Re: Fast insertion indexes: why no developments - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: Fast insertion indexes: why no developments
Date
Msg-id 20131106005904.GJ5809@eldon.alvh.no-ip.org
Whole thread Raw
In response to Re: Fast insertion indexes: why no developments  (Greg Stark <stark@mit.edu>)
List pgsql-hackers
Greg Stark escribió:

> I think minmax indexes are more akin to bitmap indexes. They will be very
> effective for columns with low-cardinality, especially for columns that are
> very clustered. In the extreme if all the values in some regions of the
> table are the same then minmax indexes would be optimal. I wouldn't expect
> them to be very effective for a highly selective column that isn't well
> clustered.

I think clustering is more important than cardinality.  I mean you can
have a very useful minmax index on a float column, on which maybe there
are no two identical values.

I certainly don't think minmax indexes will solve all indexing problems.
In the end, they're just one more tool in your DBA toolbox.

-- 
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services



pgsql-hackers by date:

Previous
From: Greg Stark
Date:
Subject: Re: Fast insertion indexes: why no developments
Next
From: Josh Berkus
Date:
Subject: Re: List of "binary-compatible" data types