In case you aren't comfortable running unreleased planner patches from
pgsql-hackers, a workaround was discussed on this list recently:
http://archives.postgresql.org/pgsql-performance/2009-09/msg00036.php
On Wed, 09 Sep 2009 06:05:22 -0400, Heikki Linnakangas
<heikki.linnakangas@xxxxxxxxxxxxxxxx> wrote:
Kevin Kempter wrote:
Hi all I have a large table (>2billion rows) that's partitioned by date
based
on an epoch int value. We're running a select max(id) where id is the
PK. I
have a PK index on each of the partitions, no indexes at all on the base
table.
If I hit a partition table directly I get an index scan as expected:
The planner isn't smart enough to create the plan you're expecting.
There was discussion and even a patch posted recently about that:
http://archives.postgresql.org/pgsql-hackers/2009-07/msg01115.php
It seems the thread petered out, but the concept seems sane.
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
--
Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
