Re: =?iso-8859-2?B?T2Rwb3bs7zog?=Re: backend don't use index when querying by indexed column

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: zhabala(at)telecom(dot)cz
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: =?iso-8859-2?B?T2Rwb3bs7zog?=Re: backend don't use index when querying by indexed column
Date: 2001-08-07 13:33:24
Message-ID: 29211.997191204@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

<zhabala(at)telecom(dot)cz> writes:
> first of all thank You for Your response. But here is the
> reality. Table formula have about 450000 rows at this time.
> The id_formula column is int4 defined as not null primary key.

It is eh? Have you done a VACUUM ANALYZE on this table lately?
4500 rows (1%) would be the default match estimate in the absence of
any statistics.

> And the index have clause unique, but backend still don't use it.

7.2 will know enough to realize that presence of a unique index means
the column is unique, but existing releases only look at the VACUUM
ANALYZE statistics. No statistics = bad plan.

regards, tom lane

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Peter Eisentraut 2001-08-07 17:40:31 Re: v7.0.2: can't use 'EXTRACT(EPOCH FROM'
Previous Message pgsql-bugs 2001-08-07 10:20:49 ecpg: unstable INSERT operation