Hi,

> Thanks for acknowledging that this is a problem Ann.
>
> I feel a lot better knowing that's recognised as something 'bad'.
> Gives you confidence that the developers are receptive.

Indeed.

=> Ann

Let's suppose CHECK constraints now operate in the system transaction
context, don't we still miss something to cover all the possibilities, in
that CHECK constraints are not triggered for deletions.

For example, if you want to implement custom RI, master table must search
the detail table for detail records, and if any are found either cascade
delete them or trigger an error.

Regards.

--
Jerome