Message-ID: <4D593458.1020004@peak6.com>
Date: Mon, 14 Feb 2011 07:55:36 -0600
From: Shaun Thomas <sthomas@peak6.com>
Reply-To: sthomas@peak6.com
Organization: Peak6
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US;
	rv:1.9.2.7) Gecko/20100720 Lightning/1.0b2 Thunderbird/3.1.1
MIME-Version: 1.0
To: "gnanam@zoniac.com" <gnanam@zoniac.com>
CC: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>
Subject: Re: How to boost performance of queries containing pattern
	matching characters
References: <004601cbcc14$acfe99d0$06fbcd70$@com>
In-Reply-To: <004601cbcc14$acfe99d0$06fbcd70$@com>
Content-Type: text/plain; charset="ISO-8859-1"; format=flowed
content-transfer-encoding: 7bit

On 02/14/2011 12:59 AM, Gnanakumar wrote:

> QUERY:  DELETE FROM MYTABLE WHERE EMAIL ILIKE '%domain.com%'
> EMAIL column is VARCHAR(256).

Honestly? You'd be better off normalizing this column and maybe hiding 
that fact in a view if your app requires email as a single column. Split 
it like this:

So user@gmail.com becomes:

email_acct (user)
email_domain (gmail)
email_tld (com)

This would let you drop the first % on your like match and then 
traditional indexes would work just fine. You could also differentiate 
between domains with different TLDs without using wildcards, which is 
always faster.

I might ask why you are checking email for wildcards after the TLD in 
the first place. Is it really so common you are trying to match .com, 
.com.au, .com.bar.baz.edu, or whatever? At the very least, splitting the 
account from the domain+tld would be beneficial, as it would remove the 
necessity of the first wildcard, which is really what's hurting you.

-- 
Shaun Thomas
OptionsHouse | 141 W. Jackson Blvd. | Suite 800 | Chicago IL, 60604
312-676-8870
sthomas@peak6.com

______________________________________________

See  http://www.peak6.com/email_disclaimer.php
for terms and conditions related to this email