Archives of the TeradataForum

Message Posted: Thu, 28 Aug 2008 @ 15:27:13 GMT

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

Chinmay:

Unfortunately, there is no way to re-write this query to make it faster. The problem lies in a diverse design of the underlying tables. It appears that one is partitioned and the other is not. In order for it to be totally optimized, both tables would need to have the PI on the join columns and be partitioned on the same column. That way, the rows are AMP local (PI) eliminating the need for redistribution into spool for both tables and sequenced the same to eliminte the need for spooling so that a sort can be performed.

I am sure this is not the simple answer you hoped to receive.

Regards,

Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor

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