Date: Wed, 01 May 2013 12:48:07 +0000
Message-Id: <E1UXWS7-0007V8-N2@wrigleys.postgresql.org>
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Subject: BUG #8130: Hashjoin still gives issues
To: pgsql-bugs@postgresql.org
From: stefan@konink.de
Precedence: bulk
Sender: pgsql-bugs-owner@postgresql.org

The following bug has been logged on the website:

Bug reference:      8130
Logged by:          Stefan de Konink
Email address:      stefan@konink.de
PostgreSQL version: 9.2.4
Operating system:   Linux
Description:=20=20=20=20=20=20=20=20

We figured out that two very close query give a massive difference
performance between using select * vs select id.

SELECT *
FROM ambit_privateevent_calendars AS a
	,ambit_privateevent AS b
	,ambit_calendarsubscription AS c
	,ambit_calendar AS d
WHERE c.calendar_id =3D d.id
	AND a.privateevent_id =3D b.id
	AND c.user_id =3D 1270
	AND c.calendar_id =3D a.calendar_id
	AND c.STATUS IN (
		1
		,8
		,2
		,15
		,18
		,4
		,12
		,20
		)
	AND NOT b.main_recurrence =3D true;

With some help on IRC we figured out that "there was a bugfix in hash
estimation recently and I was hoping you were older than that", but since we
are not:
PostgreSQL 9.2.4 on x86_64-pc-linux-gnu, compiled by x86_64-pc-linux-gnu-gcc
(Gentoo 4.7.2-r1 p1.6, pie-0.5.5) 4.7.2, 64-bit

...there might still be a bug around.

We compare:
http://explain.depesz.com/s/jRx
http://explain.depesz.com/s/eKE

By setting "set enable_hashjoin =3D off;" performance in our entire
application increased 30 fold in throughput, which was a bit unexpected but
highly appreciated. The result of the last query:

http://explain.depesz.com/s/AWB

What can we do to provide a bit more of information?


--=20
Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs