Benedict Holland <benedict.m.holland@gmail.com> writes:
> I am scratching my head over this one. I have a basic function which
> populates a table inside of it (declared outside of the function but that
> shouldn't matter) and it requires something like
> insert into table1 (col1, ...)
> select * from foo
> and I get the error "query has no destination for result data". This is
> surely not the case as it's going directly into the insert statement which
> doesn't have output. Any idea how to get around this particularly annoying
> problem?
What PG version? Could we see the exact text of the function? I'm
wondering about aliasing problems stemming from function variables named
similarly to the target table or its columns --- newer PG versions are
brighter about that sort of conflict than older ones.
> I have to say, making the switch from t-sql to postgresql has been
> fairly easy except for function debugging.
RAISE NOTICE is the usual substitute for what I think you were doing on
t-sql. Also, I think EDB is still supporting their plpgsql debugger, so
you might consider experimenting with that.
regards, tom lane