Re: thinking about UPDATE

"Mikito Harakiri" <mikharakiri_at_iahu.com> wrote in message news:S2ULc.50$X73.342_at_news.oracle.com...
> "Marshall Spight" <mspight_at_dnai.com> wrote in message
> news:QIQLc.139454$a24.125173_at_attbi_s03...
> > First, the case where the key is zero attributes. In this case,
> > there is no project operation that can reduce the number of
> > attributes in the key, so again nothing to do.
>
> A key with zero attributes means that the relation is allowed to have one
> tuple only, right?
>
>
I might be missing something here, but can one have a relation variable extensionally defined with one or more attributes where the key is the empty set?

By definition, in an informal sense as it may be, a relation value is composed of a set of tuples. I had the (perhaps wrong) impression that he minimal criterion for tuple identification (and uniqueness) is that the values of the set of attributes, irregardless of key, be unique, not the other way around. The key is merely a, perhaps unnecessary, artifact and is not necessarily the consequence of the definition of a relation. If there is at least one attribute, then that attribute would be the key by definition, right?

Given my understanding, how would you ever define a relation with attributes that has a key with zero attributes (an empty set), except in the case where one has defined a relation with zero attributes (versus the key). Is this what you meant?

Follow-up conversations to this question with Jan would indicate that it is not. If I understand the implications correctly and a relation with attributes can have an empty set key, then there is an implicit empty set key component in any relation.

Set me straight someone.

• Dan