Chapter 7. Modeling the Business Using Star Schemas

When working with data from the data warehouse, it's likely that star schemas are used to deliver the data to the end user. Not directly, however; typically, a reporting or analysis tool is used to access the data, while more advanced users might use a query tool directly. Nevertheless, it's important to note that star schemas are the vehicle of choice whether working with a Kimball-style data bus architecture or an Inmon-style corporate information factory.


There is one exception to this rule. An advanced analyst or data miner often needs access to the complete set of data in the data warehouse, thereby circumventing the created star schemas in an Inmon-style data warehouse solution. This discussion, however, will be limited to star schema access.

What Is a Star Schema?

The first thing you might ask is: Why are these database models called "star schema"? Probably because the entity-relationship diagram of this kind of schema resembles a star. The center of the star consists of a large fact table and the points of the star are the dimension tables. Most users first encounter a star schema in a sales data mart with customers, products, stores, promotions, and time, as displayed in Figure 7-1.


Although we use five points to make the model resemble a star, it is by no means required to use five points. In fact, even a fact table with only one or two dimension is also called a star schema.

Figure 7.1. Star schema diagram

It's the ...

Get Pentaho® Solutions: Business Intelligence and Data Warehousing with Pentaho and MySQL® now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.