Merlin Moncure <mmoncure@gmail.com> writes:
> On Wed, Nov 7, 2012 at 5:16 AM, Rodrigo Rosenfeld Rosas
> <rr.rosas@gmail.com> wrote:
>> This was a great guess! Congrats, Merlin:
> Heh -- that was tom's guess, not mine. What this does is confirm the
> planner regression and that elevates the importance of Tom's request
> to get sample data so we (he) can fix it.
Well, the fact that it's a planner runtime problem and not a
quality-of-plan problem is new information (I'd been assuming the
latter). Given that, it's possible it's already fixed:
http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=ca2d6a6cef5740b29406980eb8d21d44da32634b
but I'd still want to see a test case to be sure. In any case,
it's not clear what's the critical difference between the "fast" and
"slow" versions of the query.
regards, tom lane