On Wed, 24 Sept 2025 at 04:25, Kirk Parker <khp@equatoria.us> wrote:
> I'm a big fan of maintenance-free functions. What would you think about adding the following as an alternative
triggerfunction, or as a replacement for the current function, to
> https://www.postgresql.org/docs/current/ddl-partitioning.html#DDL-PARTITIONING-INHERITANCE-EXAMPLE , item #5?
>
> CREATE OR REPLACE FUNCTION measurement_insert_trigger()
> RETURNS TRIGGER AS $$
> BEGIN
> EXECUTE format('INSERT INTO measurement_%s VALUES (NEW.*)', to_char( NEW.logdate, 'YYYYMM'));
> RETURN NULL;
> END;
> $$
> LANGUAGE plpgsql;
I've somewhat mixed feelings about that. While I do agree that it
might be a good way to code things to help prevent a DBA from a
midnight callout, I'm just not sure I'm that onboard with adding the
example. About 10 years ago, I'd likely just have agreed, but since
then we've got declarative partitioning and the legitimate use cases
for using inheritance partitioning over the newer method are very
limited. Today when I look at that page in the documents, I wonder how
we could write less about inheritance partitioning or if we could move
the inheritance section out into another page rather than having it
mixed up with the declarative partitioning sections, perhaps headed up
with a note to redirect people to the declarative partitioning
section. I fear adding your proposed example might increase the
chances of someone landing on that section if they're skimming the
page.
Overall, I'm about -0.01 on your idea. I might be in favour of it if
the inheritance section had a dedicated page.
David
