Re: How is a union different from an inheritence in ER modelling

Abhishek wrote:
> Here's a ER modelling situation.
> An investor can either be an individual, family or an institution.
> What I have seen the way it has been modelled is that there are there
> are three entity types called, INDIVIDUAL, INSTITUTION, FAMILY. Now, a
> union has been created to show that the union of these three entity
> types is the INVESTOR enitity type at the bottom.
> I would like to know how the above implementation is different from the
> implementation where we can have the INVESTOR as the superclass entity
> type and then derive three sub classes entity types called INDIVIDUAL,
> INSTITUTION, FAMILY.
In the second case there can be INVESTORS that are not either an INDIVIDUAL, INSTITUTION or FAMILY, in the second case (unless you add an extra constraint) there cannot be such INVESTORS.

> Which one is more advantageous and under what circumstances?

The first question should be which of the models describes correctly your Universe of Discourse, because they cannot be both correct.

• Jan Hidders