Introduction to Databases
Normalization
Functional Dependency
Given a relation R, we say that the attribute y of R is
functionally dependent on attribute x of R, if and only if,
each value of x has association with precisely one y value in
R at one time.
Supplier-no. has FD with Sup-name, and City.
Also (Sup-no and Part-no) has FD with QTY.
Normalization
Normalization is a step by step process of replacing the
relationships between data by two dimension tables.
The normalized relation must have the following properties:
1- Each entry in a table represents one data item, no
repeating groups.
2- They have column homogeneous.
3- Each column assigned a distinct name.
4- All rows are distinct.
5- Both the rows and the columns can be viewed in any
shape.
Steps
1- The initial or the unnormalized relation, and this step
must be converted into the First Normal Form (1NF).
2- The (1NF) is converted into Second Normal Form
(2NF).
3- The (2NF) is converted into Third Normal Form (3NF).
4- The (3NF) is converted into Optimized Third Normal
Form (Optimized 3NF).
Unnormalized relation  1NF relations  2NF relations 
3NF relations Optimized relations.
*A relation R is in 1NF if and only if all underlying
domains contain atomic values only (No repeating groups).
*A relation R is in 2NF if and only if every non key
attribute is fully dependent on the primary key. It means:
1- For each field in the 1NF relation, ask if it relates to the
whole key or only part of the key.
2- If any field relates to only part of the key, remove that
field from the 1NF relation to a new relation, which
also contains the dependent key.
*Converting from 2NF to 3NF is similar to the conversion
to 2NF, in that we are looking for dependencies between
each field in the relation.
Example:
An Organization needs to manage their training
courses using the following data items:
Employee Number, EmployeeName, Employee
Address, CourseCode, CourseName, CoursePrice,
CourseLocation, CoursePrerequisite.
Having the following semantic assumptions:
1- Employee Number is not duplicated.
2- CourseCode is not duplicated
3- Courses have fixed prices and locations.
4- Each EmployeeName has a unique address.
It is required to accomplish an optimized 3NF database.
Unnormalized relation=
(e-n,e-na, e-a, cc, cn, cp, cl, cpre)
1NF
R1= (e-n, e-na, e-a) &
R2= (e-n, cc, cn, cp, cl, cpr)
2NF
R1 & R21= (cc, cp) &
R22= (cc, cl) &
R23= (e-n, cc, cn, cpre)
3NF
R11= (e-na, e-a) &
R12= (e-n, (e-n, or e-a)) &
R21 & R22 & R23
Optimized 3NF
R11 & R12 & R23 & R4= (cc, cp, cl)
Example 2
An Equipment company needs to convert its documents into
optimized Data bases. The body of the company reports the
following attributes:
Salesperson-number, S-name, S-working-area, Customernumber, C-name, Warehouse-number, W-location, Sales.
We have the following semantic assumptions:
1- The salesperson-number is unique.
2- No two S-names have the same S-working-area.
3- No two W-locations have the same Warehouse-number.
Example 3:
A company needs to computerize their purchase orders, and
they have a special document must be filled from the
employee, and this document contains the following fields:
Order-number, O-date, O-due-date, Supplier-number, Sname, S-address, Part-number, P-name, P-color, P-price, PQuantity.
The company has the following semantic assumptions:
1- Order-number are never duplicated
2- No two suppliers have the same address.
3- Same parts have the same price.
4- Same parts have the same color.