Question : Relational Database Cardinality (How is it expressed)

Greetings,

I am trying to understand how cardinality is expressed in a relational database schema as opposed to a schema which is object oriented.

For instance, when creating a few tables(relational) for a bus company where the corprate office table has a relation with branch offices table, how can the one to many relation be expressed?  Here are the two tables:

create table XYZ_Bus_Company   --Corp. Office
(
  Branch_Name varchar(20)not null,  
  Employee_SSN char(9)not null,
  primary key(Branch_Name, Employee_SSN),
  foreign key(Branch_Name) references(Branch_Offices),
  foreign key(Employee_SSN) references(Employee)
);

create table Branch_Offices
(
  Branch_Name varchar(20),
  Address varchar(30),

The XYZ Bus Co. has many branches but the Branch Office table has only one corprate office.  With the attributes I have above, how does the Branch_Offices table relate to the corprate office?

In other words, what field/attribute do I need to add to complete the relation?  I want to add the following line to the Branch_Offices table but I don't know if it would be the best way because the field would be somewhat redundant in each tuple:

Corp_Office char(11), --XYZ Company

 Thanks

John

Answer : Relational Database Cardinality (How is it expressed)

Hi John,

let's see if I can make it clear to you...

1 - Your 2 tables...

I think you want to have a "XYZ_Company" that has many "Branch_Offices", and one of them is the "Corporate_Office" (like a main office).

If it is like that, you must have 2 tables (you already have them).
You want a relationship between them:  "Company" - 1:m - "Branch_Office". How do you represent it on the tables?

From the Company side you would have to link it to a variable number of branches; can't do it with an attribute (column). But from the Branch_Office side you have to link it to 1 Company. That can be done by putting the "Company" key/id attribute in the Branch_Office table, making it a foreign key to the "Company" table.
You solved your problem! A "Branch_Office" points to a "Company".

How can you get all the offices of the company? You search the "Branch_Office" table for every row that "points" to the "Company" you want!

Now, let's go to the point where a "Company has one corporate office":
you want a relationship "Company" - 1:1 - "Branch_Office". For this, you put the "Branch_Office" key/id attribute (koreign key to Branch_Office) in the "Company" table.
Like this, any Company you insert points to a Branch_Office (the corporate office). You could also put the relationship attribute in the other side, but why would you do this when most of the Branch_Offices are not main offices, and would have the relationship attribute filled with NULL value?

As you might see, this 2 kinds of relationship are represented with an attribute. You can only figure out what kind of relationship it is, by knowing/looking the conceptual model.

Hope this helps,

PMT