Re: How to normalize this?
Date: Sat, 4 May 2013 16:32:18 -0700 (PDT)
Message-ID: <dd044d03-5c42-4078-94e2-5d4b98c58dfa_at_googlegroups.com>
On Thursday, May 2, 2013 6:23:07 AM UTC-7, Jan Hidders wrote:
> But most
> textbooks I know are correct and clear on this and will state that the
> goal is for example a lossless-join and dependency-preserving
> decomposition.
Can you name some? Because I don't think books are clear about normalization within design, which involves generated predicates and constraint preservation. Normalization per se assumes the generated components are projections of the original, ie are suitably constrained, without addressing the management of those constraints (eg introducing FKs for lost FDs).
> 1. Apply the 3NF normalization procedure we discussed earlier
It's not clear to me what you are describing or how it relates to arguing against stopping at 3nf.
> 2. Include with each component only the FDs that generated that component
An information-equivalent FD-anomoly-free bcnf and even JD-anomoly-free 5nf schema is always possible with suitable inter-component constraints. So decomposition to 3nf as a goal is just wrong/confused.
philip Received on Sun May 05 2013 - 01:32:18 CEST