On 9/29/05, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Tony Caduto <tony_caduto(at)amsoftwaredesign(dot)com> writes:
> > Please don't take this the wrong way, but don't you think even if a
> > single param is declared as OUT it should return the name of the OUT param?
> Not really, because "create function foo (in x int, out y float)" is
> supposed to have the same external behavior as "create function foo
> (in x int) returns float". I agree it's a bit of a judgment call, but
> I do not see a case for changing it.
Just my $0.02, but that seems inconsistent. In my mind, the
difference between functions with OUT params and functions that return
a RECORD (or a specific rowtype) is syntactic sugar. I'm pretty sure
that this was used to explain the implementation when it was being
discussed, in fact.
Using that logic, a functions with one OUT param would be the same as
a function returning a rowtype with only one column, and the one
column in such a rowtype certainly has a name of it's own.
GPLS -- PINES Development
In response to
pgsql-hackers by date
|Next:||From: Martin Pitt||Date: 2005-09-29 22:06:21|
|Subject: Re: horology regression test failure|
|Previous:||From: Alvaro Herrera||Date: 2005-09-29 22:02:47|
|Subject: Re: pg_total_relation_size() could not open relation with OID X|