Brian Dick wrote:
> On 28 Sep 2004 08:11:44 -0700, Tony Andrews wrote:
> > Of course, having the CID in the A table is a viable work-around,
but
> > that doesn't mean there was a relational design flaw before.
>
> You need to brush up on your normalization rules. Rule 3 states that
> columns must be dependent on nothing but the key. Your ATXT column is
not
> solely dependent on AID, so your design is not normalized.
Wrong. ATXT is solely dependent on AID, it cannot be determined from
CID. But there is an additional constraint that ATXT must be unique
w.r.t. the CID for some reason. If you still think I'm wrong, then
please show the NORMALIZED database design that resolves the issue.
Not that it was MY column or design anyway!
Received on Wed Sep 29 2004 - 12:21:38 CDT
