Dear Tom, (or anyone who followed the belowmentioned thread)
I read that thread (more-or-less), but couldn't have noticed the same
symptoms in my analyze output. So, to summarize my reading on this (please
confirm or fix):
* The symptom is the differing width in 7.2 and 7.3
* This causes more hdd work, that takes lots of time (indeed, the hdd was
going crazy)
* The query is probably good as it is; it's 7.3 that's slow (but more
reliable than 7.2) and 7.4 will most likely fix the problem.
If all these are correct, that's enough info to me. Hopefully it'll move
from a Cel333 (the developers' server) to an IBM 2x2.4 Xeon with 5-HDD SCSI
Raid (the business server).
G.
------------------------------- cut here -------------------------------
----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
Sent: Thursday, May 22, 2003 6:02 PM
Subject: Re: [PERFORM] ugly query slower in 7.3, even slower after vacuum
full analyze
> "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" <surrano@mailbox.hu> writes:
> > This is a rather nasty query, built up from several parameters, and it
> > proved to be 7--15 times slower in 7.3 than in 7.2.
>
> I think you are running into the same subselect-in-targetlist
> shortcoming as Eugene Fokin did:
> http://archives.postgresql.org/pgsql-performance/2003-05/msg00204.php
>
> regards, tom lane
>