Re: ER Model - different answers for different join paths

Walton,

I, too am baffled by that statement. Especially since,

1. Entities are not tables and are not 'joined'
2. You don't query ER models with SQL conventions, e.g. a right-hand-side conformity of TRUE for a failed subquery.
3. etc.

Ralph? Are you out there listening?

Matt Townsend

Walton Chan wrote:

> Ralph Kimball's book, 'The Data Warehouse Toolkit', indicated that one
> of
> the relational database myths was that there would be the same answer
> for
> different join paths chosen. In order words, depending on which tables
> you
> chose to join, the ER model will yield different answers, i.e. wrong
> answer.
>
> Does it mean we can no longer trust our ER modeled relational
> database?
> Unfortunately, Ralph did not give any example of how this could
> happen. In
> the real life, I have not encounter such instance with those properly
> defined ER models.
>
> Can anyone help me to understand the problem as Ralph stated (with
> examples)?
> Thanks ...
Received on Wed Aug 06 1997 - 00:00:00 CDT