pg.ddx.io pgsql-performance@postgresql.org mailing list archive help / color / mirror / Atom feed From: Andrzej Zawadzki <zawadaa@gmail.com>
To: postgres performance list <pgsql-performance@postgresql.org>
Subject: Re: Why query plan is different?
Date: Mon, 10 Oct 2016 23:17:09 +0200
Message-ID: <8aa28a9c-95c1-a8bb-e542-e7e4fd499c27@gmail.com> (raw )
In-Reply-To: <bb1d234e-62ee-70e8-44ef-9b0919ec1268@wp.pl >
References: <bb1d234e-62ee-70e8-44ef-9b0919ec1268@wp.pl >
List-Unsubscribe: <mailto:majordomo@postgresql.org?body=unsub%20pgsql-performance>
<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 10.10.2016 17:31, Andrzej Zawadzki
wrote:<br>
</div>
<blockquote cite="mid:bb1d234e-62ee-70e8-44ef-9b0919ec1268@wp.pl"
type="cite">
<meta http-equiv="content-type" content="text/html; charset=utf-8">
Hi,<br>
Today, I noticed strange situation:<br>
<br>
The same query run on different servers has very different plan:<br>
<br>
Q: SELECT b.* FROM kredytob b WHERE pesel = '22222222222' ORDER
BY b.id DESC LIMIT 1 <br>
<br>
Slow plan:<br>
<br>
"Limit (cost=0.43..28712.33 rows=1 width=4) (actual
time=2574.041..2574.044 rows=1 loops=1)"<br>
" Output: id"<br>
" Buffers: shared hit=316132 read=110001"<br>
" -> Index Scan Backward using kredytob_pkey on
public.kredytob b (cost=0.43..3244444.80 rows=113 width=4)
(actual time=2574.034..2574.034 rows=1 loops=1)"<br>
" Output: id"<br>
" Filter: (b.pesel = '22222222222'::bpchar)"<br>
" Rows Removed by Filter: 433609"<br>
" Buffers: shared hit=316132 read=110001"<br>
"Planning time: 0.414 ms"<br>
"Execution time: 2574.139 ms"<br>
<br>
<br>
Fast plan:<br>
"Limit (cost=115240.66..115240.66 rows=1 width=4) (actual
time=463.275..463.276 rows=1 loops=1)"<br>
" Output: id"<br>
" Buffers: shared hit=14661 read=4576"<br>
" -> Sort (cost=115240.66..115240.94 rows=112 width=4)
(actual time=463.271..463.271 rows=1 loops=1)"<br>
" Output: id"<br>
" Sort Key: b.id DESC"<br>
" Sort Method: top-N heapsort Memory: 25kB"<br>
" Buffers: shared hit=14661 read=4576"<br>
" -> Index Scan using
kredytob_pesel_typkred_opclass_idx on public.kredytob b
(cost=0.43..115240.10 rows=112 width=4) (actual
time=311.347..463.183 rows=5 loops=1)"<br>
" Output: id"<br>
" Index Cond: (b.pesel = '22222222222'::bpchar)"<br>
" Buffers: shared hit=14661 read=4576"<br>
"Planning time: 0.383 ms"<br>
"Execution time: 463.324 ms"<br>
<br>
Data is almost equal - "slow" has a few more rows in table.
("Fast" is a copy from 1 am today).<br>
Why runtime is slower?<br>
</blockquote>
<br>
I made another INDEX, without opclass:<br>
<br>
CREATE INDEX kredytob_pesel_typkred_idx<br>
ON public.kredytob<br>
USING btree<br>
(pesel COLLATE pg_catalog."default", typkred);<br>
<br>
after that: analyze kredytob;<br>
<br>
And now:<br>
"Limit (cost=333.31..333.31 rows=1 width=4) (actual
time=0.100..0.102 rows=1 loops=1)"<br>
" Output: id"<br>
" Buffers: shared hit=8"<br>
" -> Sort (cost=333.31..333.59 rows=114 width=4) (actual
time=0.095..0.095 rows=1 loops=1)"<br>
" Output: id"<br>
" Sort Key: b.id DESC"<br>
" Sort Method: top-N heapsort Memory: 25kB"<br>
" Buffers: shared hit=8"<br>
" -> Index Scan using kredytob_pesel_typkred_idx on
public.kredytob b (cost=0.43..332.74 rows=114 width=4) (actual
time=0.046..0.065 rows=5 loops=1)"<br>
" Output: id"<br>
" Index Cond: (b.pesel = '22222222222'::bpchar)"<br>
" Buffers: shared hit=8"<br>
"Planning time: 0.438 ms"<br>
"Execution time: 0.154 ms"<br>
<br>
So, what is a reason that "SLOW" server doesn't like opclass index?<br>
<br>
-- <br>
Andrzej<br>
</body>
</html>
view thread (7+ messages) latest in thread
Message-ID: <8aa28a9c-95c1-a8bb-e542-e7e4fd499c27@gmail.com>
Permalink: ../8aa28a9c-95c1-a8bb-e542-e7e4fd499c27@gmail.com/
Also on: postgresql.org/message-id/8aa28a9c-95c1-a8bb-e542-e7e4fd499c27@gmail.com
copy link · copy postgr.es
reply Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: zawadaa@gmail.com
Subject: Re: Why query plan is different?
In-Reply-To: <8aa28a9c-95c1-a8bb-e542-e7e4fd499c27@gmail.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox