Message-ID: <4536698B.7010001@freesurf.fr>
Date: Wed, 18 Oct 2006 19:51:07 +0200
From: Arnaud Lesauvage <thewild@freesurf.fr>
User-Agent: Thunderbird 1.5.0.7 (Windows/20060909)
MIME-Version: 1.0
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-performance@postgresql.org
Subject: Re: Index on two columns not used
References: <45360E25.9010403@freesurf.fr> <28276.1161192717@sss.pgh.pa.us>
In-Reply-To: <28276.1161192717@sss.pgh.pa.us>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit

Tom Lane a écrit :
> Arnaud Lesauvage <thewild@freesurf.fr> writes:
>> When I join these two tables, the 2-column index of the first table is 
>> not used.
>> Why does the query planner think that this plan  is better ?
> 
> Hm, is gid by itself nearly unique in these tables?  If so, the merge
> join would get only marginally more efficient by using both columns as
> merge conditions.  Heikki's probably right to guess that the planner
> thinks it's better to use the smaller index.
> 
> However, if there are lots of duplicate gids, then it ought to have
> preferred the two-column merge ...

gid is the primary key of the first table, so absolutely unique.
Thanks for the information Tom !

--
Arnaud