Re: Vanishing table in 11.2.0.3
From: Mladen Gogala <gogala.mladen_at_gmail.com>
Date: Thu, 2 Feb 2012 13:17:01 +0000 (UTC)
Message-ID: <pan.2012.02.02.13.17.01_at_gmail.com>
On Thu, 02 Feb 2012 09:47:48 +0000, Jonathan Lewis wrote:
Date: Thu, 2 Feb 2012 13:17:01 +0000 (UTC)
Message-ID: <pan.2012.02.02.13.17.01_at_gmail.com>
On Thu, 02 Feb 2012 09:47:48 +0000, Jonathan Lewis wrote:
> So many questions in such a short space.
>
> Table Elimination -
> The IN subquery is transformed into EXISTS The EXISTS subquery is
> single table based on a declared primary key,
> therefore is subject "uncosted unnesting" (I think, that's the rule
> applied here)
Ah, I forgot about that. I only looked whether the column was indexed or not.
> There is no "push_subq" in this plan, by the way. Push_subq (in
> principle) means that Oracle will run the subquery at the earliest
> possible moment rather than leaving it to the end of the plan.
That is what I thought. I thought that the subquery was executed first, as a part of the parse, and turned into an explicit in-list. That's why I mentioned push_subq.
-- http://mgogala.byethost5.comReceived on Thu Feb 02 2012 - 07:17:01 CST
