Re: [PgSQL 7.4.1] Problem with subquery

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Edwin Pauli <edwin(at)epauli(dot)dyndns(dot)org>
Cc: pgsql-general(at)postgresql(dot)org
Subject: Re: [PgSQL 7.4.1] Problem with subquery
Date: 2004-02-26 18:30:44
Message-ID: 13360.1077820244@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Edwin Pauli <edwin(at)epauli(dot)dyndns(dot)org> writes:
> There are troubles with a subquery after the upgrade.

> SELECT team_naam, team_id, wpim, (
> SELECT count(thuis_uitslag)
> FROM wedstrijden
> WHERE (thuis_wed = t.team_id
> OR uit_wed = t.team_id)
> AND seizoen_id = 3 AND klasse_id = 1)
> AS wd, (
> SELECT count(t.team_id)
> FROM wedstrijden w
> WHERE (thuis_wed = t.team_id AND thuis_uitslag >
> uit_uitslag OR uit_wed = t.team_id
> AND thuis_uitslag < uit_uitslag)
> AND seizoen_id = 3
> AND klasse_id = 1 )
> AS gw
> FROM teams t
> WHERE seizoen_id = 3 AND klasse_id = 1
> GROUP BY t.team_naam, t.team_id, t.team_id, t.wpim

> ERROR: more than one row returned by a subquery used as an expression

Since t.team_id is a variable of the outer query, count(t.team_id) is an
aggregate of the outer query according to the SQL spec, and accordingly
it is just a constant from the perspective of your second sub-SELECT.
That sub-SELECT is therefore not doing any aggregation of its own and
is trying to return multiple rows.

We didn't implement this fine point per-spec before 7.4, but we do now.
See http://archives.postgresql.org/pgsql-hackers/2003-06/msg00070.php

Solution: use count(*) instead. AFAICS you have no need to check
whether t.team_id is nonnull, since the WHERE clause could not succeed
anyway if it's null.

regards, tom lane

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Jeffrey W. Baker 2004-02-26 18:56:14 PostgreSQL on XFS experiences?
Previous Message Pavel Stehule 2004-02-26 18:10:39 Re: PL/pgSQL debugger