Content-Type: text/plain; charset="us-ascii"
Date: Tue, 04 Sep 2007 20:53:21 +0200
From: "Walter Mauritz" <waltermauritz@gmx.at>
Message-ID: <20070904185321.221220@gmx.net>
MIME-Version: 1.0
Subject: join tables vs. denormalization by trigger
To: pgsql-performance@postgresql.org
Content-Transfer-Encoding: 7bit

Hi,

I wonder about differences in performance between two scenarios:

Background:
Table A, ~50,000 records
Table B, ~3,000,000 records (~20 cols)
Table C, ~30,000,000 records (~10 cols)

a query every 3sec. with limit 10

Table C depends on Table B wich depends on Table A, int8 foreign key, btree index

* consider it a read only scenario (load data only in night, with time for vacuum analyze daily)
* im required to show records from Table C, but also with some (~5cols) info from Table B 
* where clause always contains the foreign key to Table A
* where clause may contain further 1-10 search parameter


Scenario A)
simply inner join Table B + C

Scenario B)
with use of trigger on insert/update I could push the required information from table B down to table C.
-> so i would only require to select from table C.


My question:
1) From your experience ... how much faster (approximately) in percent do you regard Scenario B faster than A ?

2) any other tips for such a read only scenario

Thx for any attention :-)
Walter
-- 
GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS.
Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail