Archives of the TeradataForum

Message Posted: Tue, 07 Mar 2006 @ 20:43:23 GMT

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

I would try replacing both tables with derived tables, and move the coalesce into the select portion of the derived tables. This will remove the coalesce from the join condition which is very CPU intensive. Depending on volumes it may or may not be quicker, but I have used this method on many occasions with quite considerable performance improvements.

This will depend on whether or not your system is set to materialize derived tables (an internal DBC control setting), if you are unsure check your explain output which should show 2 retrieves.

Regards

Keith.

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