From: Dennis Bjorklund <db@zigo.dhs.org>
To: Carlos Moreno <moreno_pg@mochima.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: Seqscan/Indexscan still a known issue?
Date: Sat, 27 Jan 2007 08:26:13 +0100
Message-ID: <45BAFE95.8060409@zigo.dhs.org> (raw)
In-Reply-To: <45BAB682.6060709@mochima.com>
References: <45BAB682.6060709@mochima.com>
Carlos Moreno skrev:
> When I force it via "set enable_seqscan to off", the index scan
> takes about 0.1 msec (as reported by explain analyze), whereas
>
> For the time being, I'm using an explicit "enable_seqscan off"
> in the client code, before executing the select. But I wonder:
> Is this still an issue, or has it been solved in the latest
> version?
For most queries it has never been an issue. Every once in a while there
is a query that the planner makes a non-optimal plan for, but it's not
that common.
In general the optimizer has improved with every new version of pg.
Almost everyone I've talked to that has upgraded has got a faster
database tham before. It was like that for 7.4->8.0, for 8.0->8.1 and
for 8.1->8.2. So in your case going from 7.4->8.2 is most likely going
to give a speedup (especially if you have some queries that isn't just
simple primary key lookups).
In your case it's hard to give any advice since you didn't share the
EXPLAIN ANALYZE output with us. I'm pretty sure it's possible to tune pg
so it makes the right choice even for this query of yours but without
the EXPLAIN ANALYZE output we would just be guessing anyway. If you want
to share it then it might be helpful to show the plan both with and
without seqscan enabled.
How often do you run VACUUM ANALYZE; on the database?
/Dennis
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: db@zigo.dhs.org, moreno_pg@mochima.com
Subject: Re: Seqscan/Indexscan still a known issue?
In-Reply-To: <45BAFE95.8060409@zigo.dhs.org>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox