Erik Jones wrote:
On Mar 24, 2008, at 2:18 PM, Josh Trutwin wrote:
On Mon, 24 Mar 2008 14:02:02 -0500
Erik Jones <[EMAIL PROTECTED]> wrote:
On Mar 24, 2008, at 1:09 PM, Martijn van Oosterhout wrote:
On Mon, Mar 24, 2008 at 12:47:43PM -0500, Josh Trutwin wrote:
My code to check if an aggreg
On Mar 24, 2008, at 2:18 PM, Josh Trutwin wrote:
On Mon, 24 Mar 2008 14:02:02 -0500
Erik Jones <[EMAIL PROTECTED]> wrote:
On Mar 24, 2008, at 1:09 PM, Martijn van Oosterhout wrote:
On Mon, Mar 24, 2008 at 12:47:43PM -0500, Josh Trutwin wrote:
My code to check if an aggregate exists runs thi
On Mon, 24 Mar 2008 14:02:02 -0500
Erik Jones <[EMAIL PROTECTED]> wrote:
>
> On Mar 24, 2008, at 1:09 PM, Martijn van Oosterhout wrote:
> > On Mon, Mar 24, 2008 at 12:47:43PM -0500, Josh Trutwin wrote:
> >> My code to check if an aggregate exists runs this query:
> >>
> >> SELECT * FROM pg_catalo
On Mar 24, 2008, at 1:09 PM, Martijn van Oosterhout wrote:
On Mon, Mar 24, 2008 at 12:47:43PM -0500, Josh Trutwin wrote:
My code to check if an aggregate exists runs this query:
SELECT * FROM pg_catalog.pg_aggretate WHERE aggfnoid =
'foo'::REGPROC;
Seems to me you'd rather want the proisag
Josh Trutwin wrote:
> Is there a better way to do this? Maybe a stored proc that eats the
> error message?
Well, you can use a plpgsql function with a BEGIN/EXCEPTION/END block to
"eat" the error message. However I'm wondering whether you should be
checking the aggregate argument type(s) as wel
On Mon, Mar 24, 2008 at 12:47:43PM -0500, Josh Trutwin wrote:
> My code to check if an aggregate exists runs this query:
>
> SELECT * FROM pg_catalog.pg_aggretate WHERE aggfnoid = 'foo'::REGPROC;
Seems to me you'd rather want the proisagg column in pg_proc and forget
about pg_aggregate altogether
