From: | Christopher Browne <cbbrowne(at)acm(dot)org> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: CUBE, ROLLUP, GROUPING SETS? |
Date: | 2006-10-29 03:17:54 |
Message-ID: | 87slh7u7x9.fsf@wolfe.cbbrowne.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Centuries ago, Nostradamus foresaw when nicolas(dot)barbier(at)gmail(dot)com ("Nicolas Barbier") would write:
> 2006/10/28, Robert Treat <xzilla(at)users(dot)sourceforge(dot)net>:
>
>> On Tuesday 10 October 2006 15:19, stig erikson wrote:
>>
>>> Are there any plans to implement CUBE, ROLLUP and/or GROUPING SETS in
>>> future PostgreSQL versions? I could not find any info on the TODO-page.
>>
>> I've heard people mention it, but no one has ever come up with a
>> solid proposal or patch. Depending on what your doing there are
>> ways to simulate those things with more standard sql.
>
> CUBE, ROLLUP and GROUPING SETS are at least part of SQL 2003.
> Supporting them would probably also include optimized ways to
> execute them, like calculating all aggregations in one run for
> ROLLUP (instead of doing multiple scans).
I'm sure a good implementation would include such things.
I'm not sure it's anywhere near trivial to do so; it may be nontrivial
to efficiently do all those aggregates at once. We'll presumably
discover this if and when someone begins work on it...
--
"cbbrowne","@","gmail.com"
http://linuxdatabases.info/info/postgresql.html
"In the long run every program becomes rococo - then rubble."
-- Alan Perlis
From | Date | Subject | |
---|---|---|---|
Next Message | David Goodenough | 2006-10-29 12:12:33 | deadlock detected messages |
Previous Message | Joshua D. Drake | 2006-10-29 02:17:30 | Re: Scalability |