Thanks Tom,
alter table boden.massnahmeobjekt alter column aktennummer set statistics 1000;
fixed it.
Regards
Christian
-----Ursprüngliche Nachricht-----
Von: pgsql-performance-owner@postgresql.org [mailto:pgsql-performance-owner@postgresql.org] Im Auftrag von Tom Lane
Gesendet: Donnerstag, 17. Juni 2010 16:59
An: Kaufhold, Christian (LFD)
Cc: pgsql-performance@postgresql.org
Betreff: Re: [PERFORM] Query slow after analyse on postgresql 8.2
"Kaufhold, Christian (LFD)" <Christian.Kaufhold@blfd.bayern.de> writes:
> I have the following query that I run agains postgresql 8.2:
> ...
> But when I run analyse the same query runs for hours.
Seems like the core of the problem is here:
> -> Seq
> Scan on massnahmeobjekt m (cost=0.00..3271.88 rows=1 width=94)
> (actual
> time=0.387..577.771 rows=3694 loops=1)
> Filter: ((aktennummer)::text ~* 'M\\-2009\\-1'::text)
If that rowcount estimate weren't off by three orders of magnitude you probably would be getting a more appropriate
plan. The first thing you could try is increasing the statistics target for aktennummer. Also, if you're running in a
non-Clocale and this is 8.2.5 or older, try a more recent 8.2.x. Updating to 8.3 or 8.4 might help even more.
regards, tom lane
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance