re:help with normalization
From: edigiam <ed0946_at_ptd-dot-net.no-spam.invalid>
Date: Sun, 04 Dec 2005 02:09:53 -0600
Message-ID: <QuWdneIY1vzMOQ_eRVn_vQ_at_giganews.com>
Date: Sun, 04 Dec 2005 02:09:53 -0600
Message-ID: <QuWdneIY1vzMOQ_eRVn_vQ_at_giganews.com>
don't jump to conclusions. Just because a tuple is "all key" is no reason to think it must be split.
As far as classes, rooms and periods and no overlap; this is clearly a constraint. This is one of the things that always posed a problem in the classes that I have taught. The tables/relationships/keys do not change, what does change is the criteria for membership in one/all of the involved tables. It's the program doing the INSERT/UPDATE that has to implement this rule. The rule itself is not referential, it's semantic. Received on Sun Dec 04 2005 - 09:09:53 CET