Re: Optimizer wrongly picks Nested Loop Left Join

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: John A Meinel <john(at)arbash-meinel(dot)com>
Cc: Edin Kadribasic <edink(at)emini(dot)dk>, pgsql-performance(at)postgresql(dot)org
Subject: Re: Optimizer wrongly picks Nested Loop Left Join
Date: 2005-05-11 15:03:43
Message-ID: 4085.1115823823@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

John A Meinel <john(at)arbash-meinel(dot)com> writes:
> Unfortunately, because Hash Join doesn't report the number of rows
> (rows=0 always), it's hard to tell how good the estimator is.

This is only a cosmetic problem because you can just look at the number
of rows actually emitted by the Hash node's child; that's always exactly
the number loaded into the hashtable.

(But having said that, it is fixed in CVS tip.)

You may be confused though --- the Hash node is not the Hash Join node.
A zero report from Hash Join does actually mean that it emitted zero
rows.

regards, tom lane

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Manfred Koizar 2005-05-11 15:59:10 Re: Sort and index
Previous Message Bricklen Anderson 2005-05-11 14:23:20 Re: Prefetch