-- Unnormalized (0NF): Customer email in every order row

-- 1NF: Atomic values, separate rows for order items

-- 2NF: Customer info in ORDERS table

-- 3NF: Product info in PRODUCTS table

-- Result: Each fact stored once, update in one place

1NF: Atomic values, no repeating groups

2NF: 1NF + no partial dependencies

3NF: 2NF + no transitive dependencies

BCNF: Every determinant is candidate key

Junction tables for many-to-many

Excellent support for normalization, foreign keys, constraints

InnoDB engine required for foreign keys

Full normalization support with constraints

Core references in this topic include BETWEEN, ANY / ALL, ROWS / RANGE. Learn what each one does, when to use it, and the execution or engine rules that matter.

BETWEEN

Checks whether a value falls inside an inclusive lower/upper range.

order_total BETWEEN 100 AND 500

ANY / ALL

Compares one value against every or at least one value from a subquery result.

salary > ALL (SELECT salary FROM interns)

ROWS / RANGE

Defines how a window frame is sliced around the current row.

ROWS BETWEEN 3 PRECEDING AND CURRENT ROW

1NF: Atomic values, no repeating groups

2NF: 1NF + no partial dependencies

3NF: 2NF + no transitive dependencies

BCNF: Every determinant is candidate key

Junction tables for many-to-many

Normalize to 3NF for most applications

Denormalize only after measuring performance

Use foreign keys to maintain relationships

Document denormalization decisions

Test with realistic data volumes

Normalized: Better writes, more joins

Denormalized: Faster reads, redundant data

Measure before denormalizing

Consider materialized views for reporting

Over-normalizing (beyond 3NF rarely needed)

Denormalizing prematurely without measurements

Storing comma-separated values (violates 1NF)