Archives of the TeradataForum

Message Posted: Thu, 07 Feb 2002 @ 08:37:06 GMT

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

My experience is very similar to Matt's.

1. update without join changing 2 mio rows - 1 minute

2. same table, different update with a join changing 2 mio rows - 20 minutes

3. the same as second case but using INSERT/SELECT strategy - 2 minutes

If there was a problem with distribution or stats, 3 would not improve compared to 2.

So the only possible explanation I can see is, that TJ is created differently depending on whether there is a join ivolved or not. Is this true? Otherwise 1 would take the same time as 2, right?

Or is there another explanation?

Does anybody know, what information is inserted into TJ, when an UPDATE is performed?

Petr Horsky

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