Normalization organizes relational data to reduce duplication and prevent insert, update, and delete anomalies.

Example

• If customer name and email are repeated on every order, changing an email requires multiple updates.
• A CUSTOMERS table can store customer_id, name, and email once.
• An ORDERS table can reference customer_id.

Normal forms

• 1NF: fields contain single values from their intended domains; repeating groups are avoided.
• 2NF: in 1NF, with every non-key attribute depending on the whole candidate key rather than part of a composite key.
• 3NF: in 2NF, with no prohibited transitive dependency of non-key attributes on a candidate key.

Denormalization

• Intentionally introduces duplication or precomputed data to improve reads or simplify access patterns.
• Can help analytics and high-read workloads.
• Adds consistency, refresh, and maintenance complexity