Date: Tue, 25 Jul 2006 11:51:05 +1000 (EST)
From: Gavin Sherry <swm@linuxworld.com.au>
To: Bruce Momjian <bruce@momjian.us>
cc: Jie Zhang <jzhang@greenplum.com>, Hannu Krosing <hannu@skype.net>,
	Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@postgresql.org,
	Luke Lonergan <LLonergan@greenplum.com>
Subject: Re: On-disk bitmap index patch
In-Reply-To: <200607250104.k6P14St25388@momjian.us>
Message-ID: <Pine.LNX.4.58.0607251149250.445@linuxworld.com.au>
References: <200607250104.k6P14St25388@momjian.us>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Mon, 24 Jul 2006, Bruce Momjian wrote:

> Jie Zhang wrote:
> > > IIRC they quoted the cardinality of 10000 as something that is still
> > > faster than btree for several usecases.
> > >
> > > And also for AND-s of several indexes, where indexes are BIG, your btree
> > > indexes may be almost as big as tables but the resulting set of pages is
> > > small.
> >
> > Yeah, Hannu points it out very well -- the bitmap index works very well when
> > columns have low cardinalities and AND operations will produce small number
> > of results.
>
> What operations on columns of low cardinality produce a small number of
> results?  That seems contradictory.

WHERE a = 1 and b = 2

a = 1 may be 5% of the table and b = 2 may be 5% of the table but their
intersection may be .001%.

Luke: the URL you sent to the bitmap slides was internal to Greenplum.
Would you be able to put them on a public site?

Thanks,

Gavin