Archives of the TeradataForum
Message Posted: Fri, 15 Jun 2012 @ 15:06:50 GMT
Teradata has a "star join" optimization, which is essentially to first cross-join all dimension tables together, and then join the result to the fact table through the primary index on the fact table.
This can minimize data redistribution of the larger fact table. For this to work, the primary index of the large fact table must be a composite of the primary index columns of all of the dimension tables (Region_ID, Product_Dim_Id, etc.)
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|