pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Michael Lorenz <mlorenz1@hotmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-performance@postgresql.org
Subject: Re: Query slows after offset of 100K
Date: Thu, 14 Feb 2008 19:49:22 +0000
Message-ID: <BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl> (raw)
In-Reply-To: <17159.1203016095@sss.pgh.pa.us>
References: <BAY111-W56DE7AD29405FE5C2BF1A93250@phx.gbl>
	<17159.1203016095@sss.pgh.pa.us>


Fair enough, and I did think of this as well.  However, I didn't think this was a viable option in my case, since we're currently allowing the user to randomly access the pages (so $lastkey wouldn't really have any meaning).  The user can choose to sort on object ID, name or modification time, and then go straight to any page in the list.  With 750K records, that's around 37K pages.

Maybe a better way to phrase my question is:  how can I paginate my data on 3 different keys which allow random access to any given page, and still get reasonable performance?  Should I just force the user to limit their result set to some given number of records before allowing any paginated access?  Or is it just not practical, period?

Thanks,
    Michael Lorenz

----------------------------------------
> To: mlorenz1@hotmail.com
> CC: pgsql-performance@postgresql.org
> Subject: Re: [PERFORM] Query slows after offset of 100K 
> Date: Thu, 14 Feb 2008 14:08:15 -0500
> From: tgl@sss.pgh.pa.us
> 
> Michael Lorenz  writes:
>> My query is as follows:
>> SELECT o.objectid, o.objectname, o.isactive, o.modificationtime 
>> FROM    object o 
>> WHERE  ( o.deleted = false OR o.deleted IS NULL ) 
>> AND      o.accountid = 111 
>> ORDER BY 2 
>> LIMIT 20 OFFSET 10000;
> 
> This is guaranteed to lose --- huge OFFSET values are never a good idea
> (hint: the database still has to fetch those rows it's skipping over).
> 
> A saner way to do pagination is to remember the last key you displayed
> and do something like "WHERE key> $lastkey ORDER BY key LIMIT 20",
> which will allow the database to go directly to the desired rows,
> as long as you have an index on the key.  You do need a unique ordering
> key for this to work, though.
> 
> 			regards, tom lane
> 
> ---------------------------(end of broadcast)---------------------------
> TIP 1: if posting/reading through Usenet, please send an appropriate
>        subscribe-nomail command to majordomo@postgresql.org so that your
>        message can get through to the mailing list cleanly

_________________________________________________________________
Your Future Starts Here. Dream it? Then be it! Find it at www.seek.com.au
http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fninemsn%2Eseek%2Ecom%2Eau%2F%3Ftracking%3Dsk%3Ahet%3...


view thread (8+ messages)  latest in thread

Message-ID: <BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl>
Permalink:  ../BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl/
Also on:    postgresql.org/message-id/BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-performance@postgresql.org
  Cc: mlorenz1@hotmail.com, tgl@sss.pgh.pa.us
  Subject: Re: Query slows after offset of 100K
  In-Reply-To: <BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl>
  References: <BAY111-W56DE7AD29405FE5C2BF1A93250@phx.gbl> <17159.1203016095@sss.pgh.pa.us> <BAY111-W375B12EB37C7EF3BE82BE793250@phx.gbl>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox