> P> Until now, I used dataset's after insert event to set the coresponding
> P> value, but I think, trigger should be used for that purpose, right ?
>
> Wrong. Your current approach is just fine. If you need something to
> improve, I'd suggest you remove all those quoted identifiers (") and
> restructure your database to avoid composite primary keys instead.
> Unless that was just an example.

Thanks Nando for clarification.

What I wanted to improve is to get rid of the code, where I'm setting the fk
value. I made suggested correction to insert statement and got what I want.
And btw, that *was* just an example.


Regards,
Primoz