On Sun, Aug 13, 2006 at 03:54:18PM -0400, Tom Lane wrote:
which is that you can use aggregate functions in the RETURNING list and
get a single-row result that is aggregated across all affected rows.
It's too late to consider implementing that for 8.2, I fear, but I think
maybe we should put it
Jim C. Nasby [EMAIL PROTECTED] writes:
Aggregates sound interesting, though I'm not sure how useful they'd
actually be. I think something like
FOR v_row IN (UPDATE ... RETURNING ...)
would be a lot more useful (if it's not already in the patch).
It's not. I thought about it for a bit
Jonah's patch for INSERT/etc RETURNING included changes to allow the
RETURNING data to be assigned to plpgsql variables, similar to plpgsql's
version of SELECT INTO. For instance, collect a serial column's
assigned value with
declare x int;
begin
insert into mytab
Tom Lane wrote:
With STRICT: you get an error unless exactly one row is affected.
This prevents the incomplete-execution problem.
BTW, some googling indicates that Oracle's equivalent PL/SQL construct
supports only the exactly-one-row case. But they have an alternative,
which is that you