Date: Tue, 05 Feb 2008 11:14:39 -0500
From: "Jignesh K. Shah" <J.K.Shah@Sun.COM>
Subject: Re: Benchmark Data requested
In-reply-to: <1202225207.4252.718.camel@ebony.site>
To: Simon Riggs <simon@2ndquadrant.com>
Cc: Richard Huxton <dev@archonet.com>,
	Dimitri Fontaine <dfontaine@hi-media.com>,
	pgsql-performance@postgresql.org, Greg Smith <gsmith@gregsmith.com>
Message-id: <47A88B6F.8070905@sun.com>
Organization: Sun Microsystems
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=ISO-8859-1
Content-transfer-encoding: 7BIT
References: <1202150236.4252.521.camel@ebony.site>
	<Pine.GSO.4.64.0802041500460.799@westnet.com>
	<47A797E5.4000507@sun.com>
	<200802051506.51581.dfontaine@hi-media.com>
	<1202221496.4252.680.camel@ebony.site> <47A87618.1090203@archonet.com>
	<1202223268.4252.707.camel@ebony.site> <47A87B2E.7050205@archonet.com>
	<1202225207.4252.718.camel@ebony.site>
User-Agent: Thunderbird 2.0.0.4 (X11/20070622)

One of the problems with "Empty Table optimization" is that if there are 
indexes created then it is considered as no longer empty.

Commercial  databases have options like "IRRECOVERABLE" clause along 
with DISK PARTITIONS and CPU partitions for their bulk loaders.

So one option turns off logging, disk partitions create multiple 
processes to read various lines/blocks from input file and other various 
blocks to clean up the bufferpools to disk and CPU partitions to process 
the various blocks/lines read for their format and put the rows in 
bufferpool if successful.

Regards,
Jignesh

Simon Riggs wrote:
> On Tue, 2008-02-05 at 15:05 +0000, Richard Huxton wrote:
>
>   
>>> Only by locking the table, which serializes access, which then slows you
>>> down or at least restricts other options. Plus if you use pg_loader then
>>> you'll find only the first few rows optimized and all the rest not.
>>>       
>> Hmm - the table-locking requirement is true enough, but why would 
>> pg_loader cause problems after the first few rows?
>>     
>
> It runs a stream of COPY statements, so only first would be optimized
> with the "empty table optimization".
>
>