Jan, can you handle this TODO item?
* Evaluate INSERT rules at end of query, rather than beginning
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Is the INSERT rule re-ordering mentioned a TODO item?
>
> Darn if I know. I threw the thought out for discussion, but didn't
> see any comments. I'm not in a hurry to change it, unless there's
> consensus that we should.
>
> regards, tom lane
>
>
> >> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> >>>> I thought an INSERT rule with an UPDATE action would work on the same
> >>>> table, but that fails. Seems the rule is firing before the INSERT
> >>>> happens.
> >>
> >> Yes, a trigger is the right way to do surgery on a tuple before it is
> >> stored. Rules are good for generating additional SQL queries that will
> >> insert/update/delete other tuples (usually, but not necessarily, in
> >> other tables). Even if it worked, a rule would be a horribly
> >> inefficient way to handle modification of the about-to-be-inserted
> >> tuple, because (being an independent query) it'd have to scan the table
> >> to find the tuple you are talking about!
> >>
> >> The reason the additional queries are done before the original command
> >> is explained thus in the source code:
> >>
> >> * The original query is appended last if not instead
> >> * because update and delete rule actions might not do
> >> * anything if they are invoked after the update or
> >> * delete is performed. The command counter increment
> >> * between the query execution makes the deleted (and
> >> * maybe the updated) tuples disappear so the scans
> >> * for them in the rule actions cannot find them.
> >>
> >> This seems to make sense for UPDATE/DELETE, but I wonder whether
> >> the ordering should be different for the INSERT case: perhaps it
> >> should be original-query-first in that case.
>
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill,
Pennsylvania19026
