"Creager, Robert S" <[EMAIL PROTECTED]> writes:
> You indicate trigger, rather than rule. Going by Momjian's book, he
> indicates that rules are "...ideal for when the action affects other
> tables." Can you clarify why you would use a trigger for this?
Primarily because there's a need to get a
due
> Cc: [EMAIL PROTECTED]
> Subject: Re: [SQL] Rule/currval() issue
>
>
> Tim Perdue <[EMAIL PROTECTED]> writes:
> > This is related to the plpgsql project I was working on
> this morning. I'm
> > trying to create a rule, so that when a row is inser
Tim Perdue <[EMAIL PROTECTED]> writes:
>> You probably should be using a trigger, not a rule at all.
> OK - so another rule like this one, is probably ill-advised as well? It seems
> a lot easier than going into the triggers:
> CREATE RULE forum_delete_agg AS
> ON DELETE TO forum
> DO UP
On Wed, Mar 14, 2001 at 01:09:18PM -0500, Tom Lane wrote:
> Tim Perdue <[EMAIL PROTECTED]> writes:
> > This is related to the plpgsql project I was working on this morning. I'm
> > trying to create a rule, so that when a row is inserted into a certain table,
> > we also create a row over in a "cou
Tim Perdue <[EMAIL PROTECTED]> writes:
> This is related to the plpgsql project I was working on this morning. I'm
> trying to create a rule, so that when a row is inserted into a certain table,
> we also create a row over in a "counter table". The problem lies in getting
> the primary key value (
Entirely untested, but how about replacing currval() in your first try
with nextval()? My theory is that the compilation of the function is
happening before the INSERT happens; therefore the sequence hasn't been
incremented yet; therefore there's no currval() for this backend
instance. If you use
This is related to the plpgsql project I was working on this morning. I'm
trying to create a rule, so that when a row is inserted into a certain table,
we also create a row over in a "counter table". The problem lies in getting
the primary key value (from the sequence) so it can be inserted in tha
