Re: update before drop causes OID problems in transaction?

On Sat, 18 Mar 2006, Tom Lane wrote:

>> Is that the case whether the triggers are executed or not?
>
> If the trigger function hasn't ever been executed in the current
> session, it wouldn't have a cached plan ... but I suspect you meant
> "if it hasn't been executed in the current transaction", and that
> doesn't help.

well, actually, I sort of meant both, though of course I'd prefer the same 
transaction.  I was actually calling this script via psql -f so it would only 
be that transaction in that one session.

>> However, we drop that trigger before
>> dropping credit_card_audit, so I'd think that would be ok.
>
> IIRC you'd have to drop the underlying plpgsql function, not only
> the trigger object that connects the function to a table.  We cache
> stuff with respect to the function.

I'll try that and see if that makes the difference, since we're recreating 
(create or replace) that function in that transaction anyway, but perhaps that 
needs to happen before the update.

-- 
Jeff Frost, Owner     <jeff@frostconsultingllc.com>
Frost Consulting, LLC     http://www.frostconsultingllc.com/
Phone: 650-780-7908    FAX: 650-649-1954