Re: Trigger and Trigger function, Part 3 - success!

On Mon, 20 Jun 2005 10:15:44 -0700, Matt Iskra wrote
> >>> Michael Fuhr <mike@fuhr.org> 06/17/05 12:34 PM >>>
> On Fri, Jun 17, 2005 at 03:05:23PM -0400, John DeSoi wrote:
> > On Jun 17, 2005, at 1:32 PM, Matt Iskra wrote:
> >
> > >Thanks Mr. Fuhr, I used the psql and it worked. The pgAdmin tool must
> > >have been having trouble with my schema qualified SQL code. The
> > >trigger and trigger function work properly now.
> >
> > I don't think there is any problem with schema qualified code in
> > pgAdmin3. At least, I tried your function on the Mac version and did
> > not have any problems.
>
> Agreed that that would be strange, especially since creating a
> schema-qualified table apparently succeeded.  Matt, if you retype
> the code in pgAdmin III, does it work?  As John mentions, you might
> have had a special character in there someplace.  We're narrowing
> down on the problem but let's not draw conclusions prematurely.
>
> Also, please don't keep changing the subject header -- that can
> mess up views of message threads.  Threads are usually easier to
> follow if all the messages have the same subject.
>
> --
> Michael Fuhr
> http://www.fuhr.org/~mfuhr/
>
> ==================
>
> Sorry about the subject header.
>
> I typed the trigger function exactly like before, I double-checked
> the code format for hidden code (its ANSI and CR/LF format), opened
> it with different editor (no weird char), and even put my editor
> into hex mode - nothing. I ran it using the pgAdmin SQL tool - error
> right at the function name - the second line in the code below. I
> run it on the psql command line and it works. Here is the source
> code again.
>
> CREATE FUNCTION
> oc.upd_last_touch_test()
> RETURNS trigger AS
> $BODY$
> BEGIN
>   NEW.last_touch := current_timestamp;
>   RETURN NEW;
> END;
> $BODY$
> LANGUAGE 'plpgsql' VOLATILE;
>
> I am, quite frankly, stumped. However this is not too important as I
> was able to create the triggers and get the test database up and
> running so the programmers can continue. I just file this sort of
> fact in my head, along with other database peculiarities. Oracle,
>  Informix and SQL Server have plenty of weirdness so I just keep
> copious notes.
>
> --Matthew
>

FWIW I copied the code into pgAdmin3 v1.2.0 on RHEL v3, created a schema
called oc hit run and voila.  No problem.

I have found that pgAdmin will mess up the schema qualified functions in a
trigger definition when viewing them.

Regards,
Keith