pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Svetlin Manavski <svetlin.manavski@gmail.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: need to repeat the same condition on joined tables in order to choose the proper plan
Date: Tue, 14 Jun 2011 12:29:05 -0400
Message-ID: <676.1308068945@sss.pgh.pa.us> (raw)
In-Reply-To: <BANLkTikSqgoucZkO0F3v8pHCopXjR=+NYg@mail.gmail.com>
References: <BANLkTikSqgoucZkO0F3v8pHCopXjR=+NYg@mail.gmail.com>

Svetlin Manavski <svetlin.manavski@gmail.com> writes:
> I am really surprised to see that the planner needs me to explicitly specify
> the same condition twice like this:

>     SD.detectorid = SS.detectorid
>     and SD.sessionid = SS.id
>     and SD.detectorid = 1
>     and SD.sessionid >= 1001000000000::INT8 and SD.sessionid <=
> 2001000000000::INT8
>     and SS.detectorid = 1
>     and SS.id >= 1001000000000::INT8 and SS.id <= 2001000000000::INT8

The planner does infer implied equalities, eg, given A = B and B = C
it will figure out that A = C.  What you are asking is for it to derive
inequalities, eg infer A < C from A = B and B < C.  That would be
considerably more work for considerably less reward, since the sort of
situation where this is helpful doesn't come up very often.  On balance
I don't believe it's a good thing for us to do: I think it would make
PG slower on average because on most queries it would just waste time
looking for this sort of situation.

(In this example, the SS.detectorid = 1 clause is in fact unnecessary,
since the planner will infer it from SD.detectorid = SS.detectorid and
SD.detectorid = 1.  But it won't infer the range conditions on SS.id
from the range conditions on SD.sessionid or vice versa.)

			regards, tom lane



view thread (4+ messages)  latest in thread

Message-ID: <676.1308068945@sss.pgh.pa.us>
Permalink:  ../676.1308068945@sss.pgh.pa.us/
Also on:    postgresql.org/message-id/676.1308068945@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 using the --to and --cc options:
  reply via email

  To: pgsql-performance@postgresql.org
  Cc: tgl@sss.pgh.pa.us, svetlin.manavski@gmail.com
  Subject: Re: need to repeat the same condition on joined tables in order to choose the proper plan
  In-Reply-To: <676.1308068945@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