From: =?utf-8?q?=C4=BDubom=C3=ADr_Varga?= <luvar@plaintext.sk>
Reply-To: luvar@plaintext.sk
Organization: LuVar
To: pgsql-performance@postgresql.org
Subject: Some question
Date: Wed, 31 Mar 2010 17:46:38 +0200
User-Agent: KMail/1.9.10
MIME-Version: 1.0
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-Id: <201003311746.38192.luvar@plaintext.sk>

Hi, stright to my "problem":

explain
SELECT * FROM t_route
	WHERE t_route.route_type_fk =3D 1
	limit 4;

"Limit  (cost=3D0.00..0.88 rows=3D4 width=3D2640)"
"  ->  Seq Scan on t_route  (cost=3D0.00..118115.25 rows=3D538301 width=3D2=
640)"
"        Filter: (route_type_fk =3D 1)"


If I try to select constant 1 from table with two rows, it will be somethin=
g=20
like this:

explain
SELECT * FROM t_route
	WHERE t_route.route_type_fk =3D (SELECT id FROM t_route_type WHERE type =
=3D 2)
	limit 4;

"Limit  (cost=3D1.02..1.91 rows=3D4 width=3D2640)"
"  InitPlan"
"    ->  Seq Scan on t_route_type  (cost=3D0.00..1.02 rows=3D1 width=3D8)"
"          Filter: ("type" =3D 2)"
"  ->  Seq Scan on t_route  (cost=3D0.00..118115.25 rows=3D535090 width=3D2=
640)"
"        Filter: (route_type_fk =3D $0)"


=46irst query is done in about milicesonds. Second is longer than 60 second=
s.=20
t_route is bigger table (~10M rows).

I think that it seq scans whole table. Is it bug? If no, how can I achieve=
=20
that second select will not take time to end of world...

Have a nice day and thanks for any reply.

=2D-=20
Odborn=C3=ADk na v=C5=A1etko je zl=C3=BD odborn=C3=ADk. Ja sa sna=C5=BE=C3=
=ADm by=C5=A5 v=C3=BDnimkou potvrdzuj=C3=BAcou=20
pravidlo.