Re: Worthwhile optimisation of position()?

From: Tim Allen <tim(at)proximity(dot)com(dot)au>
To: Thomas Hallgren <thomas(at)tada(dot)se>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Worthwhile optimisation of position()?
Date: 2006-03-24 06:38:13
Message-ID: 442393D5.1020505@proximity.com.au
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Thomas Hallgren wrote:
> Christopher Kings-Lynne wrote:
>
>> Is it worth allowing this:
>>
>> select count(*) from users_users where position('ch' in username) = 0;
>>
>> To be able to use an index, like:
>>
>> select count(*) from users_users where username like 'ch%';
>>
>> At the moment the position() syntax will do a seqscan, but the like
>> syntax will use an index.
>>
> You must compare position('ch' in username) to '%ch%' instead of 'ch%'
> in this respect.
>
> The position function must look for 'ch' everywhere in the string so
> there's no way it can use an index.

I think the '= 0' bit is what Chris was suggesting could be the basis
for an optimisation.

Tim

--
-----------------------------------------------
Tim Allen tim(at)proximity(dot)com(dot)au
Proximity Pty Ltd http://www.proximity.com.au/

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2006-03-24 06:48:18 Re: Worthwhile optimisation of position()?
Previous Message Thomas Hallgren 2006-03-24 06:32:41 Re: Worthwhile optimisation of position()?