Archives of the TeradataForum

Message Posted: Tue, 16 Jul 2002 @ 11:05:46 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Matthew,

In regards to the data warehouse I work with, a majority of our large tables have NUPIs; whereas, the majority of our smaller tables have UPIs. Our smaller tables are typically duplicated to all AMPs during our join processes; therefore, it's easier just to go ahead and define these tables as UPIs based upon the logical primary key. It's the larger tables that we spend more time on when determining the primary index.

My 2 cents...

	<Prev	Next>		<<First	<Prev	Next>	Last>>