From: =?iso-8859-2?Q?Artur_Zaj=B1c?= <azajac@ang.com.pl>
To: "'Robert Haas'" <robertmhaas@gmail.com>
Cc: <pgsql-performance@postgresql.org>
References: <000001cb849e$1dce2720$596a7560$@ang.com.pl>
	<AANLkTim=wpohtfWkNx-=22jcagXndrfTwjvogNt9=f0D@mail.gmail.com>
In-Reply-To: <AANLkTim=wpohtfWkNx-=22jcagXndrfTwjvogNt9=f0D@mail.gmail.com>
Subject: Re: Difference between explain analyze and real execution time
Date: Mon, 15 Nov 2010 21:43:48 +0100
Message-ID: <009d01cb8505$ce738900$6b5a9b00$@ang.com.pl>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-2"
Content-Transfer-Encoding: quoted-printable
Thread-Index: AQFsDnYhYIkerMawLWBzle0OUvs2zwGxik20lCRiAyA=
Content-Language: pl


2010/11/15 Artur Zaj=B1c <azajac@ang.com.pl>:
> Why there is so big difference between explain analyze (0.710 ms) and=20
> real execution time (3309 ms)? Any suggestions?

> Could it be that it takes a long time to plan for some reason?  How =
fast
is a plain EXPLAIN?

Yes! That is it :) Planning is painful. I'm so stupid. I didn't check =
VACUUM
without ANALYZE :)

time psql  -c 'explain SELECT te.idt FROM t_positions AS te    WHERE =
te.idtr
IN (347186)'

real    0m1.087s
user    0m0.004s
sys     0m0.001s

I've changed default_statistics_target to 10000 and I think that is a
reason. When I changed it to 1000 everything seems to be ok:

time psql  -c 'explain analyze SELECT te.idt FROM t_positions AS te    =
WHERE
te.idtr IN (347186)'

real    0m0.062s
user    0m0.003s
sys     0m0.004s

Thanks.

-------------------------------------------
Artur Zajac