Normalization is the process of structuring a relational database to reduce redundancy, dependency, and improve data integrity. It involves organizing data into multiple related tables using rules called normal forms.

Let’s walk through:

🎯 Goals of Normalization

• Eliminate redundant data (e.g., storing the same data in multiple tables)
• Ensure data dependencies make sense
• Make insert, update, delete operations efficient and safe

✅ First Normal Form (1NF)

🎯 Rule:

• Each cell contains atomic (indivisible) values.
• Each record must be unique.
• No repeating groups or arrays.

❌ Bad Table (Not in 1NF):

Courses is a multi-valued field → Not atomic.

✅ 1NF Table (Atomic values):

Split into two tables:

Students

Enrollments

✅ Second Normal Form (2NF)

🎯 Rule:

• Must first satisfy 1NF
• Remove partial dependencies (i.e., non-key columns depending on only part of a composite key)

❌ Bad Table (1NF but not 2NF):

• Composite key: (StudentID, Course)
• Instructor, Department depend only on Course, not full key → Partial dependency

✅ 2NF Solution:

Split into:

Enrollments

Courses

Now all non-key fields fully depend on their primary key.

✅ Third Normal Form (3NF)

🎯 Rule:

• Must satisfy 2NF
• No transitive dependencies (non-key depends on non-key)

❌ Bad Table (2NF but not 3NF):

• DeptName depends on DeptID, not on EmployeeID → Transitive dependency

✅ 3NF Solution:

Employees

Departments

✅ Boyce-Codd Normal Form (BCNF)

🎯 Rule:

• Stronger version of 3NF
• For any dependency X → Y, X must be a super key

Use when:

• Multiple candidate keys exist
• 3NF is not enough

📊 Summary Table

🛠️ How to Normalize in Practice

1. Start with one large table (denormalized)
2. Identify repeating groups, move them into separate tables (1NF)
3. Remove partial dependencies → move data into separate tables (2NF)
4. Eliminate transitive dependencies → create foreign key relationships (3NF)
5. Analyze candidate keys and apply BCNF if needed

🔁 Denormalization (reverse process)

Sometimes, you denormalize for:

• Reporting
• Joins performance
• Simplified queries

Use carefully, only when performance justifies it.

Pages: 1 2