It has now become the standard practice in data warehouse projects to use advance relational database technology with a star schema design. The star schema term comes from the picture that is created by the relationship of database tables to each other.

A star schema is made of a fact table surrounded by several dimension tables. This design is used to provide fast query times with minimum database storage requirements.

A Dimension Table stores all attributes pertinent to the dimension row, (i.e.: product number, product name, product type, etc.).

The Fact Table stores the relationships between the dimension tables for a single event or transaction, (i.e.: the Sales Order Fact Table stores the relationship for each sales order line item). In addition facts about each transaction are stored, sales order quantity, unit cost, etc.

To support aggregations of individual facts the fact table stores the lowest level of grain available. In the Sales Order Fact Table we store each individual sales order line item, rather than a summary by year, by customer, or by product, thus significantly reducing the processing time.

Back to Data Warehousing