Re: Oids

Mikito Harakiri wrote:
> VC wrote:
> > "Mikito Harakiri" <mikharakiri_nospaum_at_yahoo.com> wrote in message
> > news:1124335766.107730.187550_at_z14g2000cwz.googlegroups.com...
> > > VC wrote:
> > >> project({}, A) = 0x, where x is either 0 or 1 depending on whether A is
> > >> empty or not (by the projection definition)
> > >
> > > Could you please supply a reference to projection definition, which is
> > > general enough to cover the empty set of attributes case?
> > >
> >
> > 1. I do not have D&D, but the definition goes like this (the relation
> > body):
> >
> > Project(A, R) = { {(a, v) }| (a, v) in t, a in A}| t in R} where R is a
>
> I'm sorry, but the curly brackets aren't balanced.

Project(A, R) = { {(a, v) | (a, v) in t, a in A} | t in R}

>
> > relation, A is a set of attributes to project over, a,v,t are attribute,
> > value, t variables respectively. Clearly, if A ={}, then Project({}, R) =
> > {{}}
>
> Therefore, Project({}, R) = {{}} for any R ?

Clearly, Project({}, R} = {} if R is empty and {{}} otherwise. Sort of obvious, no ?

>
> > 2. Ipse dixit:
> >
> > "The corresponding relation would be given by SP{ } (the projection of SP
> > over no attributes). SP{ } evaluates to either TABLE_DEE or TABLE_DUM,
> > according to the existence or not of at least one tuple in SP.
> > "
> > ( http://www.dbdebunk.com/page/page/1043196.htm )
>
> Well, I see the proposition
>
> Project({}, R) = DEE or DUM
>
> there, but fail to see the formal argument for it.

What sort of argument would you expect to see ? The result is a straightforward application of the projection definition to an argument. What sort of argument would you expect to see in favour of, say, mod_2(x) being 0 or 1 (x in I) ? Received on Thu Aug 18 2005 - 19:26:43 CEST