Re: Identifying Relation vs Non-Identifying?

From: <>
Date: Fri, 19 Aug 2005 00:19:31 -0700
Message-ID: <UKsNe.5618$Us5.4226_at_fed1read02>

Here's another cut at the definition of weak entity. An example might be dependents which would not typically "stand alone" in an HR database without employees. The PK of dependents would not necessarily contain the employee PK, although there would be a FK to employees. Dependents, then would be a weak entity, employees a strong entity. This relationship is non-identifying.

An ID-Dependent weak entity on the other hand would need to include the PK of the other entity. It cannot exist without the associated strong entity. Example: dorm room and dorm. A dorm room BH12A is existence dependent on the dorm BH. This is an identifying relationship. CASE tools often use this fact to migrate the PK to the child table.

Dan D'Urso

"Jan Hidders" <> wrote in message

> Dan wrote:
> > On 7/5/2005 10:57 AM, Dan wrote:
> >
> >> On 7/1/2005 6:27 PM, Jan Hidders wrote:
> >>
> >>> Dan wrote:
> >>>>
> >>>> Can someone explain to be me the difference to between an
> >>>> identifying relation and a non-identifying relation? [...]
> >>>
> >>> Loosely speaking an identifying relationship is a relationship that
> >>> is necessary to identify a certain entity that plays a role in it.
> >>> [...]
> >
> > Can one just say that a weak entity is a tuple whose primary key relies
> > on the primary key of another entity? And that this type of
> > relationship and only this type is then defined as a weak relation?
> Yes. That is how you usually map such a construct in an ER diagram to
> relations. The PK of the weak entity will consist of the local
> identifying attrubutes plus the PK of the entity it depends upon. Things
> can get a bit more complicated if the identifying relationship is a
> many-to-many relationship, but that situation is very rare and some even
> disallow it.
> -- Jan Hidders
Received on Fri Aug 19 2005 - 09:19:31 CEST

Original text of this message