Siebel Customer-Centric Enterprise Warehouse Installation and Configuration Guide > Configuring Common Components of the Siebel Customer-Centric Enterprise Warehouse > Configuring Slowly Changing Dimensions >

About Identifying Historically Significant Attributes

You may want to retain a history of all the updates to a particular dimension so that you can use them in reports. These dimensions are known as historically significant attributes. For example, if a customer moves to a different region and you assign that customer a new regional salesperson and territory ID, you may want to keep records of that customer's account history with the original salesperson and territory ID. In this case, the salesperson and territory IDs are historically significant attributes. In contrast, you may have a load that populates the telephone number field. If your business does not perform data analysis on phone number history, then this information may be considered a historically insignificant attribute.

Identifying attributes as significant or insignificant allows you to determine the type of SCD you require. However, before you can select the appropriate type of SCD, you must understand their differences.

About the Extract View

The extract view of any given table in the Staging Area consists of four types of records:

• New records
• Changed records with data that is historically insignificant
• Changed records having historical significance
• Changed records whose changes have no significance of any kind and are ignored altogether

Of the four kinds of records, only the first three are of interest for the data mart. Of those three, brand new records and records whose changes are tracked as SCDs are both treated as new and become inserts into the data warehouse. Records with changes that are important but not historically tracked are overwritten in the data warehouse, based on the primary key.