From: | "scott(dot)marlowe" <scott(dot)marlowe(at)ihs(dot)com> |
---|---|
To: | "Lufkin, Brad" <brad(dot)lufkin(at)ngc(dot)com> |
Cc: | "'psql'" <pgsql-jdbc(at)postgresql(dot)org> |
Subject: | Re: Query Time |
Date: | 2003-08-13 20:39:00 |
Message-ID: | Pine.LNX.4.33.0308131436100.7764-100000@css120.ihs.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-jdbc |
How many rows in that table? 90 seconds is a long time for a select to
take, period, with or without a limit.
On Wed, 13 Aug 2003, Lufkin, Brad wrote:
> Thanks for the info. Still, I'm surprised that a limited query would take 90
> seconds.
>
> -----Original Message-----
> From: scott.marlowe [mailto:scott(dot)marlowe(at)ihs(dot)com]
> Sent: Wednesday, August 13, 2003 4:20 PM
> To: Lufkin, Brad
> Cc: 'psql'
> Subject: Re: [JDBC] Query Time
>
>
> On Wed, 13 Aug 2003, Lufkin, Brad wrote:
>
> > I'm running the following query:
> >
> > SELECT * FROM someTable LIMIT 20
> >
> > I turned on explain analyze (tip #8!) and was told that the query plan was
> > sequential (no surprise there) with an estimated cost of between 0.00 and
> > 1.07. Surprisingly, the actual time was around 90000 msec (or
> one-and-a-half
> > minutes). What's going on?
>
> costs are estimated as a percentage of a cost of a single page access in
> sequential mode. I.e. a single sequential page access is assumed to cost
> 1.0, and everything is relative to that. The cost numbers do NOT
> translate directly into any time unit.
>
>
From | Date | Subject | |
---|---|---|---|
Next Message | Lufkin, Brad | 2003-08-13 20:52:15 | Re: Query Time |
Previous Message | Tom Lane | 2003-08-13 20:29:05 | Re: Query Time |