Message-ID: <BLU0-SMTP822F4EE03686BA05451C63AC780@phx.gbl>
Date: Tue, 14 Sep 2010 07:31:33 -0400
From: Frank Bax <fbax@sympatico.ca>
User-Agent: Thunderbird 2.0.0.23 (X11/20100309)
MIME-Version: 1.0
To: pgsql-sql@postgresql.org
Subject: Re: Question regarding indices
References: <20100911122923.75010@gmx.net> <i6g60e$345$1@news.albasani.net>
	<20100913000306.202150@gmx.net>
In-Reply-To: <20100913000306.202150@gmx.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Steve wrote:
> -------- Original-Nachricht --------
>> Datum: Sat, 11 Sep 2010 11:08:00 -0400
>> Von: Lew <noone@lewscanon.com>
>> An: pgsql-sql@postgresql.org
>> Betreff: Re: [SQL] Question regarding indices
> 
>> On 09/11/2010 08:29 AM, Steve wrote:
>>> I have a small question about the order of values in a query.
>>> Assume I have a table with the following fields:
>>>    uid INT,
>>>    data BIGINT,
>>>    hits INT
>>> And an unique index on (uid, data). I use libpq C API to query
>>> data from the table. The query is something like this:
>>> SELECT uid,data,hits FROM mytable WHERE uid=2
>>> AND data IN (2033,2499,590,19,201,659)
>>>
>>> Would the speed of the query be influenced if I would sort the data?
>> What do you mean by "sort the data"?  Which data?
>>
> I mean sorting the values in the brackets. Instead of:
> SELECT uid,data,hits FROM mytable WHERE uid=2 AND data IN (2033,2499,590,19,201,659)
> 
> I would then send this here:
> SELECT uid,data,hits FROM mytable WHERE uid=2 AND data IN (19,201,590,659,2033,2499)
> 
> Off course this is a small dataset but the query usually has thousands of elements and not only the above 6 elements.


If there will be thousands; why not create a temp table containing these 
values then join to table - might that be faster?