> Will the constraint take care of the Cascade? I have not declared
> any triggers to do this. Are no other settings on the dataset
> necessary to cause the cascade?

The declaration should take care of cascading automatically. What I
was referring to about checking manually defined triggers to be sure
you did not have something overriding the MEMBERID field or preventing
it from changing.

> I agree that declaring a 'dummy' primary key would solve all
> problems, but something reminds me that the previous is the better
> way of doing things(guess I'm a sucker for punishment).

If updates are rare then it should be acceptable (and more efficient
while ever it does not need updating). It depends a little on how far
through the database the identity may spread. For example in an
eariler DOS based DBMS I used to relate on the account number, the
problem was that the account number was used in every-other table. So
on the few times that a change was required it was a real PITA.


If you cannot find a problem in manually defined triggers then I
suggest you try dropping and re-applying the constraint to see if that
makes any difference. Otherwise I am as puzzled as you are.

It does not make any difference what help I can offer, but you did not
mention what server version you are using.

--
Geoff Worboys - TeamIBO
Telesis Computing