Re: MV Keys

From: Brian Selzer <>
Date: Thu, 02 Mar 2006 19:45:51 GMT
Message-ID: <PXHNf.26352$>

"JOG" <> wrote in message
> David Cressey wrote:
>> "Jon Heggland" <> wrote in
>> > Yes, he says "repeating groups" and "nonsimple domains" are "roughly
>> > analogous", and uses a relational-valued attribute (RVA) as the (sole)
>> > example of an attribute with a non-simple domain. But he presents
>> > normalisation as something done for convenience.
>> This is a very, very important point. If you miss the point that
>> normalization was proposed as a convenience, you are going to distort
>> the
>> entire history of the adoption of the RDM both by comercial DBMS vendors
>> and
>> by database theorists. Thanks Jon.
> [snip]
> To me this highlights the difference between 1-NF and further
> normalization. The RM is based on a functional mapping of column names
> to values. If a relation is not in 1-NF its tuples cannot be modelled
> functionally, and this being the core of RM, it is definitional as
> opposed to just mere convenience.

I guess you could make the argument that if the list itself has meaning apart from its component values, that is, the ordering, juxtaposition, or cardinality of the component elements has additional significance within the universe of discourse, such as a procedure or recipe for producing a part, for example, then you can certainly define functional dependencies where a compound attribute is either the determinant or the dependent or both. On the other hand, if it did, then I would argue that each instance has identity with respect to the database. Don't get me wrong, I still think using compound attributes is insane. I follow the KISS model, and IMO injecting compound attributes into the RM increases the level of complexity and the probability for error.

> (And as such the RM is misnamed in my opinion. With Codd's shift from
> relations to relationships, it is far more a Functional Model than a
> Relational Model in terms of mathematics, but there you go.)
> The question of MV hence distills down to the following:
> What is the advantage if any of this functional mapping (one-one, where
> an attribute maps to a single value) over the more general
> mathematical-relational mapping (where an attribute may map to many
> values)?
> Is this just so the resulting constructs can be visualised as nice neat
> regular tables? I am yet to hear any other reasons, but I don't
> preclude their existence. Does it unduly affect the algebra somehow?
> Anyone?

It certainly can. I think I've read somewhere there is a recursive relational algebra that has been designed for nested relations. Received on Thu Mar 02 2006 - 20:45:51 CET

Original text of this message