| From: | Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com> |
|---|---|
| To: | Mike Rylander <mrylander(at)gmail(dot)com> |
| Cc: | David Helgason <david(at)uti(dot)is>, Federico Balbi <fbalbi(at)cs(dot)utsa(dot)edu>, postgresql-general <pgsql-general(at)postgresql(dot)org> |
| Subject: | Null comparisons (was Re: checksum) |
| Date: | 2004-09-27 14:47:37 |
| Message-ID: | 20040927074416.C70916@megazone.bigpanda.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Mon, 27 Sep 2004, Mike Rylander wrote:
> On Sun, 26 Sep 2004 20:16:52 +0200, David Helgason <david(at)uti(dot)is> wrote:
> > On a similar note, I've found myself wanting an extended '=' operator
> > meaning
> > (a = b or (a is null and b is null))
> >
>
> Setting 'transform_null_equals' to true in postgresql.conf should do
> what you want.
Unfortunately, it probably won't. That only changes the explicit token
sequence = NULL into an IS NULL, it won't help if you're doing a=b where a
or b may be NULL.
The original does appear to be equivalent to "not(a is distinct from b)",
although I'm not sure that's necessarily easier to use than the above.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Greg Stark | 2004-09-27 15:17:38 | Re: Null comparisons (was Re: checksum) |
| Previous Message | Andrew Sullivan | 2004-09-27 14:38:14 | Re: core dump on select |