On Thursday 26 Jun 2003 1:40 pm, Rich Shepard wrote:
> > Matt Browne wrote:
> >
> > Basically, we have a fairly complex database, with many tables
> > (customers, etc) that need to reference addresses that are contained in
> > a generic address table.
> >
> > So:
> > customer_addresses [table]
> > supplier_addresses [table]
> > address [table]
>
> I've stumbled late onto this thread so I may have missed something
> important. However, I need to ask: are you keeping the same address in two
> different tables? That is, are customer_addresses records duplicated in
> address, and the same for supplier_addresses?
>
> If so, you've violated a normalization rule in your schema and it's no
> wonder that you can't delete all the addresses you want.
It looks like he has a centralised "address" table with "customer_addresses"
linking "customer" to "address". Likewise for "supplier".
His problem was he wanted to remove address details when nothing referred to
them any more.
--
Richard Huxton