Sunday January 21st 2018




Slow Changing Dimensions another perspectives

Business intelligence is an analysis about business evidence so that companies can make informed decisions.

Facts record transactions, and Dimensions contain information about the transactions, such as customer name, time, weather, product and advertising promotions slowly changing dimensions (SCD) result from the fact that over time, the world changes and dimension values must also be updated to reflect those changes.[1]

Not every dimension is slowly changing and not every slowly changing should be handled in the same way. One example of dimension might be time. Time doesn’t change. October 3 1940 will always be October 3 1940. How ever marriage status, sales district, last names do change.[1]

You don’t necessarily want to update every change in the same way, you might want to maintain history in some changes, overwrite others, or simply detect when a change is attempted. For example, if the sale district for a company is reorganized, to insure to ensure that sales report and queries remain accurate, you might want to keep a record of previous sales district for all salespersons. But customer name happen so rarely and because names are simply identifiers for the same entity, you might want to overwrite last name attribute of customer dimensions. This is a fundamental problem with dimensions. [1]

Examples of Slow changing Dimensions


1- Preview this book Microsoft SQL Server 2005 Integration Services By Kirk Haselden, Trey Johnson