From: | Thomas Lockhart <lockhart(at)alumni(dot)caltech(dot)edu> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Malcolm Beattie <mbeattie(at)sable(dot)ox(dot)ac(dot)uk>, pgsql-hackers(at)postgreSQL(dot)org |
Subject: | Re: [HACKERS] Phantom row from aggregate in self-join in 6.5 |
Date: | 1999-07-23 03:33:36 |
Message-ID: | 3797E290.4B8E0ABD@alumni.caltech.edu |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
> > create table foo (a int);
> > select t1.a, count(*) from foo t1, foo t2 group by t1.a;
> > I get
> > a|count
> > -+-----
> > | 0
> > (1 row)
> > instead of zero rows.
> It's not a bug, it's a feature ... or at least there are some around
> here who claim that the behavior is OK. I think they're wrong, but
> if you want it changed you'll need to cite chapter and verse from the
> SQL92 standard, not just assert that Informix does it differently.
I don't recall which way I argued before (in fact, I don't recall this
particular example), but I do remember arguing (with righteous
conviction) that the query
select count(*) from foo;
should return a single row containing a zero value. Did we infer from
that some behavior for "group by" (I can't recall any)? istm, at least
today, that the behavior for the group-by is wrong, but we'd better
not change the behavior of my example query...
- Thomas
--
Thomas Lockhart lockhart(at)alumni(dot)caltech(dot)edu
South Pasadena, California
From | Date | Subject | |
---|---|---|---|
Next Message | Thomas Lockhart | 1999-07-23 04:21:02 | Re: [HACKERS] RFC: remote tables feature |
Previous Message | Philip Warner | 1999-07-23 03:31:08 | Re: [HACKERS] Phantom row from aggregate in self-join in 6.5 |