From: Tom Lane <tgl@sss.pgh.pa.us>
To: Karl Wright <kwright@metacarta.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: Performance query about large tables, lots of concurrent access
Date: Tue, 19 Jun 2007 10:56:15 -0400
Message-ID: <28036.1182264975@sss.pgh.pa.us> (raw)
In-Reply-To: <4677ECA9.3000402@metacarta.com>
References: <4676DD57.7070304@metacarta.com>
<14291.1182221811@sss.pgh.pa.us>
<4677B831.8060704@metacarta.com>
<27690.1182263794@sss.pgh.pa.us>
<4677ECA9.3000402@metacarta.com>
Karl Wright <kwright@metacarta.com> writes:
> [2007-06-18 09:39:49,797]ERROR Plan: Index Scan using i1181764142395 on
> intrinsiclink (cost=0.00..14177.29 rows=5 width=253)
> [2007-06-18 09:39:49,797]ERROR Plan: Index Cond: ((jobid = $2) AND
> ((childidhash)::text = ($3)::text))
> [2007-06-18 09:39:49,797]ERROR Plan: Filter: ((childid = ($4)::text)
> AND ((isnew = ($5)::bpchar) OR (isnew = ($6)::bpchar)))
>> In this case it looks like the planner is afraid that that's exactly
>> what will happen --- a cost of 14177 suggests that several thousand row
>> fetches are expected to happen, and yet it's only predicting 5 rows out
>> after the filter.
> Well, that's odd, because the hash in question that it is using is the
> SHA-1 hash of a URL. There's essentially one row per URL in this table.
What about isnew?
Also, how many rows do *you* expect out of the query? The planner is
not going to be aware of the hashed relationship between childidhash
and childid --- it'll think those are independent conditions which they
evidently aren't. So it may be that the query really does retrieve
thousands of rows, and the rows=5 estimate is bogus because it's
double-counting the selectivity of the childid condition.
regards, tom lane
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: tgl@sss.pgh.pa.us, kwright@metacarta.com
Subject: Re: Performance query about large tables, lots of concurrent access
In-Reply-To: <28036.1182264975@sss.pgh.pa.us>
* 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