From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "dcrespo" <dcrespo(at)gmail(dot)com> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Evaluate only one CASE WHEN in a select |
Date: | 2007-04-12 20:30:10 |
Message-ID: | 4956.1176409810@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
"dcrespo" <dcrespo(at)gmail(dot)com> writes:
> They are exactly the same, that's why I want to evaluate it only once
> and, depending on it, put the corresponding value into two different
> fields that must be returned, instead of evaluating once for each
> field. Any insight?
There's no solution that wouldn't cost you more than double evaluation,
for such a simple expression.
The general solution is to use two levels of SELECT:
select ..., x, x, ...
from (select ..., big-expr as x, ... from ... offset 0) ss;
You need the "offset 0" (which is otherwise a no-op) to prevent the
planner from folding the two selects into a single level and ending up
with two copies of big-expr anyway. The runtime overhead associated
with the extra plan level is about going to eat up whatever you might
save in this example, though with a seriously expensive expression
(for instance, a function that does some fairly expensive SELECT itself)
you might find it worth doing.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Scott Marlowe | 2007-04-12 20:33:59 | Re: deadlock |
Previous Message | Tom Lane | 2007-04-12 20:21:43 | Re: deadlock |