| From: | Michael Fuhr <mike(at)fuhr(dot)org> |
|---|---|
| To: | Clive Page <cgp(at)star(dot)le(dot)ac(dot)uk> |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Baffled by failure to use index when WHERE uses a function |
| Date: | 2006-03-10 09:57:54 |
| Message-ID: | 20060310095753.GA9260@winnie.fuhr.org |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Fri, Mar 10, 2006 at 09:14:27AM +0000, Clive Page wrote:
> I have a table cov3 of about 3 million rows, with a B-tree index on an
> integer column called hpix. If I do a simple select on this column it
> works in milliseconds, using the index naturally:
>
> select * from cov3 where hpixint = 482787587;
^^^^^^^
That's not the column you said you were using and that your other
example uses. Have you verified that hpix has an index?
> So I would like to use this function to find rows, and I try for example:
>
> select * from cov3 where hpix = healpix(2.85,-11.48);
>
> but it takes ages. An EXPLAIN shows why, it insists upon a sequential
> scan:
If healpix() always returns the same output for given input then
define it to be IMMUTABLE. For more information see "Function
Volatility Categories" in the documentation.
http://www.postgresql.org/docs/8.1/interactive/xfunc-volatility.html
--
Michael Fuhr
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Richard Huxton | 2006-03-10 10:36:19 | Re: PL/pgSQL question |
| Previous Message | Martijn van Oosterhout | 2006-03-10 09:53:57 | Re: Baffled by failure to use index when WHERE uses a function |