Online Analytical

Processing (OLAP)
ดร.มารุ ต บูรณรัช [email protected]
รหัสวิชา IT5105: เทคโนโลยีคลังข้อมูล (Data Warehouse Technology)
สาขาเทคโนโลยีสารสนเทศ
คณะวิทยาศาสตร์และเทคโนโลยี มหาวิทยาลัยราชภัฏเชียงราย
ภาคการศึกษาที่ 1 ปี การศึกษา 2551
Topics
 Business Intelligence (BI) Technologies
 OLAP definitions
 Data cube & hypercube
 OLAP operations
 Types of OLAP tools
 OLAP Demo

2
Business Intelligence (BI)
Technologies
 With the growth in data warehousing,
users demand for more powerful access
tools that provide advanced analytical
capabilities
 Two main types of these access tools are
 Online Analytical Processing (OLAP)
 Data mining

3
Business Intelligence (BI)
Technologies (2)
 OLAP and Data Mining differ in what they
offer the user
 complementary technologies
 Data warehouse (or data marts) together
with tools such as OLAP and /or data
mining are referred to as Business
Intelligence (BI) technologies

4
What is OLAP?
 Online Analytical Processing (OLAP) is a system that
further transforms the data into a more structured
(summarized) form than tables
 OLAP is a form of Executive Information System
(EIS) and Decision Support System (DSS)
 OLAP looks at data in multi-dimensional form (data
cube)
 OLAP can be used by multiple users to access data
in a data warehouse, e.g. via Internet
 OLAP provides managers with a quick and flexible
access to large volume of data

5
OLAP Definitions
 Codd (1993) – OLAP is “the dynamic
synthesis, analysis, and consolidation of
large volumes of multi-dimensional data.”
 OLAP technology uses a multi-
dimensional view of aggregate data to
provide quick access to strategic
information

6
Why OLAP?
 Users need powerful tools for the analysis of
large-volume of data,
 i.e. data in data warehouse
 Two main types of analysis tools for data
warehouse are:
 Online Analytical Processing (OLAP)
 “top-down” analysis

 Data Mining
 “bottom-up” analysis

7
Why OLAP? (2)
 OLAP vs. general-purpose query tools
 OLAP has ability to answer ‘what if?’ and
‘why?’ questions (not only ‘what’, ‘when’,
‘where’ and ‘how much’ questions)
 OLAP has more advanced and interactive
functionalities
 Browsing
 Calculations

 Complex analyses

8
OLAP Applications
 OLAP applications usually have the following
common features:
 Multi-dimensional views of data
 Data can be viewed from various perspectives, e.g.
product, location, time, etc.
 Support for complex calculations
 e.g. sales forecasting, moving averages, percentage
growth, etc.
 Time intelligence
 e.g. comparisons of sales performance between
different time periods

9
 Multi-dimensional views of data
 Example of 2 dimensional views of data

10
Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
 Multi-dimensional views of data (2)
 Example of 3 dimensional views of data

11
Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
Data Cube
 Multi-dimensional structures are best visualized
as cubes of data
 Cube represents data as cells in an array
 Each side of a cube is a dimension
 A cube supports matrix arithmetic
 Hypercube is a form of data cube that has more
than 3 dimensions
 Hypercube can be represented as cube that contains
cubes for other dimensions (cubes within cubes)
 As number of dimensions increases, number of the
cube’s cells increases exponentially

12
Data Cube Example
Mar 05
Feb 05
Jan 05

Product A $ $ $ $

Product B $ $ $ $

Product C $ $ $ $

Store A Store B Store C Store D

13
OLAP Operations
 Slice
 Select data on a single dimension of a data cube
 Dice
 Extracts a sub-cube from the original cube
 Roll-up (aggregation)
 Combing of cells for one dimension
 Generalization, e.g. Jan, Feb, Mar = Quarter 1
 May be used with “concept hierarchy”
 Drill-down
 Reverse of “Roll-up” operation
 Examine data at level of greater detail, e.g. Northern Region = Chiang Mai,
Chiang Rai, …
 Rotation (pivot)
 Allow user to view data from a new perspective
 Axis rotation

14
 OLAP Operations (2)
Slice
Mar 05
Feb 05
Jan 05 Mar 05
Feb 05
Product A $ Jan 05

Product B $ Product A $ $ $ $
Store A Store B Store C Store D
Product C $
Store A
Jan 05
Dice
Product A $
Store A
15
 OLAP Operations (3)

Mar 05
Quarter 1, 05 Feb 05
Jan 05
Product A $ $ $ $
Product A $ $ $ $
Product B $ $ $ $ $
Product B $ $ $
Product C $ $ $ $ Product C $ $ $ $

Store A Store B Store C Store D Store A Store BStore C Store D

Roll-up Drill-down

16
OLAP Operations (4)
Product A
Product B
Product C

$
$

$
Store A $ $ $

$
$
Store B $ $ $ $

$
Store C $ $ $
$

Store D
$

$
$ $ $

Jan 05 Feb 05 Mar 05

Rotation
17
Concept Hierarchy
 Attribute may have concept hierarchies
associated with
 Examples
Year Region

Quarter Province

Month City
18
Types of OLAP tools
 OLAP tools are categorized based on how
they store and process multi-dimensional
data
 4 main types of OLAP tools:
 Multi-dimensionalOLAP (MOLAP)
 Relational OLAP (ROLAP)
 Hybrid OLAP (HOLAP)
 Desktop OLAP (DOLAP)

19
Multi-dimensional OLAP (MOLAP)
 Use Multi-dimensional Database Management System
(MDDBMS) to organize and analyze data
 Use some efficient storage techniques to minimize disk
space requirement
 Provides good performance when data is used as
designed
 Provide a tight coupling between data structure and
presentation layer
 Access to data structure may be provided via application
programming interfaces (APIs)

20
 MOLAP Architecture

Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
21
MOLAP Issues
 MOLAP products require different skills
and tools to build and maintain the
database, thus increasing the cost and
complexity of support
 MDDBMS is a new and immature technology
(compared to RDBMS)

22
Relational OLAP (ROLAP)
 Fastest-growing type of OLAP technology
 MOLAP databases has some limitations
 Not all data can be efficiently stored in MOLAP
databases
 Uses supports from RDBMS
 avoids need to create multi-dimensional database
 creates multi-dimensional views from relational
database
 May use SQL to support multi-dimensional data
analysis

23
 ROLAP Architecture

24
Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
ROLAP Issues

 Need to create a middleware to work with

multi-dimensional applications
 The middleware must convert relational data
structure to multi-dimensional data structure
 Performance problems for complex
queries that require complex
transformations from relational data

25
Hybrid OLAP (HOLAP)

 Provide query support for both RDBMS

and MDDBMS
 Query data directly from the RDBMS using
SQL or via a MOLAP server in the form of a
data cube
 May cause data redundancy and
inefficient network usage

26
 HOLAP Architecture

Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
27
Desktop OLAP (DOLAP)
 Store and process the OLAP data on client side
 Data are held on client machines
 Database may be distributed in advance, or created
on demand (e.g. through the Web)
 The maintenance of database is usually done by a
central server
 DOLAP uses the power of desktop PC to
perform multi-dimensional calculations

28
 DOLAP Architecture

29
Source: Connelly & Begg (2001), Database Systems: A Practical Approach to Design, Implementation, and Management (3rd Edition), Addison Wesley
DOLAP Issues
 Security (access control) can be difficult
 Cannot utilize access control feature of
DBMS
 Current trends are towards thin client
machines
 Complex calculations are increasingly moved
to server machine rather than client machine

30
OLAP Benchmark
 APB-1 (OLAP Council, 1998) is a standard for OLAP
benchmark
 Measurement of OLAP server performance
 APB-1 evaluates OLAP server performance for the
following operations:
 Loading of data
 Aggregation of data
 Complex Calculations
 Time series analysis
 Complex Queries
 Drill-down through hierarchies
 Multiple online sessions
 etc.

31
OLAP Benchmark (2)
 A benchmark metric used by APB-1 is
AQM (Analytical Queries per Minute)
 AQM measures the number of analytical
queries that an OLAP server can process
per minute
 Thetime is measured from when the data is
loaded until the results are returned to user

32
OLAP Extensions to SQL
 SQL has limited capability to support complex
management queries
 ANSI adopted a set of OLAP functions as an
extension to SQL
 IBM and Oracle jointly proposed these extensions in
1999 as part of the current SQL standard
 The extensions are referred to as the ‘OLAP
package’:
 Feature T431, ‘Extended Grouping capabilities’
 Feature T611, ‘Extended OLAP operators’

33