Archives of the TeradataForum

Message Posted: Thu, 11 Dec 2008 @ 13:58:21 GMT

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

<-- Anonymously Posted: Thu, 11 Dec 2008 08:54 -->

For the case cited: duplicating the small table on all AMPs, scanning the large table, and doing a product join of each large table row with the small table (via in-memory lookup) is very likely the best performing plan.

Why do you think avoiding the product join would be desirable?

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