Re: B-tree performance improvements in 8.x - Mailing list pgsql-general

From Tom Lane
Subject Re: B-tree performance improvements in 8.x
Date
Msg-id 903.1139350409@sss.pgh.pa.us
Whole thread Raw
In response to Re: B-tree performance improvements in 8.x  (jao@geophile.com)
Responses Re: B-tree performance improvements in 8.x  (Dick Kniep <dick@kniep.nl>)
List pgsql-general
jao@geophile.com writes:
> Quoting Tom Lane <tgl@sss.pgh.pa.us>:
>> In 8.0, the descent code can do
>> either "first entry >= X" or "first entry > X", and the positioning
>> rules never need to step more than one entry to locate the desired
>> starting position (details left as exercise for the reader).

> What about a compound index (x, y), in which there are many y values
> for a given x? My query is "... WHERE x = ? and y = ?".

Doesn't really matter whether the key is simple or compound --- all
that matters is whether you have multiple entries that are "equal" to
the boundary value.

            regards, tom lane

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Why pg_hba not in table?
Next
From: Dick Kniep
Date:
Subject: Re: B-tree performance improvements in 8.x