Re: GLOBAL vs LOCAL temp tables - Mailing list pgsql-hackers

From Kevin Brown
Subject Re: GLOBAL vs LOCAL temp tables
Date
Msg-id 20030417211544.GI1833@filer
Whole thread Raw
In response to Re: GLOBAL vs LOCAL temp tables  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:
> Kevin Brown <kevin@sysexperts.com> writes:
> > But as Tom pointed out, if you delete a bunch of data from a table
> > then insert a fresh set of data, but don't end up inserting much data
> > with roughly the same keys that were in the original batch of data,
> > you'll get a lot of empty areas in your index that are unused.  VACUUM
> > marks them as being available for reuse, of course, but that doesn't
> > help you unless you insert data containing values that are appropriate
> > to the unused areas.
> 
> No, you misunderstood.  That is the problem in existing releases --- but
> in CVS tip, VACUUM can actually remove unused sections from the b-tree
> and make that space available for re-use in other key ranges.

Ah, okay.

That's quite a feat, actually.  7.4 is going to be one sweet
release...



-- 
Kevin Brown                          kevin@sysexperts.com



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Should libpq's environment settings affect the session default?
Next
From: Bruce Momjian
Date:
Subject: Re: Should libpq's environment settings affect the session