Re: How to normalize this?
From: Erwin <e.smout_at_myonline.be>
Date: Tue, 7 May 2013 07:20:58 -0700 (PDT)
Message-ID: <faf27808-5bad-410e-94e4-fdc5a9b7ffa2_at_googlegroups.com>
Date: Tue, 7 May 2013 07:20:58 -0700 (PDT)
Message-ID: <faf27808-5bad-410e-94e4-fdc5a9b7ffa2_at_googlegroups.com>
Op dinsdag 7 mei 2013 11:54:47 UTC+2 schreef nvitac..._at_gmail.com het volgende:
> The way I see it, the decomposition is not the end of the story. At that point, a much more complex process starts: you must discover how to keep the whole database schema together (in the most basic case, this means you have to discover which INDs hold). This is where normalization theory falls short, in my opinion. For example:
>
>
>
> S(A,B,C), {A -> BC}
>
> R(A,B), {}
>
>
>
> are each in BCNF. Now, suppose that you determine that R[A,B] ⊆ S[A,B]. Does that introduce any redundancy? Does it change the keys of R? Is it still a good database schema (in some sense to be defined)?