pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Marten Feldtmann <marten@feki.toppoint.de>
Cc: pgsql-sql@postgreSQL.org
Subject: Re: [SQL] index usage ... strange !?
Date: Wed, 19 Jan 2000 13:07:26 -0500
Message-ID: <10412.948305246@sss.pgh.pa.us> (raw)
In-Reply-To: <200001191746.SAA09643@feki.toppoint.de>
References: <200001191746.SAA09643@feki.toppoint.de>

Marten Feldtmann <marten@feki.toppoint.de> writes:
>  Ok, indeed it was the wrong statement, therefore I rewrote it to:

>  SELECT P3AT.AO,P3AT.AT,P3AT.AV FROM P3AT
>  WHERE
>    EXISTS
>      (SELECT B.AO FROM P3AT AS B WHERE B.AO=P3AT.AO AND ....)

>  There're indices on AO,AT,AV. Size of table about 23000 rows.

>  He uses indices only for the statement within EXISTS, which seems
> to be pretty fast but for the outer SELECT statement he uses a
> sequential scan ... which brings the the statement down from 23 ms to
> 2000ms.

>  Actually I don not understand it. The "B.AO=P3AT.AO" should create
> internally a join (?) and therefore he could use the index on AO in
> the outer select to create the result - but the seq scan seems to
> be wrong.

You are right, this would be better done as a join, but the system is
not currently smart enough to rewrite a sub-SELECT into a join.  You
have to do it yourself :-(.

I think preserving the semantics of this exactly would require outer
joins, which we don't yet have, but you could get close with DISTINCT:

SELECT DISTINCT A.AO,A.AT,A.AV FROM P3AT A, P3AT B
WHERE A.AO = B.AO AND ...;

Both of these issues are on the TODO list, and probably will get
addressed in a release or three...

			regards, tom lane



view thread (5+ messages)

Message-ID: <10412.948305246@sss.pgh.pa.us>
Permalink:  ../10412.948305246@sss.pgh.pa.us/
Also on:    postgresql.org/message-id/10412.948305246@sss.pgh.pa.us

 · 

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 from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-sql@postgresql.org
  Cc: tgl@sss.pgh.pa.us, marten@feki.toppoint.de
  Subject: Re: [SQL] index usage ... strange !?
  In-Reply-To: <10412.948305246@sss.pgh.pa.us>
  References: <200001191746.SAA09643@feki.toppoint.de> <10412.948305246@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