OLAP Cubes Ming-Nu Lee OLAP (Online Analytical Processing)

 Performs multidimensional analysis of business data

 Provides capability for complex calculations, trend analysis, and sophisticated modelling

 Foundation of many kinds of business applications OLAP Cube

 Multidimensional

 Method of storing data in a multidimensional form

 Optimized for and OLAP apps

 Structure:

 Data(measures) categorized by dimensions

 Dimension as a hierarchy, set of parent-child relationships

 Not exactly a “cube” Schema

1)

• Every dimension is represented with only one dimension table

• Dimension table has attributes and is joined to the using a foreign key

• Dimension table not joined to each other

• Dimension tables are not normalized

• Easy to understand and provides optimal disk usage Schema

1)

• Dimension tables are normalized

• Uses smaller disk space

• Easier to implement

• Query performance reduced

• Maintenance increased Dimension Members

 Identifies a data item’s position and description within a dimension

 Two types of dimension members

1) Detailed members

• Lowest level of members

• Has no child members

• Stores values at members intersections which are stored as fact data

2) Aggregate members

• Parent member that is the total or sum of child members

• A level above other members Fact Data

 Data values described by a company’s business activities

 Exist at Member intersection points

 Aggregation of transactions integrated from a RDBMS, or result of Hierarchy or cube formula calculation Operations

 Slice

 Dice

 Roll-up (Drill up)

 Roll-down (Drill down)

 Pivot Slice The act of picking a rectangular subset of a cube by choosing a value for one of its dimensions Dice Similar to slice, but it allows analyst to pick specific values of multiple dimensions which produces a subcube Roll-Up (Drill up) Also known as consolidation or aggregation 1) Reducing dimensions 2) Climbing up concept hierarchy Roll-Down (Drill Down) Fragment data into smaller parts Opposite of roll-up 1) Increasing a dimension 2) Moving down the concept hierarchy Pivot Provides a different look or presentation of data by rotating the cube or axes Multidimensional Expressions (MDX)

 Originally developed in the late 1990s

 Language for expressing analytical queries

 Extension of SQL language

 A MDX expression returns a multi-dimensional result set that consist of axis data and cell data Advantages

 Information and calculations are consistent in an OLAP cube

 “What if” scenarios can be quickly created and analyzed

 Broad or specific terms can be easily searched for in the database

 Slice and dice cube data by various dimensions, measures and filters

 Good for analyzing time series Disadvantages

 Data must be organized into a star or snowflake schema which hard to implement and administer

 A single OLAP cube cannot have a large number of dimensions

 Transactional data unable to be accessed with OLAP cubes

 Modification of a cube requires a full update of the cube References

 What is OLAP (Online Analytical Processing): Cube, Operations & Types, https://www.guru99.com/online-analytical-processing.html#8  (2017). Operations – SQL Queries, https://blogs.perficient.com/2017/08/02/data-cube-operations-sql-queries/  http://olap.com/  https://en.wikipedia.org/wiki/OLAP_cube  Rouse, M. (2012). OLAP cube, https://searchdatamanagement.techtarget.com/definition/OLAP-cube  Rouse, M. (2012). multidimensional expressions (MDX), https://searchsqlserver.techtarget.com/definition/multidimensional-expressions- MDX  Star and SnowFlake Schema in Data Warehousing, https://www.guru99.com/star- snowflake-data-warehousing.html