SQL Server Coding Standards GF-IS ADM Applications Performance Excellence

(APE)

—
SOFTWARE ENGINEERING STANDARDS & PRACTICES

SQL Server Coding Standards

9AAD134842

Department GF-IS ADM Applications Performance Excellence (APE)

Approver Giulio Bitella, Global Department Manager
Owner Tomasz Jastrzębski, Global Leader for Software Engineering (SE)

For the latest distributable version of this and other Software Engineering standards please
visit this link to ABB Library.
SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

WHAT IS THIS?
This document contains SQL Server programming standards.

WHAT IS THE PURPOSE OF THIS DOCUMENT?

The purpose of the document is to establish basic SQL Server programming stand-
ards, practices and guidelines.

i © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

—
TABLE OF CONTENTS

WHAT IS THIS? ............................................................................................................................... I

WHAT IS THE PURPOSE OF THIS DOCUMENT? ............................................................................ I
DATABASE DESIGN ....................................................................................................................... 1
SECURITY ...................................................................................................................................... 1
DATA TYPES .................................................................................................................................. 2
NULL VALUES ................................................................................................................................ 2
INDEXES......................................................................................................................................... 2
NAMING CONVENTIONS ............................................................................................................... 2
CODING STYLE .............................................................................................................................. 3
DATABASE ACCESS ....................................................................................................................... 4
SAMPLE CODE - VIEW ................................................................................................................... 4
SAMPLE CODE – STORED PROCEDURE........................................................................................ 5
REFERENCES ................................................................................................................................. 6
REVISION HISTORY ....................................................................................................................... 6

ii © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

DATABASE DESIGN
1. SQL Server is highly sophisticated product and should be used to its full po-
tential, rather just as a simple table store.
2. Pay attention to proper database normalization. OLTP database should sat-
isfy at least Third Normal Form (3NF) – see references for details.
3. Do not mix OLTP and OLAP tables in the same database schema.
4. SQL Server is multi user server, always consider other users may try to mod-
ify the same data the same time and use transactions whenever appropriate.
Within transaction lock as little data as possible, particularly in long-lasting
transactions using large datasets where transaction can block multiple other
transactions. Use appropriate transaction Isolation Level and Lock Level (e.g.
tablock vs. rowlock, vs. nolock).
(ref. Table Hints, Transaction Isolation Levels)
5. Make sure transactions follow ACID (Atomicity, Consistency, Isolation, Dura-
bility) principle.
6. Consider using stored procedures for time sensitive routines which would
otherwise require numerous database call and round trips. Use of temporary,
especially in-memory tables within stored procedures is encouraged.
7. To ensure data integrity use table relations (foreign keys) whenever appropri-
ate. Whenever feasible, use the same column name as primary and foreign
key column. E.g. Customers.CustomerId ↔ Orders.CustomerId.
8. Consider using table/view triggers and relationships (foreign keys) with
“on delete/update cascade” option. Such approach greatly reduces number
of required database calls and, albeit particular database operation takes
longer, overall system performance is increased.
9. Building triggers always bear in mind trigger is fired just once for multiple in-
serted, deleted or updated records.

SECURITY
1. Grant database object permissions to roles, rather than particular users. De-
fine database roles matching access needs. In simple cases use bulit-in roles
like: db_datareader, db_datawriter, db_denydatawriter.
2. Always follow “the least privilege principle”. Under no circumstances applica-
tion or its middle layer should access database as administrator or database
owner (db_owner).
3. Define database users using “contained user” model without server login for
better database portability, rather than traditional model.

1 © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

4. Avoid using passwords for authentication. Use Managed Service Iden-

tity/Principal and Azure AD (for Azure PaaS) or certificates and Azure AD (for
IaaS).

DATA TYPES
1. Data types used in data store should match types used in higher application
layers as closely as possible.
2. Avoid using varchar(max), nvarchar(max) and similar without real need to
store very large amounts of data.
3. Avoid using Unicode nvarchar and nchar types where ASCII varchar and char
types are sufficient.
4. Do not use datetime type. Use datetime2() instead, typically datetime2(0)
– no second fraction.
5. Storing date and datetime2 use UTC date/time. If timezone information is
required use datetimeoffset type. If column always contains date/datetime
other than UTC add timezone code as column name suffix. Example: name
field containing CET timezone DateCreatedCet.
6. Avoid using float point numerical types (real, float) unless stored values re-
ally are float point. Use decimal type instead. Always use decimal type to
store cost, price, quantity and other values of defined/limited precision.

NULL VALUES
1. Allow for NULL values in columns only when NULL values are required.
2. Typically in text, ntext, char, nchar, varchar, nvarchar columns NULL values
are not required and empty string can be used as a default value.

INDEXES
1. Define table indexes wisely as they slow down insert/update operations.
2. Field sequence in indexes does matter. Typically, the most specific fields
should be specified first. Example: LastName, FirstName (in this order).
3. Make informed decision whether to use the clustered index as table primary
key PK. It is often the best choice, however in cases of e.g. sequential range
scans, keeping often scanned ranges in clusters may be by far the best
choice performance-wise.
Example: OrderDetails records are usually read by a given order. Therefore it
may be a good idea to create clustered, non-unique index on OrderId column,
rather than OrderDetailId (PK).

NAMING CONVENTIONS
1. Use lowercase for all keywords to provide similar coding experience for full-
stack developers.
2 © ABB Asea Brown Boveri 2018-2019
SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

2. Literals (table names, column names etc., stored procedure and function
public parameters) should follow Pascal casing, local procedure/function pa-
rameters: camel casing. It is especially important for DB-first approach that
literals follow C# Coding Standards. (ref. 9AAD134036)
3. Use plural noun form for table names. E.g. Sessions.
4. Record ID field (usually primary key) should be named “Id”.
5. Index names should consist of:
a. IX_ or PK_ prefix for primary key indexes
b. Table name
c. Column names separated by “-“
Example: PK_Sessions, IX_Sessions_TimeStart-TimeFinish
6. View names should contain plural noun and end with “View” suffix.
Example: ActiveUsersView
7. Trigger names should consist of table name and trigger type.
Example: Users_InsertTigger
8. No prefixes or suffixes should be used for functions or stored procedures.

CODING STYLE
1. Stored procedure, view, trigger, function header should contain copyright
and description.
2. Every “select” statement clause (select, from, where, order by) on a sepa-
rate line. Exception: derived tables with short definition.
3. Long predicates in “where” clause should be placed on a separate line.
4. Use Inline comments wherever it is helpful and applicable.
5. Indent content of control blocks (statement block). Place if/while and begin
on the same line. Example:
if @isActive = 1 begin
[…]
end else begin
[…]
end
6. Indent content of transaction block (statement block), place begin
transaction and commit on separate line.
7. Table/view/function schema must be always provided (e.g. “dbo”) – it is very
important for performance since otherwise SQL Server is unable to share ex-
ecution plans between sessions and requests.
8. Do not select all the columns in views, stored procedures or queries using
select * […]. Explicit field list is always required. Adding a new column or
changing table/view column order should not be a breaking change.
9. Column names in views, stored procedures, triggers and functions should be
prefixed with table alias. Example: s.SessionId

3 © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

10. Long calculated fields should be placed on a separate line starting with field
name. Example:
[Minutes] = datediff(mi, s.TimeStart, s.TimeFinish)
11. Reserved names used as field name should be placed in square brackets, so
they do not get highlighted in source code editors.

DATABASE ACCESS
For SQL Server database access guidelines refer to:
1. Best Coding Practices - 9AAD135446
2. C# Best Practices - 9AAD134037
3. Java Coding Standards - 9AAD135383

SAMPLE CODE - VIEW

-- Copyright © ABB Inc. 2018
-- Returns session list with details.
create view dbo.SessionsView
as
select top 100 percent s.SessionId, u.LoginName, s.TimeStart, s.TimeFinish, c.ParticipantCount,
[Minutes] = datediff(mi, s.TimeStart, s.TimeFinish)
from dbo.[Sessions] s
inner join Users u on u.UserId = s.HostUserId
-- include session participants count (c)
left join (
select su.SessionId, count(0) ParticipantCount
from dbo.SessionUsers su
group by su.SessionId
) c on c.SessionId = s.SessionId
where
-- exclude sessions hosted by test users
u.LoginName not like 'testuser%'
-- exclude deleted sessions
and s.IsDeleted = 0
order by s.TimeStart

4 © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

SAMPLE CODE – STORED PROCEDURE

-- Author: Tomasz Jastrzębski, 2008
-- CreativeCommons CC0 Universal License
-- Procedure defragments table indexes.
create procedure dbo.DefragmentIndexes
@Table nvarchar(128),
@IndexCount int = null out,
@DefragCount int = null out
as

set nocount on;

declare row_cursor cursor fast_forward for

select [name], avg_fragmentation_in_percent
from sys.dm_db_index_physical_stats(db_id(), object_id(@Table), null, null, null) as a
inner join sys.indexes as b on a.object_id = b.object_id and a.index_id = b.index_id

declare @name sysname = null;

declare @fragmentation float = 0;

set @IndexCount = 0
set @DefragCount = 0

open row_cursor;
fetch next from row_cursor into @name, @fragmentation;

while @@fetch_status = 0 begin

set @IndexCount = @IndexCount + 1

if @fragmentation <= 5 begin

print @name + ' fragmented ' + cast(@fragmentation as varchar) + '% - skipping';
end

if @fragmentation > 5 and @fragmentation <= 30 begin

print @name + ' fragmented ' + cast(@fragmentation as varchar) + '% - reorganizing';
set @DefragCount = @DefragCount + 1
exec (N'alter index [' + @name + '] on ' + @Table + ' reorganize;');
end

if @fragmentation > 30 begin

print @name + ' fragmented ' + cast(@fragmentation as varchar) + '% - rebuilding';
set @DefragCount = @DefragCount + 1
exec (N'alter index [' + @name + '] on ' + @Table + ' rebuild;');
end

fetch next from row_cursor into @name, @fragmentation;

end

close row_cursor;
deallocate row_cursor;

5 © ABB Asea Brown Boveri 2018-2019

SQL Server Coding Standards GF-IS ADM Applications Performance Excellence (APE)

REFERENCES
1. Transaction Isolation Levels
https://docs.microsoft.com/en-us/sql/database-engine/transaction-
isolation-levels?view=sql-server-2014
2. Table Hints
https://docs.microsoft.com/en-us/sql/t-sql/queries/hints-transact-sql-
table?view=sql-server-2017
3. Index Architecture and Design Guide
https://docs.microsoft.com/en-us/sql/relational-databases/sql-server-
index-design-guide?view=sql-server-2017
4. Database normalization
https://en.wikipedia.org/wiki/Database_normalization
5. IS Standard for Managing Transactional Data - 9AAD135312
6. Best Coding Practices - 9AAD135446
7. C# Best Coding Practices - 9AAD134037
8. C# Coding Standards - 9AAD134036
9. Java Coding Standards - 9AAD135383
10. Source Code Management Standards - 9AAD134843

REVISION HISTORY

Rev. Page Change Description Author(s) Date

A all approved Tomasz Jastrzębski et al. 2019-04-09

6 © ABB Asea Brown Boveri 2018-2019

Dev
Information Systems
Applications Performance Excellence Department (APE)
Software Engineering Standards & Practices