[Pgpool-general] stored procedures

[Tatsuo Ishii]

> I assume that if I create a stored procedure that modifies the value in
> some table, pgpool won't propagate the change to all nodes if I access
> the procedure with a select:
> 
> 	select foo_func('a', 'b');
> 
> unless I have parallel queries enabled (which I don't want to do). Is
> this true? Will it work correctly if I structure the query as an update
> instead, like this
> 
> 	update t set z=NULL where id in (select foo_func('a', 'b'));
> 
> and 'id in ...' never matches anything so that t is never updated?

This should work but something like:

/*REPLICATE*/select foo_func('a', 'b');

would be more handy(the SQL comment can be anything you like).
--
Tatsuo Ishii
SRA OSS, Inc. Japan