Re: Completely wrong row estimates

From: Björn Lindqvist <bjourne(at)gmail(dot)com>
To: Scott Mead <scott(dot)lists(at)enterprisedb(dot)com>
Cc: Magnus Hagander <magnus(at)hagander(dot)net>, pgsql-general <pgsql-general(at)postgresql(dot)org>
Subject: Re: Completely wrong row estimates
Date: 2010-04-06 13:21:24
Message-ID: k2g740c3aec1004060621u6b217eefi916610ac253bec8d@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Den 6 april 2010 14.22 skrev Scott Mead <scott(dot)lists(at)enterprisedb(dot)com>:
>
> 2010/4/6 Björn Lindqvist <bjourne(at)gmail(dot)com>
>>
>> Den 5 april 2010 11.57 skrev Magnus Hagander <magnus(at)hagander(dot)net>:
>> >> Note how the planner estimates that there are 766 rows in the table
>> >> that matches the word 'tagtext'. In reality 43374 does. I've tried to
>> >> get postgres to refresh the statistics by running with
>> >> enable_statistics_target=100, running VACUUM, VACUUM FULL, VACUUM FULL
>> >> ANALYZE etc but nothing works. Postgres seem stuck with its bad
>> >> statistics and unwilling to change them. There are many other strings
>> >> that also matches tens of thousands of rows in the table which
>> >> postgres only thinks matches 766.
>
> Have you tried running :
>  'EXPLAIN ANALYZE <your query>'
> ?
>   This will show you the estimates and the actuals (for each operation) side
> by side.

Yes, see my first message where I post the EXPLAIN ANALYZE output for the query.

--
mvh Björn

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Andy Colson 2010-04-06 14:57:09 Re: Fuzzy string matching of product names
Previous Message Scott Mead 2010-04-06 12:22:00 Re: Completely wrong row estimates