Re: RA with MV attributes

Jon Heggland wrote:
> David wrote:
> > Jon Heggland wrote:
> >> David wrote:
> >>> r1 |X| r2 (Names,Cars,Colours)
> >>> bill car1,car4 red
> >>> bill car2 green
> >>> john,fred car3 red
> >>> john,fred green
> >>>
> >>> [...]
> >>> For example the
> >>> last tuple of r1 |X| r2 above doesn't imply that John and Fred
> >>> don't own any cars.
> >> So what exactly does that last tuple mean?
> >
> > It would appear that this tuple only states that John and Fred exist
> > and the colour Green exists. This information is redundant with
> > respect to the other tuples.
>
> "It would appear"? You are guessing?

No, but I'm a little cautious!

> Anyway, I assume the other tuples /do/ state more, i.e. "bill car2
> green" does not merely state that Bill exists, Car2 exists and Green
> exists? (I'll refrain from asking what exactly the meaning of this
> "exists" business is.) So the tuples have different kinds of meaning?

No, there is an underlying consistency. The role of the Colour attribute is understood to be relevant to the cars in the tuple. This role is implicit in the r2 relation and remains the case after the join. The attribute could have instead been named "CarColour". There is an implicit universal quantification here. ie

    for all cars in tuple, car is green

Since there are no cars, the tuple is vacuous.

This reminds me of true statements like

     for all elephants on the Moon, elephant wears pink panties

> What determines this meaning? The presence or absence of empty sets?
> What is simple, intuitive and elegant about this?

>


> > I don't even think the tuple should be regarded as implying John and


> > Fred don't have any green cars.  I would avoid using individual

>