| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Randall Lucas <rlucas(at)tercent(dot)net> |
| Cc: | Rod Taylor <rbt(at)rbt(dot)ca>, pgsql-sql(at)postgresql(dot)org |
| Subject: | Re: merge-join for domain with underlying type text |
| Date: | 2003-05-11 03:19:49 |
| Message-ID: | 10881.1052623189@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-sql |
Randall Lucas <rlucas(at)tercent(dot)net> writes:
> Hackers, why is this? It seems to me that since everything else (or
> everything I've run into so far, up to the full outer joins) is
> implicitly the same for a simple domain and its underlying base type,
> that it would make sense if this, too, Just Worked.
It does Just Work ... in CVS tip. I can't make it magically work in
7.3 --- at least not without back-patching a lot of stuff that hasn't
been through beta-testing yet.
The problem is that mergejoin in existing releases will only work on
plain "Var = Var" clauses. Your domain case doesn't look like that
because of the runtime cast operators.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Matthew Horoschun | 2003-05-11 04:11:01 | Knowing when it is safe to call currval() |
| Previous Message | Randall Lucas | 2003-05-11 00:53:21 | Re: merge-join for domain with underlying type text |