Re: Normalization: primary key or any key

From: Bert Blink <bblink_at_pcug.org.au>
Date: Mon, 18 Nov 2002 07:52:35 +1100
Message-ID: <h20gtuc8mfipf2rttqpr9rmr0l1qjuug9t_at_4ax.com>


The proper term should be "Candidate Key(s)". See Elmasri & Navathe 3rd Edition "Fundamentals of Database Systems". It has become common in poorer textbooks to just refer to 'Primary Key" (one of many possible "Candidate Key(s)").

On 17 Nov 2002 09:23:25 -0800, sicotom_at_eresmas.com (Juan Pardillos) wrote:

>Hi,
>
>I've read several texts about normalization. Sometimes, the "primary
>key" term is used to establish the normal forms constraints. However,
>other times only "key" is used. I think the second one is the right
>one, as it doesn't matter if the determinant of a functional
>dependency is the primary key or any other possible key in the
>relation.
>
>Am I right?. Please, clarify me this little question I have.
>
>Thanks in advance

Bert Blink
PO Box 720
Belconnen ACT 2616 Australia
+61-02-62541341 (h)
  0427 541344 (mobile) Received on Sun Nov 17 2002 - 21:52:35 CET

Original text of this message