Star schema is the most basic data warehousing dimensional structure and database schema, consisting of one or more fact tables referencing any number of dimension tables.
Star schema is the most basic data warehousing (data mart) dimensional structure and database schema, consisting of one or more fact tables referencing any number of dimension tables. It is organized like:
- Facts - event that is counted or measured. They can be at a very atomic level. Generally assigned a surogate key to ensure every row can be identified.
- Dimensions - reference information about facts. The records contain detailed information and attributes describing the fact data. Usually assigned a surrogate primary key.
Within the data warehouse and data mart the dimension table is associated with fact tables using foreign key relationship.
Advantages
In well-designed schemas multidimensional data can be analyzed easily. They are good in decision-support environments. Some of the advantages are query speed, data load and administration. The built-in referential integrity is good but not as in highly normalized databases though.
Examples
An example star schema could have a Sales Fact with Date, Product, and Store Dimensions. Any numerical performance measurements related to Sales would also be stored in the fact, like "Quantity Sold" and "Dollar Amount of Sale". The remaining columns in the Sales Fact would be foreign keys linking to the Dimensions.
More information
