Re: Ternary Relationship cardinality

From: Jan Hidders <hidders_at_gmail.com>
Date: 19 Sep 2006 04:47:15 -0700
Message-ID: <1158666435.167059.327550_at_b28g2000cwb.googlegroups.com>


miklesw_at_gmail.com wrote:
> > Ah, wait, now I understand your sub/superclass annotation. You say that
> > Child is a subclass of Parent (btw. letting both be a subclass of Dog
> > or whatever might be more logical) and a dog is in the class Child iff
> > it has parents.In that case there is no problem.
>
> I don't think I can generalize both in 1 class... the actual heirarchy
> is abit more complex than the parent/child example.
>
> Person
> IS A
> CoolerPerson(who can take part in breeding)
> IS A
> ChildPerson OR EvenCoolerPerson

It depends a bit on which ER dialect you happen to be using. Strictly speaking, if you specify that A IS-A C and B IS-A C this only means that every A is a C, and every B is a C. So it does not mean that A and B are disjoint (i.e., there can be entities in both A and B) or that every C is either an A or a B. Usually there are special combined isa-edges to denote the latter two constraints. But, again, this may depend on the dialect your instructor / professor is using, so be sure to check the IS-A semantics in the notation you have to use.

  • Jan Hidders
Received on Tue Sep 19 2006 - 13:47:15 CEST

Original text of this message