Message-ID: <44D8CE92.1080407@avior.ca>
Date: Tue, 08 Aug 2006 13:49:06 -0400
From: Patrice Beliveau <pbeliveau@avior.ca>
User-Agent: Thunderbird 1.5.0.4 (X11/20060614)
MIME-Version: 1.0
To: pgsql-performance@postgresql.org
Subject: Optimizing queries
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi,

I have a query that use a function and some column test to select row. 
It's in the form of:

SELECT * FROM TABLE
   WHERE TABLE.COLUMN1=something
      AND TABLE.COLUMN2=somethingelse
      AND function(TABLE.COLUMN3,TABLE.COLUMN4) > 0;

The result of the function does NOT depend only from the table, but also 
from some other tables.

Since it's long to process, I've add some output to see what's going on. 
I find out that the function process every row even if the row should be 
rejected as per the first or the second condition. Then , my question 
is: Is there a way to formulate a query that wont do all the check if it 
does not need to do it ? Meaning that, if condition1 is false then it 
wont check condition2 and that way the function will only be called when 
it's really necessary.

Thanks