Article: Multidimensional Database Technology

Question 1

Multidimensional data model

Answer 1

Categorize data as facts with associated numerical measures or textual dimensions that characterize facts.
Three important implications:
- Data warehouses
- Online analytical processing (OLAP)
- Data mining applications

Question 2

Queries

Answer 2

Aggregate measure values over a range of dimensions values to provide results such as total sales per month of a given product.

Question 3

Pivot table

Answer 3

A two-dimensional spreadsheet with associated subtotals and totals that support viewing more complex data by using several dimensions on the x- or y-axis and displaying data on multiple pages.

Question 4

Cubes

Answer 4

Support hierarchies in dimensions and formulas without duplicating their definitions. Combinations of dimensions define a cube’s cells. Depending on the specific application, the cells in a cube range from sparse to dense. Cubes tend to become sparser as dimensionality increases and as the dimension values granularities become finer.

Question 5

Dimensions

Answer 5

Are used for selecting and aggregating data at the desired level of detail. The goal of dimensions is to provide as much context for facts as possible. In contrast to relational databases, controlled redundancy is appropriate in multidimensional databases if it increases the data’s information value. Mostly, there is only redundancy in dimensions and not in facts.

Question 6

Facts

Answer 6

Implicitly defined by their combination of dimension values. Three types of facts:

• Events
• Snapshots
• Cumulative snapshots

Question 7

Events

Answer 7

One fact represents the same instance on an underlying phenomenon.

Question 8

Snapshots

Answer 8

Model an entity’s state at a given point of time.

Question 9

Cumulative snapshots

Answer 9

Handles information about an activity up to a certain moment.

Question 10

Measure

Answer 10

Consists of two components:

• A fact’s numerical property.
• A formula

Question 11

Three classes of measures

Answer 11

• Additive
• Semi additive
• Non additive

Question 12

Additive measure

Answer 12

Can be meaningfully combined along any dimension. Can occur from any kind of fact.

Question 13

Semi additive measure

Answer 13

Cannot be combined along one or more dimensions. Generally occurs for snapshots and cumulative snapshots.

Question 14

Non additive measure

Answer 14

Cannot be combined along any dimensions. Can occur from any kind of fact.

Question 15

Ways of implementing multidimensional modelling

Answer 15

• MOLAP: includes provisions for handling sparse arrays and apply advanced indexing and hashing to locate the data when performing queries. MOLAP is more flexible in redefinitions of the cube and handling updates.
• ROLAP: uses relational database technology for storing data, and they also employ specialized index structures to achieve good query performance. ROLAP is better in scaling in the number of facts that it needs to store.