| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Michael Fuhr <mike(at)fuhr(dot)org> |
| Cc: | Peter Cottingham <cottingh(at)rci(dot)rutgers(dot)edu>, pgsql-bugs(at)postgresql(dot)org |
| Subject: | Re: BUG #1555: bug in GROUP BY? |
| Date: | 2005-03-27 06:46:23 |
| Message-ID: | 2987.1111905983@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
Michael Fuhr <mike(at)fuhr(dot)org> writes:
> On Mon, Mar 21, 2005 at 07:20:04PM +0000, Peter Cottingham wrote:
>> Appears that the GROUP BY clause is not working correctly on new server
>> (GROUP BY should order the result set in ASCENDING order, NOT DESCENDING
>> order.) See below
> As far as I know, the SQL standards don't specify that GROUP BY is
> supposed to give a particular order.
> Somebody please correct me if I'm mistaken
You're not. Peter is apparently accustomed to databases that only know
how to implement GROUP BY with a sort-and-uniq kind of implementation
... that would include Postgres from a few years ago ... but there is
no such requirement in the spec.
Current PG versions are likely to implement GROUP BY with a hashing
method, which will yield a completely unsorted output. It is pure
coincidence that it happened to look like descending order in Peter's
example.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Stephan Szabo | 2005-03-27 07:02:37 | Re: BUG #1555: bug in GROUP BY? |
| Previous Message | Matthew T. O'Connor | 2005-03-27 06:24:07 | Re: BUG #1567: can't hide password with pg_autovacuum |