From: | Bruno Wolff III <bruno(at)wolff(dot)to> |
---|---|
To: | Markus Wollny <Markus(dot)Wollny(at)computec(dot)de> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Tuning random_page_cost |
Date: | 2004-07-13 12:43:54 |
Message-ID: | 20040713124354.GC3710@wolff.to |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Tue, Jul 13, 2004 at 13:30:29 +0200,
Markus Wollny <Markus(dot)Wollny(at)computec(dot)de> wrote:
> Hi!
>
> I've got a query that has a where clause on a timestamp field:
>
> select t.board_id
> , t.thread_id
> from public.board_thread t
> where t.last_reply <= now()-'6 months'::interval
> limit 1
>
> I've got random_page_cost set to 1.4 which is fine for most queries; yet
> here the planner prefers a (slower) sequential scan:
If you know that an index scan is better you can fudge the query to
change the planner's estimate of the number of rows that will be returned.
The normal fudge is to add a >= check for some timestamp that is earlier
than any in the column, so that you will have a range condition, but not
any change in the rows returned.
From | Date | Subject | |
---|---|---|---|
Next Message | Markus Wollny | 2004-07-13 12:57:50 | Re: Tuning random_page_cost |
Previous Message | Tony Reina | 2004-07-13 12:06:44 | Re: Need Libpq.dll w/ SSL |