Archives of the TeradataForum

Message Posted: Fri, 09 Jan 2004 @ 10:25:16 GMT

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

Ganga,

Thanks for experimenting with both syntax forms. We see that the semantically equivalent queries are indeed optimized differently but, fortunately, are equally fast.

As you are on R4 I take back my suggestion on in-memory derived table processing. That feature was introduced in R5 and does not apply to you yet. I think what happened to the query that I introduced was that it was rewritten by the optimizer to the form equivalent to the one of the original slow query. It is a trifle strange as your final query seems to be more complicated for the optimizer to figure out the fast solution than the query explicitly suggesting the derived table. But this is how it goes.

Good luck with the migration.

Victor

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