Database Normalization

What Is Normalization?

Database normalization is the process of organizing data to minimize redundancy and improve data integrity. It involves breaking down tables into smaller, related tables.

Normal Forms

First Normal Form (1NF)

• Remove repeating groups
• Each column contains atomic (indivisible) values
• No duplicate columns

Second Normal Form (2NF)

• Satisfy 1NF
• Remove partial dependencies
• Non-key attributes must depend on entire primary key

Third Normal Form (3NF)

• Satisfy 2NF
• Remove transitive dependencies
• Non-key attributes depend only on primary key

Boyce-Codd Normal Form (BCNF)

• Stricter than 3NF
• Every determinant is a candidate key

Example: Normalization Process

Unnormalized (Bad)

Students
StudentID | Name  | Courses
1         | Alice | Math, Physics, Chemistry
2         | Bob   | English, History

1NF (Better)

Students
StudentID | Name
1         | Alice
2         | Bob

StudentCourses
StudentID | Course
1         | Math
1         | Physics
1         | Chemistry
2         | English
2         | History

Benefits

• Reduces data redundancy
• Improves data integrity
• Easier to maintain and update
• Better query performance (fewer joins needed)
• Reduces storage space

Trade-offs

• More tables = more joins
• More complex queries
• Slightly slower reads
• Need careful schema design

Denormalization

Sometimes intentionally denormalize for performance:

• Pre-computed aggregates
• Cached calculations
• Duplicate data for read performance

Always measure impact before denormalizing.

Best Practices

• Normalize to at least 3NF by default
• Denormalize only when measured need exists
• Document your normalization decisions
• Use surrogate keys (auto-increment IDs)
• Define proper foreign keys with constraints
• Regularly review schema for optimization