Full Functional Dependency in Database Normalization

Full Functional Dependency is the Same as Second Normal Form

A full functional dependency is a state of database normalization that equates to the normalization standard of Second Normal Form (2NF). In brief, this means that it meets the requirements of First Normal Form (1NF), and all non-key attributes are fully functionally dependent on the primary key.

This is not as complicated as it may sound. Let's look at this in more detail.

Summary of First Normal Form

Before a database can be fully functionally dependent, it must first comply with First Normal Form.

All this means that that each attribute must hold a single, atomic value. 

For example, the following table does not comply with 1NF, because the employee Tina is linked to two locations, both of them in a single cell:

First Normal Form Non-Compliance
EmployeeLocation
JohnLos Angeles
TinaLos Angeles, Chicago

 

 

 

Allowing this design could negatively impact data updates or entries. To ensure 1NF compliance, rearrange the table so that all attributes (or column cells) hold a single value: 

First Normal Form Compliance
EmployeeLocation
JohnLos Angeles
TinaLos Angeles
TinaChicago

 

 

 

 

But 1NF is still not enough to avoid problems with the data.

How 2NF Works to Ensure Full Dependency

To be fully dependent, all non candidate key attributes must depend on the primary key. (Remember, a candidate key attribute is any key (for example, a primary or foreign key) used to uniquely identify a database record.

Database designers use a notation to describe the dependent relationships between attributes:

If attribute A determines the value of B, we write this A -> B — meaning that B is functionally dependent on A. In this relationship, A determines the value of B, while B depends on A.

For example, in the following Employee Departments table, EmployeeID and DeptID are both candidate keys: EmployeeID is the table's primary key while DeptID is a foreign key.

Any other attribute — in this case, EmployeeName and DeptName — must depend on the primary key to obtain its value.

Employee Departments
EmployeeIDEmployeeNameDeptIDDeptName
Emp1JohnDept001Finance
Emp2TinaDept003Sales
Emp3CarlosDept001Finance

 

 

 

 

In this case, the table is not fully dependent because, while the EmployeeName depends on the primary key EmployeeID, the DeptName depends instead on the DeptID. This is called partial dependency.

To make this table conform to 2NF, we need to separate the data into two tables:

Employees
EmployeeIDEmployeeNameDeptID
Emp1JohnDept001
Emp2TinaDept003
Emp3CarlosDept001

 

 

 

 

We remove the DeptName attribute from the Employees table and create a new table Departments:

Departments
DeptIDDeptName
Dept001Finance
Dept002Human Resources
Dept003Sales

 

 

 

 

Now the relations between the tables are fully dependent, or in 2NF.

Why Full Dependency is Important

Full dependency between database attributes helps ensure data integrity and avoid data anomalies.

For example, consider the table in the section above that adheres only to 1NF. Here it is, again:

First Normal Form Compliance
EmployeeLocation
JohnLos Angeles
TinaLos Angeles
TinaChicago

 

 

 

 

Tina has two records. If we update one without realizing that there are two, the result would be inconsistent data.

 

Or, what if we want to add an employee to this table, but we don't yet know the Location? We might be disallowed to even add a new employee if the Location attribute does not allow NULL values.

Full dependency is not the whole picture, though, when it comes to normalization. Next up, Third Normal Form, or 3NF.