Re: Bitmap indexes? - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: Bitmap indexes?
Date
Msg-id 200203140207.g2E27fm18562@candle.pha.pa.us
Whole thread Raw
In response to Bitmap indexes?  (Greg Copeland <greg@CopelandConsulting.Net>)
Responses GIST  ("Christopher Kings-Lynne" <chriskl@familyhealth.com.au>)
List pgsql-hackers
Greg Copeland wrote:

Checking application/pgp-signature: FAILURE
-- Start of PGP signed section.
> One of the reasons why I originally stated following the hackers list is
> because I wanted to implement bitmap indexes.  I found in the archives,
> the follow link, http://www.it.iitb.ernet.in/~rvijay/dbms/proj/, which
> was extracted from this,
>
http://groups.google.com/groups?hl=en&threadm=01C0EF67.5105D2E0.mascarm%40mascari.com&rnum=1&prev=/groups%3Fq%3Dbitmap%2Bindex%2Bgroup:comp.databases.postgresql.hackers%26hl%3Den%26selm%3D01C0EF67.5105D2E0.mascarm%2540mascari.com%26rnum%3D1,
archivethread.
 
> 
> At any rate, that was some number of months ago.  I've started looking
> at the results posted from their bitmap GiST efforts and found that they
> were being tested rather poorly to be of real value (I also found it
> annoying that the project seems to quote other people's work without
> giving credit).  Nonetheless, I thought I'd post to find out if anyone
> feels there is still a need for this?  That is, I'm not really sure that
> data warehousing or DDS systems are currently very common with Postgres.
> 
> If the group here still see value in adding various types of bitmap
> support, can someone please point me to some documentation.  I had
> several bookmarked but lost then when X crashed.  Anything that outlines
> cache strategy, index support, am overview, and any other documentation
> that would help excel my understanding of the code as well as the
> various structure relationships would be wonderful?

The only thing I know is that there is discussion of bitmap indexes on
the TODO list linked to from the 'bitmap index' item.  I also remember
that the intarray code in /contrib sort of simulates bitmapped indexes,
or something like that.  :-)

> Oh yes, one last question, is the required method for adding index
> support via GiST?  I ask because it seems to me that inserts could be
> exceptionally expensive, though as usual, I still have more to look at.

I think we would recommend GIST because it is easier.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026
 


pgsql-hackers by date:

Previous
From: Vince Vielhaber
Date:
Subject: Re: insert statements
Next
From: Peter Eisentraut
Date:
Subject: Re: insert statements