Re: How to normalize this?
From: Erwin <e.smout_at_myonline.be>
Date: Mon, 6 May 2013 11:27:36 -0700 (PDT)
Message-ID: <d4b55d1f-de0d-41b8-bed2-f252fb1a3285_at_googlegroups.com>
Date: Mon, 6 May 2013 11:27:36 -0700 (PDT)
Message-ID: <d4b55d1f-de0d-41b8-bed2-f252fb1a3285_at_googlegroups.com>
Op maandag 6 mei 2013 17:19:36 UTC+2 schreef nvitac..._at_gmail.com het volgende:
> Finally I ask them: does this decomposition preserve FDs? (“yes, by construction”). Is C -> B implied by F? (“yes, by transitivity”). Is C -> B preserved or not in the decomposition? (panic) :)
>
>
>
> Nicola
You've added a bit of another dimension.
c->b is not an explicitly stated FD (not member of the canonical cover ???) but a "derivable" one.
If so, then what about the "derived" FDs that express what the candidate keys are in a relation schema (say, ae->abcde) ? What is _their_ preservation status in any decomposition ? Received on Mon May 06 2013 - 20:27:36 CEST