Date: Wed, 19 Aug 2009 08:27:52 +0200
To: "Karl Denninger" <karl@denninger.net>, "Tom Lane" <tgl@sss.pgh.pa.us>
Subject: Re: [SQL] SQL Query Performance - what gives?
From: =?utf-8?Q?Pierre_Fr=C3=A9d=C3=A9ric_Caillau?= =?utf-8?Q?d?=
	<lists@peufeu.com>
Cc: "Kevin Grittner" <Kevin.Grittner@wicourts.gov>,
	pgsql-performance@postgresql.org, pgsql-sql@postgresql.org
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: <4A8B04C7.6060702@denninger.net>
	<4A8AC75C0200002500029C8A@gw.wicourts.gov>
	<4A8B1606.2040209@denninger.net>
	<4A8ADDEA0200002500029CA3@gw.wicourts.gov>
	<4A8B334F.1010801@denninger.net>
	<9800.1250647338@sss.pgh.pa.us> <4A8B67DD.2020608@denninger.net>
Content-Transfer-Encoding: 7bit
Message-ID: <op.uywb8qyucke6l8@soyouz>
In-Reply-To: <4A8B67DD.2020608@denninger.net>
User-Agent: Opera Mail/9.62 (Linux)


> The bitmask allows the setting of multiple permissions but the table
> definition doesn't have to change (well, so long as the bits fit into a
> word!)  Finally, this is a message forum - the actual code itself is
> template-driven and the bitmask permission structure is ALL OVER the
> templates; getting that out of there would be a really nasty rewrite,
> not to mention breaking the user (non-developer, but owner)
> extensibility of the current structure.
>
> Is there a way to TELL the planner how to deal with this, even if it
> makes the SQL non-portable or is a hack on the source mandatory?

	You could use an integer array instead of a bit mask, make a gist index  
on it, and instead of doing "mask & xxx" do "array contains xxx", which is  
indexable with gist. The idea is that it can get much better row  
estimation. Instead of 1,2,3, you can use 1,2,4,8, etc if you like. you'd  
probably need a function to convert a bitmask into ints and another to do  
the conversion back, so the rest of your app gets the expected bitmasks.  
Or add a bitmask type to postgres with ptoper statistics...