002 in this example, does not have to
be an actual value in the data. The RANGEATTRVAR statement associates the
attribute map with the data column Density and assigns the name RANGEVAR to
the association. The SCATTERPLOT statement references the named association in
its MARKERCOLORGRADIENT= option.
 RANGEATTRMAP Statement 1529

Example Program
proc template;
define statgraph attrmap;
begingraph;
entrytitle "Height and Weight Distribution" ;

/* Define the attribute map and assign the name "densityrange." */

rangeattrmap name="densityrange" ;
range MIN - 0.0004 / rangealtcolor=blue ;
range 0.0004 < - 0.0014 / rangealtcolormodel=(lightpurple
lightred) ;
range 0.0014 < - 0.002 / rangealtcolor=red ;
endrangeattrmap ;

/* Associate the attribute map with input data column Density and
assign variable name RANGEVAR to the named association */
rangeattrvar attrvar=rangevar var=density attrmap="densityrange" ;

/* Reference RANGEVAR in the SCATTTERPLOT statement */

layout overlay;
scatterplot x=height y=weight / markercolorgradient=rangevar
markerattrs=(symbol=squarefilled size=6px) name="scatter";
continuouslegend "scatter" / orient=vertical
halign=right title="Density";
endlayout;
endgraph;
end;

ods graphics / reset width=475px;

 1530 Chapter 19 / Range Attribute Map Statements

proc sgrender data=sashelp.gridded(where=(count>0)) template=attrmap;

run;

Example 2: Using a Range Attribute Map to Control Fill Colors

The following graph was generated by the “Example Program” on page 1530. The
example defines colors to associate with data ranges in an input column that is
used in a surface plot. The RANGEATTRMAP statement starts the attribute map
definition and assigns a name to it. The RANGE statements define the value ranges
and the colors to associate with those ranges. Because the SURFACEPLOTPARM
statement uses the Color style attribute for fill colors, the RANGECOLORMODEL=
option is used in the RANGE statement to define the range colors.

The RANGEATTRVAR statement associates the attribute map with the data
column Depth and assigns the name DEPTHRANGE to the association. The
SURFACEPLOTPARM statement references the named association in its
COLORRESPONSE= option.

Example Program
proc template;
define statgraph surfaceplot;
dynamic type;
begingraph;
/* Define the attribute map and assign the name "depth." */
rangeattrmap name='depth';
range -10 - MAX / rangecolormodel=(green white);
range -20 - -10 / rangecolormodel=(blue green);
range MIN - -20 / rangecolormodel=(darkred blue);
endrangeattrmap;

/* Associate the attribute map with input data column Depth and
 RANGEATTRVAR Statement 1531

assign variable name DEPTHRANGE to the named association. */

rangeattrvar attrvar=depthrange var=Depth attrmap='depth';
entrytitle "Surface Plot of Lake Bed";
layout overlay3d /
xaxisopts=(griddisplay=on)
yaxisopts=(griddisplay=on)
zaxisopts=(griddisplay=on);

/* Reference DEPTHRANGE in the SURFACEPLOTPARM statement */

surfaceplotparm x=length y=width z=depth / name="depth"
reversecolormodel=true
colorresponse=depthrange;
continuouslegend "depth";
endlayout;
endgraph;
end;

/* Use the SAS/GRAPH G3GRID procedure to create gridded data for

the surface plot. */
proc g3grid data=sashelp.lake out=gridded;
grid width*length = depth / naxis1=75 naxis2=75;
run;

proc sgrender data= gridded template=surfaceplot;

run;

RANGEATTRVAR Statement
Creates a named association between a range attribute map of numeric values or value ranges and
an input data column.

Requirements: You must place the RANGEATTRVAR statement directly inside the BEGINGRAPH
block between the BEGINGRAPH statement and the first LAYOUT statement. Do
not embed it in any other GTL statement block.
A range attribute map must be created using the RANGEATTRMAP statement.
See: “Example 1: Using a Range Attribute Map to Control Marker Colors” on page 1528
for an example.

Syntax
RANGEATTRVAR ATTRVAR=attrvar-name
VAR=data-column | expression | dynamic
ATTRMAP="attrmap-name";

Required Arguments
ATTRVAR=attrvar-name
specifies a SAS name for the map. This name must be unique among the
RANGEATTRVAR statements within the template. The assigned name can be
1532 Chapter 19 / Range Attribute Map Statements

referenced by other statements that can be associated with the range attribute
map.the attrvar-name

Restriction Using an attribute variable in an expression is not supported.

VAR=data-column | expression | dynamic

specifies a numeric input data column to be associated with an attribute map at
run time. If an expression is used, then a new column of transformed values is
created and then matched with the attribute map.

Interaction If the column is not found or the column is of the wrong type for the
attribute map, then the RANGEATTRVAR statement is ignored.

ATTRMAP="attrmap-name "
specifies the name of an existing range attribute map.

Restriction The attrmap-name is case sensitive, must be enclosed in quotation

marks, and must be the name that was assigned to the attribute map
in the RANGEATTRMAP statement’s NAME= argument.

Details
The RANGEATTRVAR statement creates and names an association between colors
that are specified in a RANGEATTRMAP block and a numeric column that is in the
data. Attribute maps can be useful for controlling the application of gradient color
in a graph or specifying data values that are independent of the actual data. The
name that is assigned to the association in the RANGEATTRVAR statement is the
name that plot statements must reference to use the attribute map.

Defining and using a numeric-range attribute map requires you to coordinate

settings on several statements. For more information, see the RANGEATTRMAP
statement’s “Creating and Using a Range Attribute Map” on page 1527 .

The RANGEATTRVAR statement must be located within the BEGINGRAPH block

but outside of the outermost layout block. It cannot be nested in a
RANGEATTRMAP statement.
 1533

PART 12

Run-Time Programming Features

Chapter 20
Dynamic Variables and Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1535

Chapter 21
Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1539

Chapter 22
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543

Chapter 23
Conditional Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1555
1534
 1535

20
Dynamic Variables and Macro
Variables

Template Types on PROC TEMPLATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1535

DYNAMIC, MVAR, and NMVAR Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536
Dynamic Variables Compared to Macro Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1537

Template Types on PROC TEMPLATE

PROC TEMPLATE supports different template types such as COLUMN, TABLE,
HEADER, FOOTER, and STATGRAPH. All of these template types support run-time
variable substitution via dynamic variables or macro variables. For STATGRAPH
templates such variables should be declared within the scope of the template
definition before the BEGINGRAPH block.

PROC TEMPLATE;
DEFINE STATGRAPH template-name ;

DYNAMIC variable-1<"text-1"> <... variable-n<"text-n">>;

MVAR variable-1<"text-1"> <... variable-n<"text-n">>;
NMVAR variable-1<"text-1"> <... variable-n<"text-n">>;
NOTES "text";

BEGINGRAPH;
GTL statements;
ENDGRAPH;
END;
RUN;
 1536 Chapter 20 / Dynamic Variables and Macro Variables

DYNAMIC, MVAR, and NMVAR

Statements
Each of the DYNAMIC, MVAR, and NMVAR statements can define multiple
variables and an optional text-string denoting its purpose or usage. For example:

DYNAMIC YVAR "required" YLABEL "optional";

MVAR LOCATE "INSIDE or OUTSIDE" SYSDATE;
NMVAR TRANS "transparency factor";

Note: For template readability, it is helpful to adopt a naming convention for these
variables to distinguish them from actual option values or column names. Common
conventions include capitalization, or adding leading or trailing underscores to their
names.

Dynamic variables and macro variables can be referenced within the template
definition as
n argument or option values. For example:

seriesplot x=date y=YVAR / curvelabel=YLABEL

curvelabellocation=LOCATE datatransparency=TRANS;

n parts of some text strings. For example:

entrytitle "Time Series for " YLABEL;

entryfootnote "Created on " SYSDATE;

Dynamic variables and run-time macro variable references cannot resolve to

statement or option keywords.

Note that macro variable references should not be prefaced with an ampersand (&)
if you want them to resolve at run time.

Macro variables defined by MVAR are strings when they resolve, as with SYMGET()
in the DATA step.

Macro variables defined by NMVAR are converted to numeric tokens when they
resolve, as with SYMGETN() in the DATA step.

The values for a dynamic variable do not have to be provided by the data source.
Rather, you can provide the values in the DYNAMIC statement in PROC
SGRENDER, specifying the values as a space delimited list, enclosed in quotation
marks. Do not use parentheses in the specification.
 Dynamic Variables Compared to Macro Variables 1537

In the following example, the graph template specifies a dynamic variable named
TICKS, which is referenced on the XAXISOPTS= option in LAYOUT OVERLAY. The
DYNAMIC statement in PROC SGRENDER provides values for TICKS:
proc template;
define statgraph regress;
dynamic TICKS ;
begingraph;
layout overlay /xaxisopts=(linearopts=(tickvaluelist=TICKS));
scatterplot x=age y=weight;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.class template=regress;

dynamic TICKS="11 13 16" ;
run;

If your template uses a dynamic variable to specify a required attribute, such as a

variable name, and the name is misspelled or is not provided in the SGRENDER
procedure, then a warning is issued and the respective plot statement drops out of
the final graph. A graph is produced, but it might be a blank graph, or it might show
the results of all statements except those that are in error.

For more information about using dynamic variables and macro variables in your
templates, see SAS Graph Template Language: User’s Guide.

Dynamic Variables Compared to Macro

Variables
The main difference between dynamic variables and macro variables is how they
are initialized.

For dynamic variables, use the DYNAMIC statement with PROC SGRENDER. For
example,

proc sgrender data=sashelp.class template=timeseries;

dynamic yvar="inflation" ylabel="Inflation Rate";
run;

Values for dynamic variables that resolve to column names or strings should be
quoted. Numeric values should not be quoted.

For macro variables, use the current symbol table (local or global) to look up the
macro variable values at run time. For example,

%let locate=inside;
%let trans=0.3;
1538 Chapter 20 / Dynamic Variables and Macro Variables

proc sgrender data=sashelp.class template=timeseries;

dynamic yvar="inflation" ylabel="Inflation Rate";
run;

No initialization is needed for system macro variables like SYSDATE.

 1539

21
Expressions

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1539
GTL Expressions Compared to WHERE Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1540
An Expression in Statement Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1540

Overview
In Graph Template Language (GTL), as in Base SAS, an expression is an arithmetic
or logical expression that consists of a sequence of operators, operands, and
functions. An operand is a dynamic, a macro variable, a column, a function, or a
constant. An operator is a symbol that requests a comparison, logical operation, or
arithmetic calculation. In GTL, the expression must be enclosed in an EVAL
function.

Expressions can be used to set the following types of option values:

n a constant

n a column

n part of the text for ENTRYTITLE, ENTRYFOOTNOTE, and ENTRY statements

Here is an example of computing constants:

/* create reference lines at computed positions */

referenceline y=eval(mean(height)+2*std(height)) /
curvelabel="+2 STD";
referenceline y=eval(mean(height));
referenceline y=eval(mean(height)-2*std(height)) /
curvelabel="-2 STD";

Here is an example of creating a new column:

/* create a new column as a log transformation */

scatterplot x=date y=eval(log10(amount));

Here is an example of building a footnote text string:

 1540 Chapter 21 / Expressions

/* create a date and time stamp as a footnote */

entryfootnote eval(put(today(),date9.)||" : "||
put(time(),timeampm8.));

When you are building a text string, you can use the || operator to concatenate the
substrings if none of the substrings are a numeric character value. If one or more of
the substrings are a numeric character value such as “1.5”, then the || operator
might produce unexpected results. In that case, use the CATS function to
concatenate the strings instead. Here is an example of using the CATS function to
build a footnote text string that includes the value of dynamic variable maxWeight,
which stores a numeric character value:
/* indicate the maximum weight in a footnote */
entryfootnote eval(cats("Maximum weight: ", maxWeight," lbs"));

GTL Expressions Compared to WHERE

Expressions
Valid GTL expressions are identical to valid WHERE expressions. See the WHERE
statement documentation in Base SAS for a comprehensive list of operators and
operands. However, GTL expressions do not perform subset operations as WHERE
expressions do. The major difference in the result of a logical GTL expression on a
column is that a Boolean value is returned for each observation without changing
the number of observations.

For example, in the following line of code, the expression for the Y= argument does
not reduce the number of observations plotted.

scatterplot x=name y=eval(height between 40 and 60);

Instead, the computed numeric column for the Y= argument consists of 0s and 1s,
based on whether each observation’s Height column value is between 40 and 60.
Whenever expressions are used to create new columns, a new column name is
internally manufactured so that it does not collide with other columns in use.

An Expression in Statement Syntax

Throughout GTL documentation, you see expression used in statement
documentation:

BOXPLOT X=column | expression

Y=numeric-column | expression < /options>;
 An Expression in Statement Syntax 1541

For the X= argument, expression means any EVAL(expression) that results in either
a numeric or character column. An expression that yields a constant is not valid
because the X= argument does not accept constants.

Similarly, for the Y= argument, expression means any EVAL(expression) that results
in a numeric column. The expression cannot result in a character column or any
constant because the Y= argument accepts only a numeric column.

On the following REFERENCELINE statement, the X= argument can be a constant

(single line) or a column (multiple lines) that has the same data type as the axis.
This means that EVAL(expression) can result in a numeric or character column or
constant that agrees with the axis type.

REFERENCELINE X= x-axis-value | column | expression </options>;

Automatic Type Conversion. Although expressions that are used in a DATA step
perform automatic type conversion, GTL expression evaluation does not. Thus, you
must use function(s) to perform required type conversions in an expression.
Otherwise, the expression generates an error condition without warning when the
template is executed.

For example, consider the following GTL expression:

if(substr(value, 1, 2) = "11")

This expression uses the SUBSTR function to determine whether the first two
characters from VALUE evaluate to the string value "11". If VALUE is a string, then
the expression works properly. However, if VALUE is numeric, then the expression
generates an error condition. For a numeric, you must convert the value to a string
before passing it to the SUBSTR function. The following modification uses the
CATS function to perform the type conversion when necessary:

if(substr(cats(value, 1, 2)) = "11")

1542 Chapter 21 / Expressions
 1543

22
Functions

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543
SAS Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1544
SAS Functions That Can Be Used in a GTL Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1544
SAS Functions That Can Be Used to Create Flexible Templates . . . . . . . . . . . . . . . . . . 1544
Examples of Using the IFC and IFN SAS Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1545
Functions Defined Only in GTL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547
GTL Functions Used with the EVAL Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1548
Using the TYPEOF SAS Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1549
GTL Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1550
Commonly Used Summary Statistic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1550
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1552

Overview
GTL supports a large number of functions, including:
n SAS functions that can be used in the context of a WHERE expression

n functions that are defined only in GTL

n summary statistic functions

 1544 Chapter 22 / Functions

SAS Functions

SAS Functions That Can Be Used in a GTL

Template
Most of the SAS functions that are available in WHERE expressions can be used in
a GTL template. These SAS functions include:
n character-handling functions

n date and time functions

n mathematical and statistical functions

Not all SAS functions are available in WHERE expressions. Call routines and other
DATA-step-only functions (for example, LAG, VNAME, OPEN) are some examples
of functions that cannot be used. Not all functions that are available in WHERE
expressions are supported in GTL templates in all cases. The following form of the
PUT function is an example:
markercharacter = eval(put(amount, dollar7.2 -L))

This form results in an error when the template is compiled. However, the following
form is supported.
markercharacter = eval(put(amount, dollar7.2))

If you want to justify a string that is generated by the PUT function, use the LEFT or
RIGHT function with the PUT function as shown in the following example:
markercharacter = eval(left(put(amount, dollar7.2)))

Functions that accept null parameter values also might not be supported when you
specify a null parameter value.

For more information about SAS functions, see “Dictionary of Functions and CALL
Routines” in SAS Functions and CALL Routines: Reference.

SAS Functions That Can Be Used to Create Flexible

Templates
The following table shows some of the SAS functions can be used to increase the
flexibility of your template code.
 SAS Functions 1545

SAS Function Name Description

IFC(logical-expression, Returns the character value true-value if logical-expression

"true-value", "false- resolves to TRUE, false-value if it resolves to FALSE, or
value" <,"missing- missing-value if it resolves to a missing value. The TRUE,
value">) FALSE, and MISSING values must be enclosed in quotation
marks.

IFN(logical-expression, Returns the numeric value true-value if logical-expression

true-value, false-value resolves to TRUE, false-value if it resolves to FALSE, or
<,missing-value>) missing-value if it resolves to a missing value.

Examples of Using the IFC and IFN SAS Functions

The IFC and IFN functions return one of two character or numeric values based on
whether a conditional expression resolves to TRUE or FALSE. They can also return
an optional third value if the conditional expression resolves to a missing value.
These functions enable you to specify a value based on a conditional expression,
effectively creating a new data column. In some cases, these functions can be used
in place of IF-THEN-ELSE statements in your template code. As with other
functions, you must enclose the IFC and IFN functions in the EVAL function.

Here is an example that uses both the IFN and IFC functions for creating a sales-
based commission chart for employees in a sales group. Each employee in the
group works in one of two sales units: Products and Services. The data for this
example includes the employee ID, total sales, and sales unit code for each member
of the sales group. Here is the data.
data sales;
input empID totalSales salesUnit $18;
format totalSales dollar9.;
datalines;
112876 129489.44 P
112421 169842.97 S
115331 108763.51 S
110765 181009.22 P
113722 147688.78 P
;

The TotalSales column contains the total sales for each employee. The SalesUnit
column contains a code that identifies the sales unit in which each employee works.
The codes are P for the Products unit and S for the Services unit.

Here is the output for this example.

1546 Chapter 22 / Functions

The two bar charts show the total sales and earned commission for each employee.
The IFN function is used to compute commission for each employee based on his or
her total sales. Employees that achieved a sales total of $120,000 or more earn a
commission of 5% of their total sales. All other employees earn a commission of
2.5% of their total sales.

An axis table along the X axis shows the sales unit for each employee. The IFC
function is used to convert the P and S SalesUnit codes into more descriptive
values in the axis table. Because only two sales unit codes are used in this case, the
IFC function can be used for this purpose. This eliminates the need to add a new
column to the data in a DATA step or to create and apply a custom format to the
SalesUnit column.

Here is the SAS code that defines the template and generates the graph.
proc template;
define statgraph commission;
begingraph;
entrytitle "Sales-Based Commission";
layout overlay /
xaxisopts=(label="Employee ID")
yaxisopts=(label="Total Sales")
y2axisopts=(label="Commission"
linearopts=(viewmax=15000 tickvalueformat=dollar9.));
/* Generate the sales bar chart. */
barchart category=empID response=totalSales /
name="Sales" legendlabel="Total Sales" barwidth=0.3
discreteoffset=-0.2 fillattrs=graphData1;

/* Generate the commission bar chart. */

barchart category=empID
/* Use IFN to compute the commission. */
response=eval(ifn(totalSales >= 120000,
 Functions Defined Only in GTL 1547

totalSales * 0.05, /* 5% if TRUE */

totalSales * 0.025)) / /* 2.5% if FALSE */
name="Commission" legendlabel="Commission"
barwidth=0.3 yaxis=y2 discreteoffset=0.2
fillattrs=graphData2;

/* Add an axis table that shows the sales unit for each
employee. */
innermargin / align=bottom;
axistable x=empID
/* Use IFC to convert the codes to meaningful
values. */
value=eval(ifc(salesUnit = 'P',"Products",
"Services")) / display=(values);
endinnermargin;
discretelegend "Sales" "Commission";
endlayout;
endgraph;
end;
run;

proc sgrender data=sales template=commission;

run;

Functions Defined Only in GTL

GTL Functions Used with the EVAL Function

The following table shows some functions that are used only in GTL. As with other
functions, these must be enclosed within an EVAL. In all these functions, column
can be either the name of a column in the input data set or a dynamic / macro
variable that resolves to such a column.

Function Name Description

COLC("string-1", Converts a list of comma-separated string values into a

"string-1"<, "string-n" … temporary character column. Starting with SAS 9.4M2,
>) you can use this function to specify values in options that
accept a character column.

COLN(n-1, n-1<, n-N… Converts a list of comma-separated numeric values into a

>) temporary numeric column. You can use this function to
specify values in options that accept a numeric column.

COLNAME(column) Returns the case-sensitive name of the column.

 1548 Chapter 22 / Functions

Function Name Description

COLLABEL(column) Returns the case-sensitive label of the column. If no label

is defined for the column, then the case-sensitive name of
the column is returned.

EXISTS(item) Returns 1 if specified item exists, 0 otherwise. If item is a

column, then it tests for the presence of the column in the
input data set. If item is a dynamic / macro variable, then
it tests whether there has been a run-time initialization of
the variable.

EXPAND(numeric- Creates a new column as (numeric-column * frequency-

column, freq-column) column) .

ASORT(column, Sorts all columns of the data object by the values of

RETAIN=ALL) column in ascending order. SORT is an alias for ASORT.
Warning: if the RETAIN=ALL argument is not included,
column alone is sorted, not the other columns, causing
rowwise information to be lost.
Limitation: only one sort operation (whether an ASORT()
or DSORT() function) can be used within a single template
definition.

DSORT(column, Sorts all columns of the data object by the values of

RETAIN=ALL) column in descending order.
Warning: if the RETAIN=ALL argument is not included,
column alone is sorted, not the other columns, causing
rowwise information to be lost.
Limitation: only one sort operation (whether an ASORT()
or DSORT() function) can be used within a single template
definition.

NUMERATE(column) Returns a column that contains the ordinal position of

each observation in the input data set (similar to an Obs
column).

Examples

/* arrange bars in descending order of response values */

barchartparm category=region response=eval(dsort(amount,retain=all));

/* label outliers with their position in the data set */

/* it does not matter which column is used for NUMERATE() */
 Functions Defined Only in GTL 1549

boxplot x=age y=weight / datalabel=eval(numerate(age));

/* add information about the column being processed,

which is passed by a dynamic */
entrytitle "Distribution for " eval(colname(DYNVAR));

Using the TYPEOF SAS Function

The TYPEOF function returns the type of a specified column at run time.

TYPEOF(column)

This function returns the character ‘C’ if the specified column is a character column
or ‘N’ if it is a numeric column.

You can use the TYPEOF function to take specific actions in your template at run
time based on the input data type. Here is an example that creates a graph of two
columns and uses the TYPEOF function to select a graph type that is appropriate
for the column types. The result returned by the TYPEOF function determines the
graph type as follows:
n If both columns are numeric, then it creates a scatter plot.

n If the X column is character and the Y column is numeric, then it creates a

vertical bar chart.
n If the X column is numeric and the Y column is character, then it swaps the
category and response columns in the BARCHART statement and orients the
chart horizontally.

Here is the output for the third case, a numeric X column and a character Y column.

Here is the SAS code.

/* Define the graph template. */
 1550 Chapter 22 / Functions

proc template;
define statgraph plot;
dynamic cat resp; /* Category and response columns. */
begingraph;
entrytitle "Graph of " eval(collabel(resp)) " and "
eval(collabel(cat));
layout overlay;
/* If cat and resp are numeric, then generate a scatter plot.
Otherwise, generate a bar chart. */
if (typeof(cat) = "N" and typeof(resp) = "N")
scatterplot x=cat y=resp;
else
/* If cat is a character column, then generate a vertical bar
chart. Otherwise, generate a horizontal bar chart. */
if (typeof(cat) = "C")
barchart category=cat response=resp / stat=mean;
else
barchart category=resp response=cat /
stat=mean orient=horizontal;
endif;
endif;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.cars template=plot;

dynamic cat="MPG_CITY" resp="TYPE";
run;

Note: See “Functions Defined Only in GTL” on page 1547 for information about the
COLLABEL function.

GTL Summary Statistic Functions

Commonly Used Summary Statistic Functions

Several GTL summary statistic functions are available that return a numeric
constant, based on a summary operation on a numeric column. The results are the
same as if the corresponding statistics were requested with PROC SUMMARY.
These functions take a single argument that resolves to the name of a numeric
column. Here is an example.
number = EVAL(function-name(numeric-column))
 GTL Summary Statistic Functions 1551

The GTL summary statistic functions take precedence over similar multi-argument
DATA step functions. The following table lists the GTL summary statistic functions
that you can use.

Function Name Description

CSS Corrected sum of squares

CV Coefficient of variation

KURTOSIS Kurtosis

LCLM One-sided confidence limit below the mean

MAX Largest (maximum) value

MEAN Mean

MEDIAN Median (50th percentile)

MIN Smallest (minimum) value

N Number of nonmissing values

NMISS Number of missing values

P1 1st percentile

P5 5th percentile

P25 25th percentile

P50 50th percentile

P75 75th percentile

P90 90th percentile

P95 95th percentile

P99 99th percentile

PROBT p-value for Student’s t statistic

Q1 First quartile

Q3 Third quartile

QRANGE Interquartile range

 1552 Chapter 22 / Functions

Function Name Description

RANGE Range

SKEWNESS Skewness

STDDEV Standard deviation

STDERR Standard error of the mean

SUM Sum

SUMWGT Sum of weights

T Student’s t statistic

UCLM One-sided confidence limit above the mean

USS Uncorrected sum of squares

VAR Variance

Example
The following example uses GTL summary statistic functions to dynamically
construct reference lines and a table of statistics for a numeric variable, which is
supplied at run time.

Here is the graph for this example.

 GTL Summary Statistic Functions 1553

Here is the SAS code.

proc template;
define statgraph expression;
dynamic NUMVAR "required";
begingraph;
entrytitle "Distribution of " eval(colname(NUMVAR));
layout overlay / xaxisopts=(display=(ticks tickvalues line));
histogram NUMVAR;

/* create reference lines at computed positions */

referenceline x=eval(mean(NUMVAR)+2*std(NUMVAR)) /
lineattrs=(pattern=dash) curvelabel="+2 STD";
referenceline x=eval(mean(NUMVAR)) /
lineattrs=(thickness=2px) curvelabel="Mean";
referenceline x=eval(mean(NUMVAR)-2*std(NUMVAR)) /
lineattrs=(pattern=dash) curvelabel="-2 STD";

/* create inset */
layout gridded / columns=2 order=rowmajor
autoalign=(topleft topright) border=true;
entry halign=left "N";
entry halign=left eval(strip(put(n(NUMVAR),12.0)));
entry halign=left "Mean";
entry halign=left eval(strip(put(mean(NUMVAR),12.2)));
entry halign=left "Std Dev";
entry halign=left eval(strip(put(stddev(NUMVAR),12.2)));
endlayout;
endlayout;
endgraph;
end;
run;

proc sgrender data=sashelp.heart template=expression;

dynamic numvar="MRW";
1554 Chapter 22 / Functions

run;
 1555

23
Conditional Logic

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1555
Conditional Logic Determines Statement Rendering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1556
GTL Does Not Provide ELSE IF Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1557

Overview
GTL supports conditional logic that enables you to include or exclude one or more
GTL statements at run time:

IF (condition)
GTL-statement(s);
ELSE
GTL-statement(s);
ENDIF;

The IF statement requires an ENDIF statement. The IF block can be placed

anywhere within the BEGINGRAPH / ENDGRAPH block.

The condition is an expression that evaluates to a numeric constant, where all

numeric constants other than 0 and MISSING are true. There is an implied
EVAL(condition), so it is not necessary to include an EVAL as part of the condition.

Examples:
/* test a computed value */
if (weekday(today()) in (1 7))
entrytitle "Run during the work week";
else
entrytitle "Run during the weekend";
endif;

/* test for the value a numeric dynamic */

if (ADDREF > 0)
referenceline y=1;
 1556 Chapter 23 / Conditional Logic

referenceline y=0;
referenceline y=-1;
endif;

/* test for the value a character dynamic */

if (upcase(ADDREF) =: "Y")
referenceline y=1;
referenceline y=0;
referenceline y=-1;
endif;

/* test whether a dynamic is initialized */

if (exists(ADDREF))
referenceline y=1;
referenceline y=0;
referenceline y=-1;
endif;

Conditional Logic Determines Statement

Rendering
The GTL conditional logic is used only to determine which statements are rendered,
not to control what is in the data object. In the following example, the data object
contains columns for Date, Amount, and LOG(AMOUNT), but only one scatter plot
is created.

if (LOGFLAG)
scatterplot x=date y=amount;
else
scatterplot x=date y=log(amount);
endif;

Also, it is seldom necessary to test for the existence of option values set by
columns or dynamic variables. Consider the following statement:

scatterplot x=date y=amount / group=GROUPVAR;

This SCATTERPLOT statement is equivalent to the following code because option

values that are set by columns that do not exist or dynamic variables that are
uninitialized simply “drop out” at run time and do not produce errors or warnings:

if (exists(GROUPVAR))
scatterplot x=date y=amount / group=GROUPVAR;
else
scatterplot x=date y=amount;
endif;
 GTL Does Not Provide ELSE IF Syntax 1557

The GTL code that is conditional must be complete statements, or complete blocks
of statements, or both. The following IF block produces a compile error because
there are more LAYOUT statements than ENDLAYOUT statements:

/* this IF block produces a compile error */

if (exists(SQUAREPLOT))
layout overlayequated / equatetype=square;
else
layout overlay;
endif;

scatterplot x=XVAR y=YVAR;

endlayout;

This is the correct conditional construct:

if (exists(SQUAREPLOT))
layout overlayequated / equatetype=square;
scatterplot x=XVAR y=YVAR;
endlayout;
else
layout overlay;
scatterplot x=XVAR y=YVAR;
endlayout;
endif;

GTL Does Not Provide ELSE IF Syntax

GTL does not provide ELSE IF syntax, but you can create a nested IF/ELSE block as
follows:

IF (condition-1)
GTL-statement(s);
ELSE
IF (condition-2)
GTL-statement(s);
ELSE
GTL-statement(s);
ENDIF;
ENDIF;
1558 Chapter 23 / Conditional Logic
 1559

PART 13

Appendixes

Appendix 1
ODS GRAPHICS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1561

Appendix 2
Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595

Appendix 3
Display Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1599

Appendix 4
Format Support in ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1621

Appendix 5
Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 1627

Appendix 6
Memory Management for ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1633
1560
 1561

24
ODS GRAPHICS Statement

Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1561
ODS GRAPHICS Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1561

Dictionary

ODS GRAPHICS Statement

Enables or disables ODS Graphics processing and sets graphics environment options. This statement
affects ODS template-based (ODS Graphics) graphics only. The ODS GRAPHICS statement does not
affect device-based graphics (SAS/GRAPH).

Valid in: Anywhere

Category: ODS: Output Control

Default: ON. Beginning in SAS 9.4, ODS Graphics is enabled by default on all platforms
except z/OS. When running SAS in batch mode, the default is OFF.
Interaction: SAS/GRAPH device-based global statements such as GOPTIONS, SYMBOL,
PATTERN, AXIS, and LEGEND do not affect template-based graphics. The ODS
GRAPHICS statement does not affect device-based graphics.

Syntax
ODS GRAPHICS <OFF | ON> </ options> ;
1562 Chapter 24 / ODS GRAPHICS Statement

Summary of Optional Arguments

ANTIALIAS | NOANTIALIAS | ANTIALIAS= ON | OFF
specifies whether anti-aliasing is applied to the rendering of the line and
markers in any graph.
ANTIALIASMAX= n
specifies the maximum number of graphics elements before anti-aliasing
is disabled.
ATTRPRIORITY=COLOR | NONE
specifies a priority for cycling of the group attributes.
BORDER | NOBORDER | BORDER=ON | OFF
specifies whether to draw a border around each graph.
BYLINE=NOBYLINE | TITLE | FOOTNOTE
specifies how the BY line is displayed in graphs.
DATASKINMAX=n
specifies the maximum number of graphical elements allowed per plot
when a data skin is applied.
DISCRETEMAX=n
specifies the maximum number of discrete values to be shown in any
graph.
DRILLTARGET="_blank" | "_self" | "_parent" | "_top" | "frame-name"
specifies the window that displays the drill-down output.
GROUPMAX=n
specifies the maximum number of group values to be shown in any graph.
HEIGHT=dimension
specifies the height of a graph.
IMAGEMAP | NOIMAGEMAP | IMAGEMAP=ON | OFF
specifies whether data tips are generated.
IMAGENAME="filename"
specifies the base image filename.
LABELMAX=n
specifies the maximum number of labeled areas before data label
collision avoidance is disabled.
LABELPLACEMENT= GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in
the graphs.
LEGENDAREAMAX=n
specifies an integer that is interpreted as the maximum percentage of the
overall graphics area that a legend can occupy.
LINEPATTERNOBSMAX=n
specifies the maximum the number of vertices for a patterned line.
LOESSOBSMAX=n
specifies an upper limit for the number of observations that can be used
with a loess plot.
NBINSMAX=n
specifies the maximum number of bins that are processed for histograms.
NXYBINSMAX=n
 ODS GRAPHICS Statement 1563

specifies the maximum number of bins that are processed for heat maps.
OBSMAX=n
specifies the maximum number of observations that are processed.
OUTPUTFMT=file–type | STATIC
specifies the output format used to generate image or vector graphic
files.
PANELCELLMAX=n
specifies the maximum number of cells in a graph panel where the
number of cells is determined dynamically by classification variables.
PUSH
POP
pushes and pops ODS GRAPHICS settings in a stack.
RESET | RESET= option
Reset one or more ODS GRAPHICS options to its default.
SCALE | NOSCALE | SCALE=ON | OFF
specifies whether the content of a graph is scaled proportionally.
SCALEMARKERS | NOSCALEMARKERS | SCALEMARKERS=ON | OFF
specifies whether the plot markers are to be scaled with the graph size.
SHOW
writes the current ODS Graphics settings to the SAS log.
STACKDEPTHMAX=n
specifies the maximum stack depth for PUSH and POP requests.
SUBPIXEL | NOSUBPIXEL | SUBPIXEL=ON | OFF
specifies whether subpixel rendering should be used for rendering ODS
Graphics.
TIPMAX=n
specifies the maximum number of distinct data tip areas allowed before
data tips are disabled.
TOTALCELLMAX=n
specifies the maximum number of total cells in a graph where the number
of cells is determined dynamically by classification variables.
WIDTH=dimension
specifies the width of any graph.

Without Arguments
If ODS Graphics is currently disabled, then specifying the ODS GRAPHICS
statement without options enables it. If ODS Graphics is currently enabled, then
specifying the ODS GRAPHICS statement leaves it enabled.

Required Arguments
ON
enables ODS Graphics processing. This is the default if no argument is used.

Note: Beginning in SAS 9.4, ODS Graphics is enabled by default on all platforms
except z/OS.
1564 Chapter 24 / ODS GRAPHICS Statement

Alias YES

OFF
disables ODS Graphics processing.

Alias NO

Optional Arguments
ANTIALIAS | NOANTIALIAS | ANTIALIAS= ON | OFF
specifies whether anti-aliasing is applied to the rendering of the line and
markers in any graph. Anti-aliasing smooths the appearance of lines and some
markers. Text displayed in the graph is always anti-aliased. For graphical
displays that plot large numbers of points it is recommended that
ANTIALIAS=OFF be specified for performance considerations.

ANTIALIAS
smooths jagged edges of all components in the graph.

NOANTIALIAS
does not smooth jagged edges of components other than text in the graph.

ANTIALIAS=ON | OFF
specifies whether anti-aliasing is applied to the rendering of the line and
markers in the graph.

ON
smooths jagged edges of all components in the graph.

Alias YES

OFF
does not smooth jagged edges of components other than text in the
graph.

Alias NO

Default ANTIALIAS or ANTIALIAS=ON | YES

Restrictions If the number of markers or lines in the plot exceeds the number
specified by the ANTIALIASMAX= option, then the ANTIALIAS
option is disabled. This is true even if you specify the option
ANTIALIAS=ON or ANTIALIAS.

This option is ignored by the SGPIE and SGMAP procedures.

ANTIALIASMAX= n
specifies the maximum number of graphics elements before anti-aliasing is
disabled. For example, if there are more than 400 scatter point markers to be
anti-aliased and ANTIALIASMAX=400, then no markers are anti-aliased. The
default value is 4000.

Note: Prior to SAS 9.4M3, the ANTIALIASMAX= option specifies the maximum
number of observations in the graph data to be anti-aliased before anti-aliasing
 ODS GRAPHICS Statement 1565

is disabled. The default is 4000. When the graph data contains more than 4000
observations, anti-aliasing is disabled for the entire graph. Starting with SAS
9.4M3, the ANTIALIASMAX= option specifies the maximum number of graphics
elements to be anti-aliased in each plot on a per-plot basis. The default remains
at 4000. If any plot in a graph contains more than 4000 elements, anti-aliasing
is disabled for that plot. Anti-aliasing is enabled for the rest of the graph in that
case.

n
specifies a positive integer.

Default 4000

Restriction This option is ignored by the SGPIE and SGMAP procedures.

ATTRPRIORITY=COLOR | NONE
specifies a priority for cycling of the group attributes.

COLOR
assigns priority to the color attribute rotation by cycling through the list of
colors while holding the marker symbol and line pattern constant. When all
of the colors are exhausted, the marker symbol and line style attributes
increment to the next element, and then the colors in the list are repeated.
This pattern repeats as needed.

NONE
does not use an attribute priority in the rotation pattern, even if one is set in
the active style’s AttrPriority attribute. The rotation pattern cycles
progressively through the attribute lists.

Default The AttrPriority attribute of the graph style element, or NONE if the
current style does not define the AttrPriority style attribute.

Interaction The default lists of data colors, contrast colors, marker symbols, and
line patterns are set in the active style’s GraphData1–GraphDataN
elements.

Tip Use the ATTRPRIORITY=NONE option if you want groups to be

distinguished by color, marker, and line changes for all styles that
use color.

BORDER | NOBORDER | BORDER=ON | OFF

specifies whether to draw a border around each graph.

BORDER
draws a border around the graph.

NOBORDER
does not draw a border around the graph.

BORDER=ON | OFF
specifies whether to draw the graph with a border on the outermost layout.

ON
draws a border around the graph.
1566 Chapter 24 / ODS GRAPHICS Statement

Alias YES

OFF
does not draw a border around the graph.

Alias NO

Default BORDER or BORDER=ON | YES

BYLINE=NOBYLINE | TITLE | FOOTNOTE

specifies how the BY line is displayed in graphs. The option specifies how the
BY line is displayed when an analysis is run with a BY statement. By default, no
BY line is displayed.

The following code is an example of how the placement of the BY line is

controlled in most graph templates:
if (_BYTITLE_)
entrytitle _BYLINE_ / textattrs=GraphValueText;
else
if (_BYFOOTNOTE_)
entryfootnote halign=left _BYLINE_;
endif;
endif;

You can modify the graph template if you want to change how the BY line is
displayed. Because most graphs have titles and few graphs have footnotes, the
BY line looks better when it is displayed as a footnote. For complete
documentation about the Graph Template Language, see SAS Graph Template
Language: User’s Guide.

When the BYLINE= option is specified, and there are BY groups, ODS creates a
BY line and sets the appropriate special dynamic variables. The following table
lists the special dynamic variables for BY lines. For complete documentation
about special dynamic variables, see “Special Dynamic Variables” in SAS Graph
Template Language: User’s Guide.
Table A12.1 Special Dynamic Variables for BY Lines

_BYFOOTNOTE This variable is set to 1 when you specify a BY statement and

_ the ODS GRAPHICS BYLINE= option is set to FOOTNOTE.
Otherwise, the variable is set to 0 or is NULL.

_BYTITLE_ This variable is set to 1 when you specify a BY statement and

the ODS GRAPHICS BYLINE= option is set to TITLE.
Otherwise, the variable is set to 0 or is NULL.

_BYLINE_ This variable contains the text string that can be displayed as a
title or footnote.

The variables in the table are set automatically only for analytical procedures
that support ODS Graphics. For all other procedures, the variables are not set
 ODS GRAPHICS Statement 1567

automatically (NULL). To determine whether the procedure that you are using
supports ODS Graphics, refer to the procedure documentation.

NOBYLINE
specifies that no BY line is displayed. NOBYLINE is the default.

FOOTNOTE
specifies that the BY line is displayed as a left-justified graph footnote. This
is the recommended setting.

TITLE
specifies that the BY line is displayed as a centered graph title. Specifying
TITLE is not recommended because graphs are not designed to have
additional title lines.

Default NOBYLINE

Restriction This option does not work with the ODS Graphics procedures such
as SGPLOT and SGPANEL. To remove BY lines in those procedures,
use the NOBYLINE SAS system option. Example: options
nobyline;

DATASKINMAX=n
specifies the maximum number of graphical elements allowed per plot when a
data skin is applied.

Note: This feature applies to SAS 9.4M1 and later releases.

n
specifies a positive integer.

Default 200

DISCRETEMAX=n
specifies the maximum number of discrete values to be shown in any graph. Bar
charts and box plots are examples of affected plot types. Scatter plots and
other plot types can be affected if the data to be plotted is discrete or the axis is
discrete.

n
specifies a positive integer.

Default 1000

Tips Some plot layers might be unaffected by the DISCRETEMAX= option,

and those layers are rendered. If all layers are affected, a blank graph is
rendered.

If the value specified by the DISCRETEMAX= option is exceeded by any

plot layer in the graph, that layer is not drawn and a warning message is
issued. In that case, use the DISCRETEMAX= option to increase the
maximum number of discrete values that are allowed. Starting with
SAS 9.4M5, the log message includes a suggested value for
DISCRETEMAX=.
1568 Chapter 24 / ODS GRAPHICS Statement

DRILLTARGET="_blank" | "_self" | "_parent" | "_top" | "frame-name"

specifies the window that displays the drill-down output.

Note: This option is supported only for HTML.

"_blank"
opens a new browser window to display the drilldown output.

Default _blank is the default.

Requirements You must enclose _blank in quotation marks.

You must specify _blank in lowercase.

"_self"
opens the drill-down output in the same window.

Requirements You must enclose _self in quotation marks.

You must specify _self in lowercase.

"_parent"
opens the drill-down output in the parent frame.

Requirements You must enclose _parent in quotation marks.

You must specify _parent in lowercase.

"_top"
opens the drill-down output in the full body of the window.

Requirements You must enclose _top in quotation marks.

You must specify _top in lowercase.

"frame-name"
opens the drill down output in the named frame in the current window. If the
name does not exist, the output is opened in a new window.

Requirement You must enclose frame–name in quotation marks.

GROUPMAX=n
specifies the maximum number of group values to be shown in any graph. Any
graph that supports the GROUP= option is affected.

n
specifies a positive integer.

Default 1000

Note This option is ignored by the SGMAP procedure.

Tip If the value specified by the GROUPMAX= option is exceeded by any

plot layer in the graph, that layer is rendered. The system ignores the
GROUP= option and issues a warning message. In that case, use the
 ODS GRAPHICS Statement 1569

GROUPMAX= option to increase the maximum number of group values

that are allowed. Starting with SAS 9.4M5, the log message includes a
suggested value for GROUPMAX=.

HEIGHT=dimension
specifies the height of a graph.

dimension
is a nonnegative number followed by one of these units of measure:
Table A12.2 Units of Measure for Dimension

cm Centimeters

in Inches

mm Millimeters

pct or % Percentage

pt Point size (72 points = 1 inch)

px Pixels

Defaults The value of the SAS registry entry "ODS > ODS GRAPHICS >
Design Height" or the value of the DesignHeight= option in a
STATGRAPH template. Typically, the value is 480px.

For the PRINTER destination, units of 1/150 of an inch are used by

default.

Tips If only the HEIGHT= option is specified, then the default aspect ratio of
the graph is maintained.

When you change the size of a graph, by default graph elements such as
markers, lines, text, and titles are scaled proportionally. This scaling takes
effect even when you specify dimensions for the graph elements. To
disable the scaling, specify the NOSCALE option.

IMAGEMAP | NOIMAGEMAP | IMAGEMAP=ON | OFF

controls data tips and drill down generation. Data tips are pieces of explanatory
text that appear when you hold the mouse pointer over the data portions of a
graph contained in an HTML page.

IMAGEMAP
specifies to generate data tips.

NOIMAGEMAP
specifies not to generate data tips.

IMAGEMAP= ON | OFF
controls data tips generation.
1570 Chapter 24 / ODS GRAPHICS Statement

ON
specifies to generate data tips.

Alias YES

OFF
specifies not to generate data tips.

Alias NO

Default NOIMAGEMAP or IMAGEMAP=OFF | NO

Restrictions This option applies only when one of the ODS HTML* destinations
is used.

Prior to SAS 9.4M5, an image map is not generated using SVG with
ODS Graphics. The image map data that is used to produce tooltips
and links is written directly in the SVG and is not part of the HTML.
Using HTML5 with the inline SVG mode (the default value), the
tooltips and links are written in the SVG portion of the document.
(If you are using SAS 9.4M5, see the following Note.)

Interaction When IMAGEMAP | IMAGEMAP=ON is specified and the ODS

HTML destination is used, the IMAGE_DPI option in the ODS HTML
destination is ignored, if specified, and the default image resolution
of 96 DPI is used.

Note Starting with SAS 9.4M5, image maps are supported with SVG
output using HTML5. However, image maps are supported only
when the HTML5 SVG mode is INLINE (the default value).

IMAGENAME="filename"
specifies the base image filename. If more than one image is generated, each is
assigned a filename that consists of a base name followed by a number in order
to create unique names. This numbering can be reset with the RESET=INDEX
option. Path information (if needed) can be set with the GPATH= option on the
ODS destination statement. The default path is the current output directory. A
file extension for filename is automatically generated based on the
OUTPUTFMT= option.

Note: Starting with SAS 9.4M5, if a BY statement is in effect, you can uniquely
name image files based on BY groups. To do this, insert #BY text into the image
name. For more information, see “Substituting BY Values in the Image Name” on
page 1585.

Default The name of the output object.

Restrictions filename must be a single name. It must not include any path
specification or image-format name extension.

filename can contain only the following types of characters: letter,

digit, underscore, and space.
 ODS GRAPHICS Statement 1571

Requirement You must enclose filename in quotation marks.

See “Specifying and Resetting the Image Name” on page 1585

LABELMAX=n
specifies the maximum number of labeled areas before data label collision
avoidance is disabled. For example, if there are more than 50 points to be
labeled and LABELMAX=50, then collision avoidance is turned off and the
labels are all displayed at the top right of the data points.

n
specifies a positive integer.

Default 200

Restriction Data label collision avoidance is turned off under the following
conditions:
n The number of observations with
nonmissing labels exceeds the
value specified by LABELMAX=.
n The number of observations
exceeds five times the value
specified by LABELMAX=.
A message is then sent to the SAS log.

Tip To turn off collision avoidance, specify LABELMAX=0.

LABELPLACEMENT= GREEDY | SA
specifies the label-placement algorithm to use for positioning labels in the
graphs. The following labels are affected:
n data labels for needle plots, scatter plots, series plots, step plots, and vector
plots
n vertex labels for line charts

n curve labels when the curve label is positioned at the start or end of the
curve

GREEDY
specifies the Greedy method for managing label collision. The Greedy
method tries different placement combinations in order to find an optimal
approximation that avoids collisions. Label placement using this method is
often less optimal than label placement using the Simulated Annealing (SA)
method. However, depending on the number of data points and the potential
for label collisions, the Greedy process can be significantly faster.

SA
specifies the Simulated Annealing method for managing label collision. The
SA method attempts to determine the global minimization-of-cost function,
which is based on a simulated annealing algorithm. The resulting label
placement is usually better than placement using the Greedy method.
However, depending on the number of data points and the potential for label
collisions, the SA method can be significantly slower.
1572 Chapter 24 / ODS GRAPHICS Statement

Restriction For BANDPLOT and LINECHART, the SA method has no effect

on the curve labels when the CURVELABELPOSITION= option
specifies START or END.

Default GREEDY

LEGENDAREAMAX=n
specifies an integer that is interpreted as the maximum percentage of the
overall graphics area that a legend can occupy.

Note: This feature applies to SAS 9.4M3 and later releases.

LEGENDAREAMAX= replaces MAXLEGENDAREA=. However,
MAXLEGENDAREA= is supported as an alias. It is recommended that you use
LEGENDAREAMAX=.

n
specifies a positive integer.

Alias MAXLEGENDAREA=

Default 20

Range 0–100

Tip To turn off the legend, specify LEGENDAREAMAX=0. No warning is

issued when the legend is turned off in this way.

LINEPATTERNOBSMAX=n
specifies the maximum the number of vertices for a patterned line. If the
number of vertices exceeds the specified limit, the plot is not drawn.

n
specifies a positive integer.

Note: This feature applies to SAS 9.4M5 and later releases.

Default 10000

Restriction This option is ignored by the SGMAP and SGPIE procedures.

Note If the number of vertices exceeds the specified limit, the plot is not
drawn and a note is written to the SAS log. In that case, increase
LINEPATTERNOBSMAX= to the value suggested in the log message,
or change the line pattern to SOLID.

LOESSOBSMAX=n
specifies an upper limit for the number of observations that can be used with a
loess plot.
 ODS GRAPHICS Statement 1573

Note: This feature applies to SAS 9.4M3 and later releases. LOESSOBSMAX=
replaces LOESSMAXOBS=. However, LOESSMAXOBS= is supported as an alias.
It is recommended that you use LOESSOBSMAX=.

If the number of observations of the loess plot exceeds the specified limit, the
loess plot is not drawn.

For example, the following specifies that the most observations a loess plot can
have is 1000.
LOESSOBSMAX=1000

Alias LOESSMAXOBS=

Default 5000

NBINSMAX=n
specifies the maximum number of bins that are processed for histograms. This
option affects both computed and parameterized histograms, when available.

Note: This feature applies to SAS 9.4M4 and later releases.

n
specifies a positive integer.

Default 10000

Requirement The NBINSMAX= value must be greater than or equal to the

default number of bins in the histogram. If the NBINXMAX= value
is less than the default number of bins, an error occurs and the
histogram is not drawn. In that case, an error message is written to
the SAS log, which provides a minimum NBINSMAX= value that
you can use.

Interaction If you specify the number of bins in the histogram plot statement
NBINS= option, that option is honored if the NBINS= value
specified is less than or equal to the NBINXMAX= value. If the
NBINS= value specified exceeds the NBINSMAX= value, an error
occurs, and the histogram is not drawn.

NXYBINSMAX=n
specifies the maximum number of bins that are processed for heat maps. This
option affects both computed and parameterized heat maps. The option applies
to the product of the X and Y bins.

Note: This feature applies to SAS 9.4M4 and later releases.

n
specifies a positive integer.

Default 100000
1574 Chapter 24 / ODS GRAPHICS Statement

Interaction If you specify the number of bins in the heat map statement, that
option is honored regardless of the NXYBINSMAX= option.

OBSMAX=n
specifies the maximum number of observations that are processed. If the
number of observations in the data set exceeds the value specified for
OBSMAX=, the procedure step terminates with a log message.

Note: This feature applies to SAS 9.4M5 and later releases.

n
specifies a positive integer.

Alias MAXOBS=

Default 2 million observations

Interaction There are other ways to control the number of observations:

CASDATALIMIT= system option, DATALIMIT= option in the CAS
LIBNAME statement, and DATALIMIT= data set option. If the CAS
data transfer limit is set lower than OBSMAX=, then OBSMAX= has
no effect.

OUTPUTFMT=file–type | STATIC
specifies the output format used to generate image or vector graphic files. If the
image or vector graphic format is not valid for the active output destination, the
format is automatically changed to the default format for that destination.

file-type
is the image or vector graphic format to be generated. See “Supported File
Types for Output Destinations” on page 1590.

STATIC
uses the best quality static image format for the active output destination.
This is the default output format.

TIP The STATIC keyword can be used to reset the output format to
its default state.

Default STATIC

See “Specifying the Image Format” on page 1587

PANELCELLMAX=n
specifies the maximum number of cells in a graph panel where the number of
cells is determined dynamically by classification variables. If the number of cells
in the panel exceeds the specified limit, the panel is not drawn.

n
specifies a positive integer.
 ODS GRAPHICS Statement 1575

Default 1000

Note Graphs with DataPanel or DataLattice templates layouts are affected.

In the ODS Graphics procedures, this option affects graphs that are
created with the SGPANEL procedure. If the value specified by the
PANELCELLMAX= option is exceeded by any of these layouts, an
empty graph is rendered and a warning message is issued.

Tip You can use the TOTALCELLMAX= option to control the total
maximum number of cells in the graph.

PUSH | POP
pushes and pops ODS GRAPHICS settings in a stack. This feature enables you
to temporarily save your custom settings in a stack and later restore those
settings.

Note: This feature applies to SAS 9.4M3 and later releases.

PUSH
pushes the current ODS GRAPHICS settings to a stack.

POP
restores the most recently pushed settings from the stack. For each PUSH
action, you can specify a POP request. ODS issues a warning if you specify
POP without a corresponding PUSH. In that case, nothing is popped because
nothing has been pushed.

The pushed settings remain in the stack in the current SAS session until they are
popped or the stack is emptied.

Interaction You can specify PUSH as many times as you like up to the limit that
is defined by the STACKDEPTHMAX= option. You can also use
STACKDEPTHMAX= to empty the stack. For more information, see
“Managing the Stack Depth” on page 1594.

Note Order of specification is important when using the PUSH and POP
options. For more information, see “About PUSH and POP” on page
1593.

Tip Use the SHOW option to show the current ODS GRAPHICS settings.

See “Temporarily Saving and Restoring ODS GRAPHICS Settings” on

page 1593

RESET | RESET= option

Reset one or more ODS GRAPHICS options to its default.

RESET
resets all options to their defaults.

RESET=
resets one of the following to its default:

ALL
resets all reset-options to their defaults.
1576 Chapter 24 / ODS GRAPHICS Statement

ANTIALIAS
resets the ANTIALIAS= option to its default.

See ANTIALIAS= on page 1564

ANTIALIASMAX
resets the ANTIALIASMAX= option to its default.

See ANTIALIASMAX= on page 1564

ATTRPRIORITY
resets the ATTRPRIORITY= option to its default.

See ATTRPRIORITY= on page 1565

BORDER
resets the BORDER= option to its default.

See BORDER= on page 1565

BYLINE
resets the BYLINE= option to its default.

See BYLINE= on page 1566

DATASKINMAX
resets the DATASKINMAX= option to its default.

See DATASKINMAX= on page 1567

DISCRETEMAX
resets the DISCRETEMAX= option to its default.

See DISCRETEMAX= on page 1567

DRILLTARGET
resets the DRILLTARGET= option to its default.

See DRILLTARGET= on page 1568

GROUPMAX
resets the GROUPMAX= option to its default.

See GROUPMAX= on page 1568

HEIGHT
resets the HEIGHT= option to its default.

See HEIGHT= on page 1569

IMAGEMAP
resets the IMAGEMAP= option to its default.

Note Not all output destinations support this feature.

 ODS GRAPHICS Statement 1577

See IMAGEMAP= on page 1569

IMAGENAME
resets the IMAGENAME= option to its default.

Note: This feature applies to SAS 9.4M3 and later releases.

See IMAGENAME= on page 1570

INDEX <(positive-integer)>
resets the index counter that is appended to static image files.

Note: This feature applies to SAS 9.4M3 and later releases.

When specifying this option, you can also specify the value for the index
counter. The number that you specify must be enclosed in parentheses.
positive-integer determines the suffix for the next subsequent image, and
increments with each new image.

See “Resetting the Image Name” on page 1585

LABELMAX
resets the LABELMAX= option to its default.

See LABELMAX= on page 1571

LABELPLACEMENT
specifies the label-placement algorithm to use for positioning labels in
the graphs.

See LABELPLACEMENT= on page 1571

LEGENDAREAMAX
resets the LEGENDAREAMAX= option to its default.

See LEGENDAREAMAX= on page 1572

LOESSOBSMAX
resets the LOESSOBSMAX= option to its default.

See LOESSOBSMAX= on page 1572

NBINSMAX
resets the NBINSMAX= option to its default.

See NBINSMAX= on page 1573

NXYBINSMAX
resets the NXYBINSMAX= option to its default.

See NXYBINSMAX= on page 1573

1578 Chapter 24 / ODS GRAPHICS Statement

OUTPUTFMT
resets the OUTPUTFMT= option to its default.

Note: This feature applies to SAS 9.4M3 and later releases.

See OUTPUTFMT= on page 1574

PANELCELLMAX
resets the PANELCELLMAX= option to its default.

See PANELCELLMAX= on page 1574

SCALE
resets the SCALE= option to its default.

See SCALE= on page 1578

SCALEMARKERS
resets the SCALEMARKERS= option to its default.

See SCALEMARKERS= on page 1579

STACKDEPTHMAX
resets the STACKDEPTHMAX= option to its default.

Note: This feature applies to SAS 9.4M3 and later releases.

See STACKDEPTHMAX= on page 1581

SUBPIXEL
resets the SUBPIXEL option to its default.

Note: This feature applies to SAS 9.4M3 and later releases.

See SUBPIXEL on page 1581

TIPMAX
resets the TIPMAX= option to its default.

See TIPMAX = on page 1582

WIDTH
resets the WIDTH= option to its default.

See WIDTH= on page 1583

SCALE | NOSCALE | SCALE=ON | OFF

specifies whether the content of a graph is scaled proportionally.

SCALE
scales the components of a graph proportionally.
 ODS GRAPHICS Statement 1579

NOSCALE
does not scale the components of a graph proportionally.

SCALE=ON | OFF
specifies whether the content of a graph is scaled proportionally.

ON
scales the components of a graph proportionally.

Alias YES

OFF
does not scale the components of a graph proportionally.

Aliases NOSCALE

NO

Default SCALE or SCALE=ON | YES

SCALEMARKERS | NOSCALEMARKERS | SCALEMARKERS=ON | OFF

specifies whether the plot markers are to be scaled with the graph size. The
scaling factor is based on the height of the graph cells and the height of the
graph.

SCALEMARKERS
scales the markers with the graph size.

NOSCALE
does not scale the markers with the graph size.

SCALEMARKERS=ON | OFF
specifies whether the plot markers are to be scaled with the graph size.

ON
scales the markers with the graph size.

Alias YES

OFF
does not scale the markers with the graph size.

Aliases NOSCALE

NO

Default SCALEMARKERS or SCALEMARKERS=ON | YES

Restriction Scaling is done only if the graph contains multiple cells or single
nested cells.

SHOW
writes the current ODS Graphics settings to the SAS log. This option enables
you to verify which settings are in effect. The option is especially useful when
you use the PUSH and POP options to restore settings. For more information,
see “Temporarily Saving and Restoring ODS GRAPHICS Settings” on page 1593.
1580 Chapter 24 / ODS GRAPHICS Statement

Note: This feature applies to SAS 9.4M3 and later releases.

If no options have been specified, then SHOW lists those options for which ODS
currently knows the default values.

The following statement resets all settings and shows the default values.
ods graphics / reset=all show;

Here are the default values displayed in the SAS log:

ODS Graphics Settings
---------------------
Output format: STATIC
By line: NOBYLINE
Antialias: ON
Maximum Loess observations: 5000
Maximum stack depth: 1024
Stack depth: 0
MaxObs: 2000000
Maximum Histogram Bins: 10000
Maximum Heatmap Bins: 100000
Maximum Obs for Patterned Lines: 10000
Maximum Total Cells per BY-group: 2000

If you have specified the settings for one or more options, then SHOW includes
those settings along with the defaults.

Order of specification is important when using the SHOW option. For example,
the following statement shows the current settings and then sets the
NOBORDER option.
ods graphics / show noborder;

However, the following statement sets the NOBORDER option and then shows
the settings. The NOBORDER setting is shown in the log along with the other
settings that are in effect.
ods graphics / noborder show;

The following statement resets all settings. It then sets the image width and
shows the default settings along with the specified width.
ods graphics / reset=all width=5in show;

Here are the default values plus the image width, as displayed in the SAS log:
ODS Graphics Settings
---------------------
Output format: STATIC
By line: NOBYLINE
Antialias: ON
Maximum Loess observations: 5000
Image width: 5in
Maximum stack depth: 1024
Stack depth: 0
MaxObs: 2000000
Maximum Histogram Bins: 10000
Maximum Heatmap Bins: 100000
 ODS GRAPHICS Statement 1581

Maximum Total Cells per BY-group: 2000

Tip If you have specified the settings for some options but want to see the
default values without losing your specified settings, issue the following
two statements. The first statement pushes your specified settings, resets
all settings, and then lists options for which ODS currently knows the
default values. The second statement restores your previous settings.
ods graphics / push reset=all show;
ods graphics / pop;

STACKDEPTHMAX=n
specifies the maximum stack depth for PUSH and POP requests. The stack is
used to temporarily store ODS GRAPHICS settings when you issue PUSH
requests. PUSH saves the current settings to the stack and increments the stack
depth. POP restores the most recently saved settings from the stack and
decrements the stack depth.

Note: This feature applies to SAS 9.4M3 and later releases.

n
specifies a positive integer.

If n is less than the current stack depth, then the stack is popped until its
depth equals n. Popping the stack does not affect other option settings.

Defaults 1024 is the default maximum depth

0 is the default depth

Tips To empty the stack and then reset it to the default maximum depth,
issue the following statement:
ods graphics / stackdepthmax=0 reset=stackdepthmax;

You can use any of the following commands to reset the stack to its
default maximum depth:
reset=stackdepthmax
reset=all
reset
stackdepthmax=1024

See “Managing the Stack Depth” on page 1594

SUBPIXEL | NOSUBPIXEL | SUBPIXEL=ON | OFF

specifies whether subpixel rendering should be used for rendering ODS
Graphics. Subpixel rendering produces smoother curves and more precise bar
spacing.

Note: This feature applies to SAS 9.4M3 and later releases.

SUBPIXEL
always uses subpixel rendering, when applicable, for rendering lines and
bars.
1582 Chapter 24 / ODS GRAPHICS Statement

NOSUBPIXEL
never uses subpixel rendering.

SUBPIXEL=ON | OFF
specifies whether subpixel rendering should be used.

ON
always uses subpixel rendering, when applicable, for rendering lines and
bars.

Alias YES

OFF
never uses subpixel rendering.

Alias NO

Default Subpixel rendering is always enabled for vector-graphics output. It

is enabled by default for image output, unless the graph contains a
scatter plot or a scatter-plot matrix. In those cases, subpixel
rendering is disabled by default.

Restriction This option is ignored by the SGPIE and SGMAP procedures.

Requirement Antialiasing must be enabled for this option to have any effect.
Antialiasing is enabled by default. To re-enable antialiasing, use
the ANTIALIAS=ON option in the ODS GRAPHICS statement.

Tip For a large amount of data, antialiasing is disabled when the

number of observations exceeds the default maximum of 4000
observations. In that case, subpixel rendering is also disabled. To
increase the maximum, use the ANTIALIASMAX= option in the
ODS GRAPHICS statement.

See “Subpixel Rendering” in SAS ODS Graphics: Procedures Guide

TIPMAX=n
specifies the maximum number of distinct data tip areas allowed before data
tips are disabled. For example, if there are more than 400 points in a
scatterplot, and TIPMAX=400, then no data tips appear. The default maximum
value is 500.

Note: Prior to SAS 9.4M3, the TIPMAX= option specifies the maximum number
of observations in the graph data to be allowed before data tips are disabled.
The default is 500. When the graph data contains more than 500 observations,
data tips are disabled for the entire graph. Starting with SAS 9.4M3, the
TIPMAX= option specifies the maximum number of data-tip areas allowed
before data tips are disabled. This threshold is applied separately for each plot.
The default remains at 500. If any plot in a graph contains more than 500 data-
tip areas, data tips are disabled for that plot. Data tips are enabled for the
remaining plots in the graph.

n
specifies a positive integer.
 ODS GRAPHICS Statement 1583

Default 500

TOTALCELLMAX=n
specifies the maximum number of total cells in a graph where the number of
cells is determined dynamically by classification variables.

Note: This feature applies to SAS 9.4M5 and later releases.

If the number of cells exceeds the specified limit, the graph is not drawn. This
option is useful when multiple classification variables or large amounts of data
result in a large number of cells.

n
specifies a positive integer.

Default 2000

Notes If the number of cells exceeds the limit specified for

TOTALCELLMAX=, the graph is not drawn and an error message is
written to the SAS log. The error message includes a suggested value
for TOTALCELLMAX=. You can use the TOTALCELLMAX= option to
increase the maximum number of cells. Note, however, that the
processing time to render the graph increases with an increase in the
maximum number of cells.

Graphs with DataPanel or DataLattice templates layouts are affected.

In the ODS Graphics procedures, this option affects graphs that are
created with the SGPANEL procedure.

Tip You can use the PANELCELLMAX= option to control the maximum
number of cells in a graph panel.

WIDTH=dimension
specifies the width of any graph.

dimension
is a nonnegative number followed by one of these units of measure:
Table A12.3 Units of Measure for Dimension

cm Centimeters

in Inches

mm Millimeters

pct or % Percentage

pt Point size (72 points = 1 inch)

px Pixels
1584 Chapter 24 / ODS GRAPHICS Statement

Defaults The value of the SAS registry entry "ODS > ODS GRAPHICS >
Design Width" or the value of the DesignWidth= option in a
STATGRAPH template. Typically, this value is 640px.

For the PRINTER destination, units of 1/150 of an inch

Tips If only the WIDTH= option is specified, then the default aspect of the
graph is maintained.

When you change the size of a graph, by default graph elements such as
markers, lines, text, and titles are scaled proportionally. This scaling takes
effect even when you specify dimensions for the graph elements. To
disable the scaling, specify the NOSCALE option.

Details

Using the ODS GRAPHICS Statement

You can enable ODS Graphics by using one of the following equivalent statements:
ods graphics on;
ods graphics;

When you specify one of these statements before your procedure invocation, Base,
SAS/STAT, SAS/ETS, and SAS/QC procedures support ODS Graphics, either by
default, or when you specify procedure options for requesting particular graphs.

To disable ODS Graphics, specify the following statement:

ods graphics off;

Note: ODS Graphics is ON by default for these procedures: SGDESIGN, SGMAP,

SGPANEL, SGPIE, SGPLOT, SGRENDER, and SGSCATTER. For other products, the
initial state of ODS Graphics is determined by a SAS Registry setting.

Using the ODS GRAPHICS Statement for Batch Jobs

To generate ODS Graphics output in UNIX batch jobs, you must set the DISPLAY
system option before creating the output. To set the display, enter the following
command:
export DISPLAY=<ip_address>:0

The ip_address is the TCP/IP address, or the name of a UNIX terminal. Usually, the
IP address of the UNIX system where SAS is running would be used. If you do not
set the DISPLAY variable, then you get an error message in the SAS log.
 ODS GRAPHICS Statement 1585

Specifying and Resetting the Image Name

Specifying the Image Name

For ODS Graphics output, by default, the ODS object name is used as the “root”
name for the image output file. The following example creates a GIF image named
REGPLOT:
ods graphics / imagename="regplot" outputfmt=gif;

The assigned name REGPLOT is treated as a "root" name and the first output
created is named REGPLOT. Subsequent graphs are named REGPLOT1, REGPLOT2,
and so on, with an increasing index counter. This numbering can be reset with the
RESET=INDEX option.

Resetting the Image Name

The RESET=INDEX option enables you to reset the file name numbering sequence.
This feature applies to SAS 9.4M3 and later releases.

For a usage example, suppose that you are developing a paper or a presentation
and it takes several submissions to get the desired output. You can use the RESET
or RESET=INDEX option to force each output to replace itself:
ods graphics / reset=index ... ;

This specification causes all subsequent images to be created with the default or
current image name.

When specifying this option, you can also specify the value for the index counter.
The value that you specify determines the suffix for the next subsequent image.
For example:
ods graphics / reset=index(100) imagename="MyName";

The next graph that you produce is named MYNAME100.

This feature is useful for creating animated graphics. For example, for a sequence of
100 images, you might begin with the following statement:
ods graphics / reset=index(1) imagename="MyName";

In the example, your program produces 100 images named MYNAME1,

MYNAME2, ..., MYNAME100. If you later add more images to the animation, you
might submit the following:
ods graphics / reset=index(101) imagename="MyName";

The next generated image is named MYNAME101.

Substituting BY Values in the Image Name

Note: This feature applies to SAS 9.4M5 and later releases.

If a BY statement is in effect for the data, then you can uniquely name image files
based on BY groups. You do this by inserting any of the #BY* substitution items
into the image file name. You can specify BY variable values or BY variable names
1586 Chapter 24 / ODS GRAPHICS Statement

for IMAGENAME=. You can combine the #BY* substitution item with other text. To
do this, insert the #BY* item in the specified text string at the position where you
want the substitution text to appear.

As explained in the IMAGENAME= description, the file extension is automatically

generated based on the OUTPUTFMT= option.

Here are descriptions of the #BY* substitution items:

#BYVALn | #BYVAL(BY-variable-name)
substitutes the current value of the specified BY variable for #BYVAL in the file
name string. Specify the variable with one of these:

n
specifies a variable by its position in the BY statement. You must provide the
position even if your BY statement contains only one variable. Otherwise, no
substitution occurs.

BY-variable-name
specifies a variable from the BY statement by its name. BY-variable-name is
not case sensitive.

Examples The following example specifies the first variable in the BY

statement. The value of that variable is used for the image file name.
imagename="#byval1"

The following example specifies the second variable in the BY

statement preceded by the text sales_:
imagename="sales_#byval2"

The following example specifies the YEAR variable in the BY

statement. The value of that variable is used for the image file name:
imagename="#byval(year)"

The following example specifies the YEAR variable in the BY

statement followed with the text _sales:
imagename="#byval(year)_sales"

#BYVARn | #BYVAR(BY-variable-name)
substitutes the name of the BY variable or the label associated with the variable
(whatever the BY line would normally display) for #BYVAR in the file name
string. Specify the variable with one of these:

n
specifies a variable by its position in the BY statement. You must provide the
position even if your BY statement contains only one variable. Otherwise, no
substitution occurs.

BY-variable-name
specifies a variable from the BY statement by its name. BY-variable-name is
not case sensitive.

Examples The following example specifies the first variable in the BY

statement. The name of that variable is used for the image file name.
imagename="#byvar1"
 ODS GRAPHICS Statement 1587

The following example specifies the second variable in the BY

statement preceded by the text sales_:
imagename="sales_#byvar2"

The following example specifies the YEAR variable in the BY

statement. The name of that variable is used for the image file name:
imagename="#byvar(year)"

The following example specifies the YEAR variable in the BY

statement followed with the text _sales:
imagename="#byvar(year)_sales"

Note: When using this feature, consider any file name length restrictions that
might apply to your platform host. The file name length can vary greatly depending
on the input data.

Specifying the Image Format

Each ODS destination uses a default format for its output. You can use the
OUTPUTFMT= option in the ODS GRAPHICS statement to change the output
format.

Note: Unless you have a special requirement for changing the image format, we
recommend that you not change it. The default PNG or vector graphic format is far
superior to other formats, such as GIF, in support for transparency and a large
number of colors. Also, PNG and vector graphics images require much less disk
storage space than JPEG or TIFF formats.

If you want to generate vector graphics images, you can use the following
OUTPUTFMT= values for each destination:

Table A12.4 Generating Vector Graphics Output with ODS

ODS Destination OUTPUTFMT=value

ODS EPUB OUTPUTFMT=SVG

ODS destination for Excel OUTPUTFMT=EMF

ODS HTML OUTPUTFMT=SVG

ODS HTML5 Vector graphics images are generated by default

ODS LISTING OUTPUTFMT=EMF

OUTPUTFMT=PDF
OUTPUTFMT=PS | EPS | EPSI
OUTPUTFMT=SVG
1588 Chapter 24 / ODS GRAPHICS Statement

ODS Destination OUTPUTFMT=value

OUTPUTFMT=PCL

ODS PDF Vector graphics images are generated by default

ODS PCL OUTPUTFMT=PCL (for PCL output)

ODS PS OUTPUTFMT=PS (for PostScript output)

ODS destination for OUTPUTFMT=EMF

PowerPoint

ODS PRINTER OUTPUTFMT=PCL (for PCL output)

OUTPUTFMT=PDF (for PDF output)
OUTPUTFMT=PS | EPS | EPSI (for PostScript output)
OUTPUTFMT=SVG

ODS RTF OUTPUTFMT=EMF

ODS Measured RTF OUTPUTFMT=EMF

Note: For information about SVG and SAS Universal Printing, see “Creating SVG
Documents Using Universal Printing” in SAS 9.4 Universal Printing.

When a vector graphics image cannot be generated for the format that you specify,
a PNG image is generated instead and is embedded in the specified output file. The
output file format and extension are not changed in that case. In the following
cases, a vector graphics image cannot be generated:
n surface plots

n bivariate histograms

n graphs that use smooth gradient contours

n graphs that include continuous legends

n graphs that use data skins

n graphs that use transparency (EMF and PS ODS destinations only)

n graphs that contain one or more rotated images

n graphs that have a broken axis

n graphs that contain outline marker characters

Starting with SAS 9.4M2, additional cases for which vector graphics output cannot
be generated for graphs are as follows:
n graphs that use gradient fill for bars in a bar chart, histogram, or waterfall chart

n graphs that use the back-light effect on text

 ODS GRAPHICS Statement 1589

n graphs that include a text plot that displays text with an outlined bounding box
or text with a filled bounding-box background
n graphs that include images (PostScript output only)

Starting with SAS 9.4M3, vector graphics output can be generated in the EMF, PDF,
and SVG output formats for the following cases:
n graphs that use data skins

Note: For the EMF, PDF, and SVG formats, vector graphics output is not
supported for graphs that use transparency and data skins. An image is
generated in that case.

n graphs that include one or more rotated images

n graphs that use gradient fills (except PDF)

n graphs that use a continuous legend

Note: For the PDF output format, vector graphics output is not supported for
graphs that use a continuous legend and data transparency. An image is
generated in that case.

Starting with SAS 9.4M5, vector graphics output can or cannot be generated in the
output formats as indicated in the following table:

Table A12.5 Support for Vector Graphics

Case EMF PCL PDF PS SVG

Surface plots No No No No No

Gradient fill contour plots No No No No No

3-D bivariate histograms No No No No No

Continuous legends without data or Yes No Yes No Yes

fill transparency

Continuous legends with data Yes No No No Yes

transparency

Data skins Yes No Yes No Yes

Data skins with transparency No No No No Yes

Data transparency, without images Yes Yes Yes No Yes

Transparent images1 Yes No Yes No Yes

Images with data transparency No No No No Yes

Rotated images Yes No Yes No Yes

 1590 Chapter 24 / ODS GRAPHICS Statement

Case EMF PCL PDF PS SVG

Text plots with backlight Yes No Yes No Yes

Text plots with fill and outline Yes No Yes Yes Yes

Scatter plots with an outlined Yes No Yes No Yes

marker character

Broken axis Yes No Yes No Yes

Fill patterns Yes No No No Yes

Fill pattern legend chiclets No No No No No

1 Starting with SAS 9.4M7, vector graphics are supported for transparent images in EMF, PDF, and SVG output. The image
itself must be transparent. If you make an image transparent by specifying transparency in your program, a vector graphic
is not created.

Note: The SGMAP procedure, which is new in SAS 9.4M5, does not support vector-
based output.

Supported File Types for Output Destinations

The following table lists all of the supported file types for some ODS output
destinations.

Table A12.6 Supported File Types for Output Destinations

Output Destination Supported File Types

EPUB, EPUB2 PNG (default), GIF, JPG, SVG

EPUB3 SVG (default), PNG, GIF, JPG

Note: EPUB3 was added in SAS 9.4M1.
Note: Starting with SAS 9.4M3, EPUB3 is an alias for
EPUB, and the EPUB3 supported file types supersede
the EPUB supported file types.

ODS destination for Excel PNG (default), JPEG, JPG, EMF

HTML PNG (default), GIF, JPEG, JPG, SVG

HTML5 SVG (default), PNG, GIF, JPEG, JPG

LISTING PNG (default), BMP, EMF, EPSI, GIF, JFIF, JPEG, JPG,
PDF, PS, SASEMF, STATIC, TIFF, WMF, PSL, SVG

PDF Native PDF (default), JPEG, JPG, GIF, PNG

 ODS GRAPHICS Statement 1591

Output Destination Supported File Types

ODS destination for PNG (default), JPEG, JPG, GIF, EMF, TIFF, BMP
PowerPoint

PS PNG (default), JPEG, JPG, GIF, EPS, PDF, PCL, PS

RTF and Measured RTF EMF (default), PNG, JPEG, JPG, JFIF

Description of Supported File Types

The following table provides descriptions of the supported file types for ODS
output destinations.

Table A12.7 Description of Supported File Types

File Type Description

BMP (Microsoft Windows Device Supports color-mapped and true color

Independent Bitmap) images that are stored as uncompressed
or run-length encoded data. BMP was
developed by Microsoft Corporation.

CGM (Computer Graphics Metafile) A free and open international standard

file format for 2-D vector graphics, raster
graphics, and text. This format is defined
by ISO/IEC 8632.

DIB (Microsoft Windows Device See the description of BMP.

Independent Bitmap)

EMF ( Enhanced Metafile Format) Supports standard Enhanced Metafile

Format.

EMF Plus (Enhanced Metafile Format Supports Enhanced Metafile Plus

Plus Extensions) Extensions that provides additional
functionality, such as support of RGBA
colors.

EMF Dual (Enhanced Metafile Format Produces both EMF and EMF Plus
and Enhanced Metafile Format Plus formats simultaneously in the same
Extensions) output.

EPS Encapsulated PostScript

EPSI (Microsoft NT Enhanced Metafile) An extended version of the standard

PostScript (PS) format. Files that use
this format can be printed on PostScript
printers and can also be imported into
1592 Chapter 24 / ODS GRAPHICS Statement

File Type Description

other applications. Notice that EPSI files

can be read, but PS files cannot be read.

GIF (Graphics Interchange Format) Supports only color-mapped images. GIF

is owned by CompuServe, Inc.

JFIF (JPEG File Interchange Format) Supports JPEG image compression. JFIF
software is developed by the
Independent JPEG Group.

JPEG or JPG (Joint Photographic Experts A file format that is used for storing
Group) noninteractive images.

PBM (Portable Bitmap Utilities) Supports gray-scale, color, RGB, and

bitmap files. The Portable Bitmap
Utilities are a set of free utility programs
that were developed primarily by Jef
Poskanzer.

PCL Printer Control Language

PDF (Portable Document Format) A file format for electronic distribution

and exchange of documents.

PNG (Portable Network Graphic) Supports true color, gray-scale, and 8-bit
images.

PS (PostScript Image File Format) The Image classes use only PostScript
image operators. A level II PS printer is
required for color images. PostScript was
developed by Adobe Systems, Inc.

PSL (PostScript) PostScript

STATIC Chooses the best image format for the

current ODS destination.

SVG (Scalable Vector Graphics) Is an XML language for describing two-

dimensional vector graphics.

TIFF (Tagged Image File Format) Internally supports a number of

compression types and image types,
including bitmapped, color-mapped,
gray-scaled, and true color. TIFF was
developed by Aldus Corporation and
Microsoft Corporation and is used by a
wide variety of applications (available if
licensed).
 ODS GRAPHICS Statement 1593

File Type Description

XBM X Window Bitmap

XPM X Window Pixmap

Temporarily Saving and Restoring ODS GRAPHICS Settings

About PUSH and POP

Note: This feature applies to SAS 9.4M3 and later releases.

Although you can use the RESET option to restore the default ODS GRAPHICS
settings, there might be times when you want to save your current custom settings
and later restore them. ODS enables you to temporarily store your custom settings
in a stack created for this purpose, perform some other task with different settings,
and then restore the previous settings.

The PUSH option saves the current ODS GRAPHICS settings to the stack and
increments the stack depth. The POP option restores the most recently stored
settings from the stack and decrements the stack depth.

This feature is useful when you write macros. Within a macro you can PUSH at the
start of the macro and POP at the end. This enables your macro to have custom
ODS GRAPHICS behaviors without affecting the calling environment.

You can specify PUSH as many times as you like up to the limit that is defined by
the STACKDEPTHMAX= option. The pushed settings remain in the stack in the
current SAS session until they are popped or the stack is emptied. For more
information, see “Managing the Stack Depth” on page 1594. For each PUSH option,
you can specify a POP option. ODS issues a warning if you specify POP without a
corresponding PUSH. In that case, nothing is popped because nothing has been
pushed to the stack.

Order of specification is important when using the PUSH option. For example, the
following statement pushes the NOBORDER option to the stack along with any
other custom settings that are in effect.
ods graphics / noborder push;

A subsequent POP option restores the pushed settings including NOBORDER.

However, the following statement pushes the current custom settings and then
sets the NOBORDER option.
ods graphics / push noborder;

Here, the subsequent POP option restores whatever border setting was in effect
when the PUSH option was specified.
1594 Chapter 24 / ODS GRAPHICS Statement

TIP Use the SHOW option to show the ODS GRAPHICS settings that are
currently in effect.

Settings That Can Be Pushed

The PUSH and POP options apply to all ODS GRAPHICS options except the
following: PUSH, POP, RESET=INDEX, and SHOW.

How Code Errors Affect the PUSH Operation

If the ODS GRAPHICS statement contains a syntax error, then the PUSH option is
ignored.

For example, the PUSH option is ignored in the following statement:

ods graphics / antialias=bogus push;

A syntax error (BOGUS) in ANTIALIAS causes the parser to ignore the remaining
options. However, a simple semantics error does not prevent the remaining options
from being handled. In the following statement, the PUSH option is honored.
ods graphics / antialiasmax=-1 push;

In this statement, ANTIALIASMAX= –1 is invalid. The option expects a zero or a

positive integer. In this case, a warning is issued to the log, but the PUSH occurs.

Note: Syntax errors in your code can have other unexpected results that are not
described here.

Managing the Stack Depth

By default, the stack supports up to 1024 pushes. You can change the default by
using the STACKDEPTHMAX= option.

If the specified STACKDEPTHMAX= value is less than the current stack depth, then
the stack is popped until its depth equals the specified value. Popping the stack
does not affect other option settings.

If you want to empty the stack, issue the following statement:

ods graphics / stackdepthmax=0 reset=stackdepthmax;

This statement first empties the stack of all PUSH requests and then restores the
stack size to 1024.
 1595

Appendix 2
Reserved Keywords and Unicode
Values

Reserved Keywords and Unicode Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595
Lowercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1596
Uppercase Greek Letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597
Special Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1598

Reserved Keywords and Unicode Values

Overview
The tables in this section show some of the reserved keywords and Unicode values
that can be used with the UNICODE text command. For information about
rendering Unicode characters, see “Managing the String on Text Statements” in SAS
Graph Template Language: User’s Guide.

Note the following:

n Keywords and Unicode values are not case-sensitive: "03B1"x is the same code
point as "03b1"x.
n The word blank is the keyword for a blank space.
 1596 Appendix 2 / Reserved Keywords and Unicode Values

Lowercase Greek Letters

Keyword Glyph Unicode Description

alpha α 03B1 lowercase alpha

beta β 03B2 lowercase beta

gamma γ 03B3 lowercase gamma

delta δ 03B4 lowercase delta

epsilon ε 03B5 lowercase epsilon

zeta ζ 03B6 lowercase zeta

eta η 03B7 lowercase eta

theta θ 03B8 lowercase theta

iota ι 03B9 lowercase iota

kappa κ 03BA lowercase kappa

lambda λ 03BB lowercase lambda

mu μ 03BC lowercase mu

nu ν 03BD lowercase nu

xi ξ 03BE lowercase xi

omicron ο 03BF lowercase omicron

pi π 03C0 lowercase pi

rho ρ 03C1 lowercase rho

sigma σ 03C3 lowercase sigma

tau τ 03C4 lowercase tau

upsilon υ 03C5 lowercase upsilon

phi φ 03C6 lowercase phi

 Reserved Keywords and Unicode Values 1597

Keyword Glyph Unicode Description

chi χ 03C7 lowercase chi

psi ψ 03C8 lowercase psi

omega ω 03C9 lowercase omega

Uppercase Greek Letters

Keyword Glyph Unicode Description

alpha_u Α 0391 uppercase alpha

beta_u Β 0392 uppercase beta

gamma_u Γ 0393 uppercase gamma

delta_u Δ 0394 uppercase delta

epsilon_u Ε 0395 uppercase epsilon

zeta_u Ζ 0396 uppercase zeta

eta_u Η 0397 uppercase eta

theta_u Θ 0398 uppercase theta

iota_u Ι 0399 uppercase iota

kappa_u Κ 039A uppercase kappa

lambda_u Λ 039B uppercase lambda

mu_u Μ 039C uppercase mu

nu_u Ν 039D uppercase nu

xi_u Ξ 039E uppercase xi

omicron_u Ο 039F uppercase omicron

pi_u Π 03A0 uppercase pi

 1598 Appendix 2 / Reserved Keywords and Unicode Values

Keyword Glyph Unicode Description

rho_u Ρ 03A1 uppercase rho

sigma_u Σ 03A3 uppercase sigma

tau_u Τ 03A4 uppercase theta

upsilon_u Υ 03A5 uppercase upsilon

phi_u Φ 03A6 uppercase phi

chi_u Χ 03A7 uppercase chi

psi_u Ψ 03A8 uppercase psi

omega_u Ω 03A9 uppercase omega

Special Characters
Keyword Glyph Unicode Description

prime ´ 00B4 single prime sign

bar ̅ 0305 combining overline1

bar2 ̿ 033F combining double overline1

tilde ̃ 0303 combining tilde1

hat ̂ 0302 combining circumflex accent1

1 This is an overstriking character that requires a Unicode font to render properly.

 1599

Appendix 3
Display Attributes

General Syntax for Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1599

Attributes Available for the Attribute Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1600
Fill Color Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1600
Fill Pattern Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1601
Line Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1602
Marker Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1604
Text Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1606
Available Line Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607
Color-Naming Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1609
Overview of Color-Naming Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1609
RGB Color Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1610
CMYK Color Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1611
HLS Color Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1612
HSV (or HSB) Color Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1613
Gray-Scale Color Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1614
Color Naming System (CNS) Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1615
SAS Color Names and RGB Values in the SAS Registry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1616
Converting Color Values between Color-Naming Schemes . . . . . . . . . . . . . . . . . . . . . . . . 1617
Converting Numeric Color Component Values to Color Names . . . . . . . . . . . . . . . . . . . . 1617
Converting Color Values from Other Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1618

General Syntax for Attribute Options

Most statements provide options that enable you to specify attributes for the fills,
lines, data markers, or text that is used in the display. For example, many plots
provide a DATALABELATTRS= option that specifies the attributes of the data
labels. This appendix discusses the general syntax for those options and the valid
values for they accept.

A statement’s attribute options use the following general syntax:

 1600 Appendix 3 / Display Attributes

attribute-option-name = style-element | style-element (options) | (options)

style-element
specifies the name of a style element. Only style attributes that are relevant for
rendering the fill, line, data marker, or text are used.

Example fillattrs = graphdata1

style-element (options)
specifies the name of a style element, plus individual options to be used as style
overrides. Options is a space-delimited list of option = value pairs. Any options
that are not specified are derived from the specified style element.

Example fillattrs = graphdata1(color = blue)

(options)
specifies individual options as a space-delimited list of option = value pairs. Any
options that are not specified are derived from the default style element.

Example datalabelattrs = (family = "Arial" size = 10pt)

datalabelattrs = (family = graphvaluetext:fontfamily)

Depending on the attribute option used, the options might be fill options , line
options , marker options , or text options .

In general, any relevant attribute that is not specified defaults to an internal value,
which is typically derived from the style element that you specify for the attributes.
When choosing a style element, you should use an element of the correct type. See
“Graph Style Elements Used by ODS Graphics” in SAS Graph Template Language:
User’s Guide for a list of style elements and their types.

Attributes Available for the Attribute

Options
Depending on the attribute option used on a statement, the available attributes
might be fill-color options on page 1600, fill-pattern options on page 1601, line
options on page 1602, marker options on page 1604, or text options on page 1606.

Fill Color Options

Use one or more of the following options to specify fill-color attributes. The
options must be enclosed in parentheses and specified as a space-delimited list of
name = value pairs.

COLOR=style-reference | color
specifies the fill color. Style-reference must specify a valid style attribute such
as Color, ContrastColor, StartColor, NeutralColor, or EndColor in the form
 Attributes Available for the Attribute Options 1601

style-element-name:attribute-name. If style-reference is not defined in the

active ODS style, the COLOR= option is ignored and the default color is used.

Color must be a valid color name, such as RED, or a color code, such as
CXFF0000 or #FF0000. The color name must not exceed 64 characters. A
color code must be a valid code for a SAS color-naming scheme, such as RGB,
CMYK, HLS, or HSV (HSB). See “Color-Naming Schemes” on page 1609.

TRANSPARENCY=number
specifies the degree of the transparency of the filled area. This setting enables
you to set the transparency for the filled elements of some graph types. You can
set just this fill transparency, or set the fill independently of the other
transparent elements in the graph. For example, you can use this setting to set
the transparency level for the filled bars of a bar chart, and use the bar chart’s
DATATRANSPARENCY= option to set a different transparency level for the bar
outlines.

Default The same as the setting of the statement’s DATATRANSPARENCY=

option.

Range 0–1, where 0 is opaque and 1 is entirely transparent

Interaction This setting overrides the statement’s DATATRANSPARENCY=

setting for the fills but not for the outlines.

Example fillattrs = (transparency = 0.5)

Fill Pattern Options

Use one or more of the following options to specify fill-pattern attributes. The
options must be enclosed in parentheses and specified as a space-delimited list of
name = value pairs.

COLOR=style-reference | color
specifies the fill color.Style-reference must specify a valid style attribute such
as Color, ContrastColor, StartColor, NeutralColor, or EndColor in the form
style-element-name:attribute-name. If style-reference is not defined in the
active ODS style, the COLOR= option is ignored and the default color is used.

Color must be a valid color name, such as RED, or a color code, such as
CXFF0000 or #FF0000. The color name must not exceed 64 characters. A
color code must be a valid code for a SAS color-naming scheme, such as RGB,
CMYK, HLS, or HSV (HSB). See “Color-Naming Schemes” on page 1609.

PATTERN=style-reference | fill-pattern
specifies the fill pattern. Style-reference is valid starting with SAS 9.4M5. It
must specify a FILLPATTERN attribute in the form style-element-
name:FILLPATTERN. If style-reference is not defined in the active ODS style, the
PATTERN= option is ignored and the default fill pattern is used, if one is
specified by the active ODS style.
 1602 Appendix 3 / Display Attributes

Fill-pattern is a two-character specification that consists of a line-direction

prefix (R for right, L for left, and X for cross hatch) and a line-identification
number, 1–5. The following table shows the patterns for each of the possible
combinations.
Table A14.1 Fill Patterns

Pattern Pattern Pattern

Name Example Name Example Name Example

L1 R1 X1

L2 R2 X2

L3 R3 X3

L4 R4 X4

L5 R5 X5

Line Options
Use one or more of the following options to specify line attributes. The options
must be enclosed in parentheses and specified as a space-delimited list of name =
value pairs.

COLOR=style-reference | color
specifies the line color. Style-reference must specify a valid style attribute such
as Color, ContrastColor, StartColor, NeutralColor, or EndColor in the form
style-element-name:attribute-name. If style-reference is not defined in the
active ODS style, the COLOR= option is ignored and the default color is used.
Color must be a valid color name, such as RED, or a color code, such as
CXFF0000 or #FF0000. The color name must not exceed 64 characters. A
color code must be a valid code for a SAS color-naming scheme, such as RGB,
CMYK, HLS, or HSV (HSB). See “Color-Naming Schemes” on page 1609.

PATTERN=style-reference | line-pattern-name | line-pattern-number

specifies the line pattern. Style-reference must specify a LineStyle attribute in
the form style-element-name:LINESTYLE. Line patterns can be specified as a
pattern name or pattern number. The following table lists commonly used line
patterns.
 Attributes Available for the Attribute Options 1603

Table A14.2 Commonly Used Line Patterns

Pattern
Number Pattern Name Example

1 Solid

2 ShortDash

4 MediumDash

5 LongDash

8 MediumDashShortDash

14 DashDashDot

15 DashDotDot

20 Dash

26 LongDashShortDash

34 Dot

35 ThinDot

41 ShortDashDot

42 MediumDashDotDot

Valid line-pattern numbers range from 1 to 46. Not all line-pattern numbers
have names. See “Available Line Patterns” on page 1607 for a list of all possible
line patterns. We recommend that you use the named patterns because they
have been optimized to provide good discriminability when used in the same
plot.

Note Anti-aliasing might alter the appearance of some line patterns that have
fine detail such as line patterns 33 through 46. For example, if you specify
the color black and the pattern 33 for a line, and anti-aliasing is enabled,
then the line might appear gray. In that case, you can use the ODS
GRAPHICS statement to disable anti-aliasing as shown in the following
example:
ods graphics / antialias = off;

THICKNESS=style-reference | dimension
specifies the line thickness. Style-reference must specify a LINETHICKNESS
attribute in the form style-element-name:LINETHICKNESS.
 1604 Appendix 3 / Display Attributes

See “dimension” on page xiii

Marker Options
Use one or more of the following options to specify data-marker attributes. You
must enclose the options in parentheses and specify the options as a space-
delimited list of name = value pairs.

COLOR=style-reference | color
specifies the line color. Style-reference must specify a valid style attribute such
as Color, ContrastColor, StartColor, NeutralColor, or EndColor in the form
style-element-name:attribute-name. If style-reference is not defined in the
active ODS style, the COLOR= option is ignored and the default color is used.
Color must be a valid color name, such as RED, or a color code, such as
CXFF0000 or #FF0000. The color name must not exceed 64 characters. A
color code must be a valid code for a SAS color-naming scheme, such as RGB,
CMYK, HLS, or HSV (HSB). See “Color-Naming Schemes” on page 1609.

Restriction This option has no effect on marker symbols that are created with
the SYMBOLIMAGE statement.

See “SYMBOLIMAGE Statement” on page 1384

SIZE=style-reference | dimension
specifies the marker size (both width and height). Style-reference must specify a
MARKERSIZE attribute in the form style-element-name:MARKERSIZE.

See “dimension” on page xiii

SYMBOL=style-reference | marker-name
specifies the name of the marker. Style-reference must specify a
MARKERSYMBOL attribute in the form style-element-
name:MARKERSYMBOL.The following table lists the SAS symbols that are
supported.
Table A14.3 Supported Marker Symbols

Symbol Name Plot Symbol Symbol Name Plot Symbol

ArrowDown StarFilled

Asterisk Tack

Circle Tilde

CircleFilled Triangle
 Attributes Available for the Attribute Options 1605

Symbol Name Plot Symbol Symbol Name Plot Symbol

Diamond TriangleFilled

DiamondFilled TriangleDown

GreaterThan TriangleDownFille
d

Hash TriangleLeft

HomeDown TriangleLeftFilled

HomeDownFilled TriangleRight

IBeam TriangleRightFille
d

LessThan Union

Plus X

Square Y

SquareFilled Z

Star

You can also specify the names of symbols that are created by the
SYMBOLCHAR and SYMBOLIMAGE statements.

See “SYMBOLCHAR Statement” on page 1377

“SYMBOLIMAGE Statement” on page 1384

TRANSPARENCY=number
specifies the degree of transparency for the plot markers.

Default The transparency that is specified by the DATATRANSPARENCY=

option, which is 0 by default.

Range 0–1, where 0 is opaque and 1 is entirely transparent

Interaction This suboption overrides the DATATRANSPARENCY= option for the

plot markers only.
 1606 Appendix 3 / Display Attributes

WEIGHT=NORMAL | BOLD
specifies the marker weight.

Restriction This option has no effect on marker symbols that are created with
the SYMBOLCHAR and SYMBOLIMAGE statements.

See “SYMBOLCHAR Statement” on page 1377

“SYMBOLIMAGE Statement” on page 1384

Text Options
Use one or more of the following options to specify text attributes. The options
must be enclosed in parentheses and specified as a space-delimited list of name =
value pairs.

COLOR=style-reference | color
specifies the line color. Style-reference must specify a valid style attribute such
as Color, ContrastColor, StartColor, NeutralColor, or EndColor in the form
style-element-name:attribute-name. If style-reference is not defined in the
active ODS style, the COLOR= option is ignored and the default color is used.
Color must be a valid color name, such as RED, or a color code, such as
CXFF0000 or #FF0000. The color name must not exceed 64 characters. A
color code must be a valid code for a SAS color-naming scheme, such as RGB,
CMYK, HLS, or HSV (HSB). See “Color-Naming Schemes” on page 1609.

FAMILY=style-reference | "string"
specifies the font family of the text. Style-reference must specify a
FONTFAMILY attribute in the form style-element-name:FONTFAMILY.

SIZE=style-reference | dimension
specifies the font size of the text. Style-reference must specify a FONTSIZE
attribute in the form style-element-name:FONTSIZE.

Restriction The font size cannot be less than the minimum font size in SAS,
which is determined by the SAS system. If you specify a font size
that is less than the minimum font size, the minimum size is used.

See “dimension” on page xiii

STYLE=style-reference | NORMAL | ITALIC

specifies the font style of the text. Style-reference must specify a FONTSTYLE
attribute in the form style-element-name:FONTSTYLE.

WEIGHT=style-reference | NORMAL | BOLD

specifies the font weight of the text. Style-reference must specify a
FONTWEIGHT attribute in the form style-element-name:FONTWEIGHT.
 Available Line Patterns 1607

Available Line Patterns

The following line patterns can be used with the Graphics Template Language. A
line pattern can be specified by its number or name. Not all patterns have names.
We recommend that you use the named patterns because they have been
optimized to provide good discriminability when used in the same plot.

Table A14.4 Full List of Line Patterns

Pattern
Number Pattern Name Example

1 Solid

2 ShortDash

4 MediumDash

5 LongDash

8 MediumDashShortDash

10

11

12

13

14 DashDashDot

15 DashDotDot

16
1608 Appendix 3 / Display Attributes

Pattern
Number Pattern Name Example

17

18

19

20 Dash

21

22

23

24

25

26 LongDashShortDash

27

28

29

30

31

32

33

34 Dot

35 ThinDot

36

37

38

39
 Color-Naming Schemes 1609

Pattern
Number Pattern Name Example

40

41 ShortDashDot

42 MediumDashDotDot

43

44

45

46

Color-Naming Schemes

Overview of Color-Naming Schemes

The valid color-naming schemes in SAS are as follows:
n RGB (red green blue)

n CMYK (cyan magenta yellow black)

n HLS (hue lightness saturation)

n HSV (hue saturation brightness), also called HSB

n Gray scale

n SAS color names (from the SAS Registry)

n SAS Color Naming System (CNS)

Table A3.5 on page 1609 shows examples of each color-naming scheme.

Table A14.5 Examples of SAS Color-Naming Schemes

Color-Naming
Scheme Example

RGB COLOR=(CX98FB98)
 1610 Appendix 3 / Display Attributes

Color-Naming
Scheme Example

CMYK COLOR=(FF00FF00)
COLOR=(CMYK00FFFF00)

HLS COLOR=(H14055FF)

HSV COLOR=(V0F055FF)

Gray Scale COLOR=(GRAY4F)

SAS Registry Colors COLOR=(palegreen)

CNS Color Names COLOR=(VeryLightPurplishBlue)

COLOR=(Very_Light_Purplish_Blue)

You can also mix color-naming schemes in the same statement. For example:
DATACOLORS=(CXEE0044 vivid_blue darkgreen);

Each of the color-naming schemes has its advantages and disadvantages based on
how the output is used. For example, if you are creating a report that will be viewed
only online, then specifying colors using the RGB naming scheme or the SAS color
names defined in the registry might produce better results. If you are creating a
report for publishing in printed form, you might want to use the CMYK color-naming
scheme.

Note: If you specify an invalid color name, the default color is used instead.

For additional information about color-naming schemes, see Effective Color

Displays: Theory and Practice by David Travis, Computer Graphics: Principles and
Practice by Foley, van Dam, Feiner, and Hughes, and http://colorbrewer2.org.

RGB Color Codes

The RGB color-naming scheme is usually used to define colors for a display screen.
This color-naming scheme is based on the properties of light. An RGB color code
defines a color by combining red, green, and blue colors in different ratios. All the
colors combined together create white. The absence of all color creates black.

Color names are in the form CXrrggbb, where the following is true:
n CX indicates to SAS that this is an RGB color specification.

n rr is the red component.

n gg is the green component.

n bb is the blue component.

 Color-Naming Schemes 1611

The components are hexadecimal numbers in the range 00–FF (0% to 100%). Each
hexadecimal number indicates how much of the red, green, or blue is included in the
color. Lower percentage values are darker and higher values are lighter. This
scheme allows for up to 256 levels of each color component (more than 16 million
different colors).

Table A14.6 Examples of RGB Color Values

Color RGB Value

Red CXFF0000

Green CX00FF00

Blue CX0000FF

White CXFFFFFF

Black CX000000

Any combination of the color components is valid. Some combinations match the
colors produced by predefined SAS color names. See “Predefined Colors” in SAS
Graph Template Language: User’s Guide.

CMYK Color Codes

The CMYK color-naming scheme is used in four-color printing. CMYK is based on
the principles of objects reflecting light. Combining equal values of cyan, magenta,
and yellow produces process black, which might not appear as pure black. The
black component (K) of CMYK can be used to specify the level of blackness in the
output. A lack of all colors produces white when the output is printed on white
paper.

To specify the colors from a printer's Pantone Color Lookup Table, you can use the
CMYK color-naming scheme. Specify colors in terms of their cyan, magenta, yellow,
and black components. Color names are in the form CMYKccmmyykk or
Kccmmyykk, where the following is true:
n CMYK or K is an optional prefix that indicates to SAS that this is a CMYK color
specification.
n cc is the cyan component.

n mm is the magenta component.

n yy is the yellow component.

n kk is the black component.

The color-value components are hexadecimal numbers in the range 00–FF, where
higher values are darker and lower values are brighter. This scheme allows for up to
 1612 Appendix 3 / Display Attributes

256 levels of each color component. When the color value begins with a letter (A–
F), you can omit the CMYK or K prefix. When the color value begins with a number
(0–9), you must use the CMYK or K prefix.

Table A14.7 Examples of CMYK Color Values

Color CMYK Value

Red CMYK00FFFF00

Green FF00FF00

Blue CMYKFFFF0000

White K00000000

Process black (using cyan, magenta, and yellow FFFFFF00

ink)

Pure black (using only black ink) K000000FF

Note: You can specify a CMY value by specifying zero (00) for kk, the color's black
component.

CMYK color specifications are for output devices that support four colors. If you
specify CMYK colors on an output device that supports three colors such as RGB,
the CMYK colors are converted to colors in the three-color space. Because the
four-color space supports many more colors than a three-color space, the CMYK
colors might map to different colors in the three-color space. To preserve your
CMYK colors in that case, specify a device that supports the CMYK color space.
See “Color Support for Universal Printers” in SAS 9.4 Universal Printing.

HLS Color Codes

The HLS color scheme specifies colors in terms of hue, lightness, and saturation
levels. It is based on the Tektronix Color Standard. HLS color names are of the form
Hhhhllss, where the following is true:
n H indicates to SAS that this is an HLS color specification.

n hhh is the hue component, which is expressed as an angle.

n ll is the lightness component.

n ss is the saturation component.

The components are hexadecimal numbers. The hue component is in the range
000–168 hexadecimal, which represents an angular value in the range 0–360
decimal. Hue starts with the primary color blue at 0 degrees, and then progresses
 Color-Naming Schemes 1613

to red at 120 degrees, to green at 240 degrees, and back to blue at 360 degrees.
Both the lightness and saturation components are in the range 00–FF hexadecimal
(0–255 decimal), which provides 256 levels that represent 0% to 100% for each
component.

Table A14.8 Examples of HLS Color Codes

Color HLS Color Code

Red H07880FF

Green H0F080FF

Blue H00080FF

Light gray H000BB00

White1 HxxxFF00, such as H000FF00

Black1 Hxxx0000 such as H0000000

1 When the saturation is set to 00, the color is a shade of gray that is determined by the lightness
value. Therefore, white is defined as HxxxFF00 and black as Hxxx0000, where xxx can be any hue.

HSV (or HSB) Color Codes

The HSV color-naming scheme specifies colors in terms of hue, saturation, and
value (or brightness) components. HSV color names are of the form Vhhhssvv,
where the following is true:
n V indicates to SAS that this is an HSV color specification.

n hhh is the hue component, which is expressed as an angle.

n ss is the saturation component.

n vv is the value or brightness component.

The components are hexadecimal numbers. The hue component is in the range
000–168 hexadecimal, which represents an angular value in the range 0–360
decimal. Hue starts with the primary color red at 0 degrees, and then progresses to
green at 120 degrees, to blue at 240 degrees, and back to red at 360 degrees. Both
the saturation and value (brightness) components are in the range 00–FF
hexadecimal (0–255 decimal), which provides 256 levels for each component.

Table A14.9 Examples of HSV (or HSB) Color Codes

Color HSV Color Code

Red V000FFFF
 1614 Appendix 3 / Display Attributes

Color HSV Color Code

Green V078FFFF

Blue V0F0FFFF

Light gray1 Vxxx00BB such as V07900BB

White1 Vxxx00FF such as V07900FF

Black1 Vxxx00000 such as V0790000

1 When the saturation is set to 00, the color is a shade of gray. The value component determines the
intensity of gray level. The xxx value can be any hue.

Gray-Scale Color Codes

Gray-scale colors are specified using the word GRAY and a lightness value in the
form GRAYhh. The value hh is the lightness of the gray and is a hexadecimal
number in the range 00–FF, which provides 256 levels on the gray scale.

Note: GRAY, without a lightness value, is a SAS color name defined in the SAS
registry. (See “SAS Color Names and RGB Values in the SAS Registry” on page
1616.) Its value is CX808080. Invalid color specifications are mapped to GRAY.

The following figure shows the gray-scale color for each hexadecimal value.
 Color-Naming Schemes 1615

Color Naming System (CNS) Values

For CNS, you specify a color value by specifying lightness, saturation, and hue, in
that order, using the terms shown in the following table.

Table A14.10 Color Naming System Values

Lightness Saturation Hue

Black Gray Blue

Very Dark Grayish Purple

Dark Moderate Red

Medium Strong Orange/Brown

Light Vivid Yellow

Very Light Green

White

Follow these rules when you are determining the CNS color name:
n You should not use the lightness values Black and White with saturation or hue
values.
n If you do not specify default values, medium is the default lightness value and
vivid is the default saturation value.
n Gray is the only saturation value that can be used without a hue.

n Unless the color that you want is black, white, or some form of gray, you must
specify at least one hue.

You can use one or two hue values in the CNS color name. When you use two hue
values, the hues must be adjacent to each other in the following list: blue, purple,
red, orange/brown, yellow, green, and then returning to blue. Two hue values result
in a color that is a combination of both colors. Use the suffix “ish” to reduce the
effect of a hue when two hues are combined. Reddish purple is less red than red
purple. The color specified with the “ish” suffix must precede the color without the
“ish” suffix.

You can write color names in the following ways:

n without space separators between words

n with an underscore to separate words

 1616 Appendix 3 / Display Attributes

Do not enclose color names in quotation marks in GTL. For example, the following
color specifications are valid:
n verylightmoderatepurplishblue

n very_light_moderate_purplish_blue

Note: If a CNS color name is also a color name in the SAS Registry, the SAS
Registry color value takes precedence. Some CNS color names and color names in
the SAS Registry have different color values. Dark blue is an example. To use a CNS
color value when the color name is also in the SAS Registry, include an underscore
to separate the words. Here is an example.
color=dark_blue

SAS Color Names and RGB Values in the SAS

Registry
The SAS Registry provides a set of color names and RGB values that you can use to
specify colors. These color names and RGB values are common to most web
browsers. You can specify the name itself or the RGB value associated with that
color name. To view the color names as associated RGB values that are defined in
the registry, submit the following code:
proc registry list
startat="COLORNAMES";
run;

SAS prints the output in the SAS log. Here is a partial listing.
NOTE: Contents of SASHELP REGISTRY starting at subkey [COLORNAMES]
[ COLORNAMES]
Active="HTML"
[ HTML]
AliceBlue=hex: F0,F8,FF
AntiqueWhite=hex: FA,EB,D7
Aqua=hex: 00,FF,FF
Aquamarine=hex: 7F,FD,D4
Azure=hex: F0,FF,FF
Beige=hex: F5,F5,DC
...

For a list of the color names that are defined in the SAS Registry, see “Predefined
Colors” in SAS Graph Template Language: User’s Guide.

You can also create your own color values by adding them to the SAS registry. For
information about viewing and modifying the list of color names, see “The SAS
Registry” in SAS Programmer’s Guide: Essentials.
 Color-Naming Schemes 1617

Converting Color Values between Color-Naming

Schemes
If the SAS/GRAPH software is included in your SAS installation, several macros are
provided with SAS/GRAPH that enable you to perform selected conversions
between color-naming schemes. The following table shows the conversions that
are possible using these macros.

From Color Value To Color Value Conversion Macro

CMY RGB %CMY

CNS HLS %CNS

RGB HLS %RGB2HLS

HLS RGB %HLS2RGB

For information about these macros, see SAS/GRAPH: Reference.

Converting Numeric Color Component Values to

Color Names
If the SAS/GRAPH software is included in your SAS installation, several macros are
provided with SAS/GRAPH that enable you to convert numeric color component
values to color names. The following table shows the macros that you can use for
this purpose.

From Numeric Color Component Values To Color Name Conversion Macro

cyan, magenta, yellow, black CMYK %CMYK

hue, lightness, saturation HLS %HLS

hue, saturation, value HSV %HSV

red, green, blue RGB %RGB

 1618 Appendix 3 / Display Attributes

Converting Color Values from Other Applications

Many software programs enable you to change or customize various colors. A
dialog box typically provides a means of selecting a different color or modifying the
attributes of an existing color. In the SAS ODS Graphics Editor, for example, the
More Colors dialog box shown in the following figure serves this purpose.

Figure A14.1 More Colors Dialog Box

Figure A3.1 on page 1618 shows the HSB and RGB numeric color component values
for the currently selected color. It also shows the RGB values in hexadecimal. To
use the RGB specification for this color in a SAS program, add the CX prefix to the
Hex value E7B3B4. If the application provides only the numeric component values,
you must convert the decimal component values to hexadecimal. In Figure A3.1 on
page 1618, the H value is in the range 0–360 degrees, and the S and B values are
each in the range 0–100 percent. The R, G, and B values are in the range 0–255
each. You can convert the component values manually. (See “Understanding
Hexadecimal Values” in SAS Graph Template Language: User’s Guide.) Be aware that
you must first convert the S and B values from percentages to 255-based values by
rounding the result of the following computation to the nearest integer:
Value
100
× 255

If the SAS/GRAPH software is included in your SAS installation, you can also use
the SAS/GRAPH color utility macros to convert the values. The %HSV color utility
macro converts HSV (HSB) numeric color component values to an HSV color name.
Likewise, the %RGB macro converts RGB numeric color component values to an
RGB color name. The following example shows how to use the %HSV color utility
macro to convert the H, S, and B color component values in Figure A3.1 on page 1618
to an HSV color name.
 Color-Naming Schemes 1619

%COLORMAC;
data _null_;
put "%HSV(357,22,90)";
run;

Note: The %COLORMAC macro compiles all of the SAS/GRAPH color utility
macros. You need to run it only once during a SAS session.

Because the %HSV macro accepts values in the range 0–100 as a percentage of 255
for saturation and value, use the S and B values as shown. The result is V16538E6.

The following example shows how to use the %RGB color utility macro to convert
the R, G, and B numeric color component values in Figure A3.1 on page 1618 to an
RGB color name.
/* Compute the RGB percentages */
data _null_;
r = 231;
g = 179;
b = 180;
call symputx("r", round((r/255)*100));
call symputx("g", round((g/255)*100));
call symputx("b", round((b/255)*100));
run;

/* Convert to RGB color name */

%COLORMAC;
data _null_;
put "%RGB(&r,&g,&b)";
run;

Because the %RGB color utility macro accepts integer values in the range 0–100 as
a percentage of 255, you must convert the 255-based values to integer percentages
in order to use them in the %RGB macro call. The result is CXE8B3B5. The result
from the %RGB macro is not exact due to rounding. If you want more exact results,
use the following program.
data _null_;
r = 231;
g = 179;
b = 180;
rgb="CX" || put(r,hex2.) || put(g,hex2.) || put(b,hex2.);
put rgb;
run;

The result is CXE7B3B4.

To convert the RGB color name to an HLS color name, use the %RGB2HLS macro as
shown in the following program.
%COLORMAC;
data _null_;
put "%RGB2HLS(CXE7B3B4)";
run;

The result is H077CD84.

1620 Appendix 3 / Display Attributes

For more information about the SAS/GRAPH color utility macros, see SAS/GRAPH:
Reference.
 1621

Appendix 4
Format Support in ODS Graphics

Assigning Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1621

Using Locale-Sensitive Decimal Separators with Numeric Formats . . . . . . . . . . . . . . . . . . . 1622
SAS Formats That Are Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1622
Unsupported Numeric Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1622
Unsupported Date and Time Formats Related to ISO 8601 . . . . . . . . . . . . . . . . . . . . . . . . 1622
Other Unsupported Date and Time Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1623
Unsupported Currency Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1623
User-Defined Formats That Are Not Supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1624
Picture Formats with Date, Time, or Datetime Directives . . . . . . . . . . . . . . . . . . . . . . . . . 1624
Unicode Values with a User-Defined ODS Escape Character . . . . . . . . . . . . . . . . . . . . . . 1625

Assigning Formats
SAS formats and user-defined formats can be assigned to input data columns with
the FORMAT statement of the SGRENDER procedure. Also, several GTL statement
options accept a format as an option value. Examples include the
TICKVALUEFORMAT= option for formatting axis tick values, and the TIPFORMAT=
option for formatting data tips. For information about SAS formats, see SAS
Formats and Informats: Reference. For information about user-defined formats, see
“FORMAT Procedure” in Base SAS Procedures Guide.

Not all SAS formats and user-defined formats are supported in GTL or with the
SGPLOT, SGSCATTER, SGPANEL, and SGRENDER procedures. See “SAS Formats
That Are Not Supported” on page 1622. When an unsupported format is
encountered, a note similar to the following is written to the SAS log:

TICKVALUEFORMAT=bestx. is invalid. The format is invalid or unsupported.

The default will be used.
 1622 Appendix 4 / Format Support in ODS Graphics

Using Locale-Sensitive Decimal

Separators with Numeric Formats
System option NLDECSEPARATOR= is not supported by ODS Graphics. When
system option NLDECSEPARATOR is in effect, syntax errors or null debug errors
might result when using an ODS Graphics enabled procedure. For locale-sensitive
decimal separators in your graph’s numeric values, specify system option
NONLDECSEPARATOR, and then use appropriate NL formats to format your
numeric variables. See “Dictionary of Formats for NLS” in SAS National Language
Support (NLS): Reference Guide.

SAS Formats That Are Not Supported

Unsupported Numeric Formats

The following numeric formats are not supported in ODS Graphics:

BESTD BESTX D FLOAT FRACT

FREE IB IBR IEEE IEEER

ODDSR PCPIB PD PIB PIBR

PK RB SSN WORDF WORDS

Z ZD

Unsupported Date and Time Formats Related to

ISO 8601
The following date and time formats are not supported in ODS Graphics:
 SAS Formats That Are Not Supported 1623

$N8601B $N8601BA $N8601E $N8601EA $N8601EH

$N8601EX $N8601H $N8601X B8601DA B8601DN

B8601DT B8601DZ B8601LZ B8601TM B8601TZ

E8601DA E8601DN E8601DT E8601DZ E8601LZ

E8601TM E8601TZ IS8601DA IS8601DN IS8601DT

IS8601DZ IS8601LZ IS8601TM IS8601TZ

Other Unsupported Date and Time Formats

The following date and time formats are not supported in ODS Graphics:

HDATE HEBDATE JDATEMDW JDATEMNW JDATEWK

JDATEYDW JDATEYM JDATEYMD JDATEYMW JDATEYT

JDATEYTW JNENGO JNENGOT JNENGOTW JNENGOW

JTIMEH JTIMEHM JTIMEHMS JTIMEHW JTIMEMW

JTIMESW MDYAMPM MINGUO NENGO NLDATEYQ

NLDATEYR NLDATEYW NLDATMYQ NLDATMYR NLDATMYW

NLSTRMON NLSTRQTR NLSTRWK PDJULG PDJULI

TWMDY XYYMMDD YYQZ

Unsupported Currency Formats

The following currency formats are not supported in ODS Graphics:

EURFRATS EURFRBEF EURFRCHF EURFRCZK EURFRDEM

EURFRDKK EURFRESP EURFRFIM EURFRFRF EURFRGBP

EURFRGRD EURFRHUF EURFRIEP EURFRITL EURFRLUF

EURFRNLG EURFRNOK EURFRPLZ EURFRPTE EURFRROL

 1624 Appendix 4 / Format Support in ODS Graphics

EURFRRUR EURFRSEK EURFRSIT EURFRTRL EURFRYUD

EURTOATS EURTOBEF EURTOCHF EURTOCZK EURTODEM

EURTODKK EURTOESP EURTOFIM EURTOFRF EURTOGBP

EURTOGRD EURTOHUF EURTOIEP EURTOITL EURTOLUF

EURTONLG EURTONOK EURTOPLZ EURTOPTE EURTOROL

EURTORUR EURTOSEK EURTOSIT EURTOTRL EURTOYUD

User-Defined Formats That Are Not

Supported

Picture Formats with Date, Time, or Datetime

Directives
ODS Graphics does not support user-defined picture formats that specify date,
time, or datetime directives such as %a, %W, %H, and so on. Using a picture format
with directives produces unexpected results in ODS Graphics. To work around this
limitation, apply the picture format to a new column in your data, and then plot the
new preformatted column instead. Here is an example that creates column
NewDate by formatting the Date column as MM/YYYY.
/* Create picture format MonYear for MM/YYYY date */
proc format;
picture monyear (min=7)
low-high='%0m/%Y' (datatype=date);
run;

/* Add preformatted date column NewDate to the data */

data stocks;
set sashelp.stocks
(where=(year(date) eq 2001 and stock eq "IBM"));
newdate = put(date, monyear.);
label newdate="Date: MM/YYYY";
run;

The NewDate column can be used in ODS Graphics plot statements instead of the
Date column.

For information about the PICTURE statement in the FORMAT procedure, see Base
SAS Procedures Guide.
 User-Defined Formats That Are Not Supported 1625

Unicode Values with a User-Defined ODS Escape

Character
ODS Graphics supports Unicode values in user-defined formats only if they are
preceded by the (*ESC*) escape sequence as shown in the following example.
"(*ESC*){unicode beta}"

ODS Graphics does not support the use of a user-defined ODS escape character to
escape Unicode values in user-defined formats.

For an example of how to use Unicode values in user-defined formats with ODS
Graphics, see “Formatting the Tick Values on a Discrete Axis” in SAS Graph
Template Language: User’s Guide.
1626 Appendix 4 / Format Support in ODS Graphics
 1627

Appendix 5
Generalized Macro for
BOXPLOTPARM Data

Generalized Macro for BOXPLOTPARM Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1627

Generalized Macro for BOXPLOTPARM

Data
The following SAS code is a generalized macro for computing input data for
BOXPLOTPARM.
%macro boxcompute(indsn=,x=,y=,outdsn=boxdata,datalabel=,
qntldef=5,table=no);
/* NOTE: INDSN, X and Y are required parameters, where
INDSN = input SAS data set
X = categorical variable (num or char)
Y = response variable (num)
OUTDSN = output dataset. It contains these variables:
STAT: Statistic names for BOXPLOTPARM
VALUE: values for STAT type
X: X variable values
DATALABEL: outlier labels from the DATALABEL= variable
N, Mean, Median, Std if TABLE=YES
DATALABEL= variable used to label outliers (num or char)
QNTLDEF = 1|2|3|4|5
(how to compute quantiles - see PROC SUMMARY)
TABLE = YES | NO
(add additional data to build table of statistics)
*/
%macro varinfo(dsid,varname,role,rc);
/* utility macro for obtaining variable info */
1628 Appendix 5 / Generalized Macro for BOXPLOTPARM Data

%local varnum;
%if %length(&varname)=0 %then %do;
%let &rc=0; %return;
%end;
%let varnum=%sysfunc(varnum(&dsid,&varname));
%if &varnum > 0 %then %do;
%let &role.label=%sysfunc(varlabel(&dsid,&varnum));
%if %length(&&&role.label)=0 %then
%let &role.label=%sysfunc(varname(&dsid,&varnum));
%let &role.fmt=%sysfunc(varfmt(&dsid,&varnum));
%let &rc=0;
%end;
%else %do;
%put ERROR: %upcase(&role) variable &varname not found.;
%let &rc=1;
%end;
%mend varinfo;

/* validate dataset and variables */

%local dsid ylabel xlabel datalabellabel
yfmt xfmt datalabelfmt rc_y rc_x rc_d;
%let dsid=%sysfunc(open(&indsn));
%if &dsid %then %do;
%varinfo(&dsid,&y,Y,rc_y)
%varinfo(&dsid,&x,X,rc_x)
%if %length(&datalabel) %then
%varinfo(&dsid,&datalabel,DATALABEL,rc_d);
%else %let rc_d=0;
%let dsid=%sysfunc(close(&dsid));
%if &rc_y or &rc_x or &rc_d %then %return;
%end;
%else %do;
%put ERROR: Input dataset &indsn not found.;
%return;
%end;

/* compute basic summary statistics */

proc summary data=&indsn(rename=(&y=VALUE &x=X))
nway qntldef=&qntldef;
class x;
var value;
output out=summary(drop=_type_ _freq_) n=N mean=Mean
median=Median q1=Q1 q3=Q3 std=STD / noinherit;
run;
proc sort data=&indsn(keep=&x &y &datalabel)
%if %length(&datalabel) %then
out=sorted(rename=(&x=X &y=VALUE &datalabel=DATALABEL));
%else out=sorted(rename=(&x=X &y=VALUE));
; by &x;
run;
/* compute fences, MIN, MAX and any outliers for X values */
data outliers;
length STAT $10;
%if %length(&datalabel) %then
%do;
keep STAT X VALUE DATALABEL;
 Generalized Macro for BOXPLOTPARM Data 1629

label VALUE="&ylabel" X="&xlabel"

DATALABEL="&datalabellabel";
format VALUE &yfmt X &xfmt DATALABEL &datalabelfmt;
%end;
%else
%do;
keep STAT X VALUE;
label VALUE="&ylabel" X="&xlabel";
format VALUE &yfmt X &xfmt;
%end;
retain lowerFence upperFence farLowerFence farUpperFence
tempmin tempmax;
merge sorted summary; by x;

/* perform computations for each X value */

if first.X then do;
lowerFence=q1-((q3-q1)*1.5);
upperFence=q3+((q3-q1)*1.5);
farLowerFence=q1-((q3-q1)*3);
farUpperFence=q3+((q3-q1)*3);
/* these computations for MIN and MAX result
in the same values produced by the BOXPLOT
statement, however they can be modified to
satify other statistical definitions */
if value <= upperFence then tempmax=value;
else tempmax=.;
if value => lowerFence then tempmin=value;
else tempmin=.;
end;
/* recompute MAX and MIN for each obs */
if 0 <= sum(upperFence,-value) then
tempmax=max(tempmax,value);
if 0 <= sum(value,-lowerFence) then
tempmin=min(tempmin,value);
/* write out both types of outliers */
if value < farLowerFence or value > farUpperFence then do;
stat="FAROUTLIER"; output;
end;
else if value < LowerFence or value > UpperFence then do;
stat="OUTLIER"; output;
end;
/* write out MIN and MAX for each X value */
if last.X then do;
value=tempmin; stat="MIN"; output;
value=tempmax; stat="MAX"; output;
end;
run;
/* transpose the stats into the columns
required by BOXPLOTPARM */
data transpose(keep=x stat value);
length STAT $10;
set summary;
array stats{*} n--std;
do i=1 to dim(stats);
stat=upcase(vname(stats{i}));
VALUE=stats{i};
1630 Appendix 5 / Generalized Macro for BOXPLOTPARM Data

output;
end;
run;
/* interleave the obs by the X variable */
data &outdsn;
set transpose outliers; by X;
run;
/* merge the output stats for building a stat table */
%if %upcase(&table)=YES %then %do;
data &outdsn;
merge &outdsn summary; by X;
run;
%end;
%mend boxcompute;

Here is the macro invocation to produce the data for the graph shown in the section
“Example: BOXPLOTPARM Statement” on page 440 .
%boxcompute(indsn=sashelp.cars,x=type,y=mpg_city,
datalabel=make);

proc template;
define statgraph boxplotparm1;
begingraph;
entrytitle "City Mileage for Vehicle Types";
layout overlay;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true ;
endlayout;
endgraph;
end;
run;

proc sgrender data=boxdata template=boxplotparm1;

run;

The following figure and code show an example of a “table” of statistics with
BLOCKPLOT statements.
 Generalized Macro for BOXPLOTPARM Data 1631

%boxcompute(indsn=sashelp.cars,x=type,y=mpg_highway,
outdsn=boxdata2,table=yes);

proc template;
define statgraph boxplotparm2;
begingraph;
entrytitle "Highway Mileage for Vehicle Types";
layout overlay / xaxisOpts=(offsetMin=0.08 offsetMax=0.08);
innerMargin / align=top;
blockplot x=x block=n /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=std /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=mean /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
blockplot x=x block=median /
display=(values label outline) valuehalign=center
labelattrs=graphdatatext valueattrs=graphdatatext;
endInnerMargin;
boxplotparm y=value x=x stat=stat /
datalabel=datalabel spread=true;
endlayout;
endgraph;
end;
run;

proc sgrender data=boxdata2 template=boxplotparm2;

run;
1632 Appendix 5 / Generalized Macro for BOXPLOTPARM Data
 1633

Appendix 6
Memory Management for ODS
Graphics

SAS Options Affecting Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1633

Managing a Java Out of Memory Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1634

SAS Options Affecting Memory

ODS Graphics uses Java technology to produce its graphs. Most of the time this
fact is transparent to you because the required Java Runtime Environment (JRE)
and JAR files are included with SAS software installation. Also, the Java
environment is automatically started and stopped for you. When Java is started, it
allocates a fixed amount of memory. The memory can grow up to the value set for
the -Xmx suboption in the JREOPTIONS option (discussed in a moment). This
memory is independent of the memory limit that SAS sets for the SAS session with
its MEMSIZE= option.

Normally, the memory limit for Java is sufficient for most ODS Graphics
applications. However, some tasks are very memory intensive and might exhaust all
available Java memory, resulting in an OutOfMemoryError condition. You might
encounter Java memory limitations in the following cases:
n the product of the output size and the DPI setting results in very large output

n a classification panel has a very large number of classifier crossings

n a scatter plot matrix has a large number of variables

n creating 3-D plots and 2-D contours, which are memory intensive to generate

n a plot has a very large number of marker labels

n a plot uses many character variables or has a large number of GROUP values

n using the SG Editor to edit a graph with a large amount of data

 1634 Appendix 6 / Memory Management for ODS Graphics

Managing a Java Out of Memory Error

If you encounter a Java OutOfMemoryError, then you can try executing your
program again by restarting SAS and specifying a larger amount of memory for Java
at SAS invocation.

To determine what the current Java memory settings are, you can submit a PROC
OPTIONS statement that shows the value of the JREOPTIONS option:

proc options option=jreoptions;

run;

After you submit this procedure code, a list of JREOPTIONS settings is written to
the SAS log. The JREOPTIONS option has many suboptions that configure the SAS
Java environment. Many of the suboptions are installation and host specific and
should not be modified, especially the ones that provide installed file locations. For
managing memory, look for the -Xmx and -Xms suboptions:

JREOPTIONS=(/* other Java suboptions */ -Xmx128m -Xms128m)

-Xms
Use this option to set the minimum Java memory (heap) size, in bytes. Set this
value to a multiple of 1024 greater than 1MB. Append the letter k or K to
indicate kilobytes, or m or M to indicate megabytes. The default is 2MB.
Examples:

-Xms6291456
-Xms6144k
-Xms6m

-Xmx
Use this option to set the maximum size, in bytes, of the memory allocation
pool. Set this value to a multiple of 1024 greater than 2MB. Append the letter k
or K to indicate kilobytes, or m or M to indicate megabytes. The default is 64MB.
Examples:

-Xmx83886080
-Xmx81920k
-Xmx80m

As a general rule, you should set the minimum heap size (-Xms) equal to the
maximum heap size (-Xmx) to minimize garbage collections.

Typically, SAS sets both -Xms and -Xmx to be about 1/4 of the total available
memory or a maximum of 128M. However, you can set a more aggressive maximum
memory (heap) size, but it should never be more than 1/2 of physical memory.
 Managing a Java Out of Memory Error 1635

You should be aware of the maximum amount of physical memory your computer
has available. Let us assume that doubling the Java memory allocation is feasible.
So when you start SAS from a system prompt, you can add the following option:

-jreoptions (-Xmx256m -Xms256m)

Alternatively, you might need to specify the setting in quotation marks:

-jreoptions '(-Xmx256m -Xms256m)'

The exact syntax varies for specifying Java options, depending on your operating
system, and the amount of memory that you can allocate varies from system to
system. The set of JRE options must be enclosed in parentheses. If you specify
multiple JREOPTIONS system options, then SAS appends JRE options to JRE
options that are currently defined. Incorrect JRE options are ignored.

If you choose to create a custom configuration file, then you would simply replace
the existing -Xms and -Xmx suboption values in the JREOPTIONS=(all Java
options) portion of the configuration file.

For more information, see the SAS Companion for your operating system.
1636 Appendix 6 / Memory Management for ODS Graphics