Re: Aggregation is Inverse of Join

Marshall wrote:
> On Jan 15, 8:41 am, "Walt" <wami..._at_verizon.net> wrote:
> > "Marshall" <marshall.spi..._at_gmail.com> wrote in
> >
> > > Well, I've run out of steam. More later.
> >
> > > Marshall
> >
> > > PS. I told you I wasn't organized.
>
> > Interesting. It's going to take me weeks to get it.
> >
> > In the meantime, what is the inverse of Project?
>
> I think the usual term is "extend."
>
> In the relational lattice, project is just a specialization of the
> inner union, with one of the operands having cardinality zero.

"Inverse" in what sense? Among the 4 operators:

1. natural join
2. inner union
3. relational division
4. anti join

it is the relational division that is inverse of join, not the "anti join"! Also, I can't quite make sense of what happens with the relations attributes:

1. natural join -- union of attributes
2. inner union -- intersection of attributes
3. relational division -- symmetric difference of attributes
4. anti join -- attributes of A

The picture is not symmetric at all! Received on Mon Jan 15 2007 - 20:40:45 CET