Modelling for SQLs
Data modelling in SQL is the design of tables, keys and relationships so data is stored once and stays consistent. Primary and foreign keys define links, constraints enforce rules, and normalisation, up to third normal form, removes redundancy that causes update anomalies. A good model makes queries simpler and prevents bad data.
Stage: Multiple Tables · Part of the SQL roadmap · Last reviewed 19 November 2026
What to learn in Modelling
- Choose a primary key and decide between natural and surrogate keys
- Model one-to-many and many-to-many relationships using foreign keys and junction tables
- Add NOT NULL, UNIQUE, CHECK and FOREIGN KEY constraints to protect data
- Normalise a flat table to third normal form step by step
- Judge when denormalisation is a sensible trade for read speed
Modelling interview questions
What is the difference between a primary key and a unique key?
A primary key uniquely identifies each row, cannot be NULL and a table has only one. A unique constraint also prevents duplicates, but a table can have several, and most databases allow NULL in them. Foreign keys normally reference the primary key.
What are update anomalies, and how does normalisation fix them?
When the same fact is stored in many rows, changing it in only some rows leaves inconsistent data, an update anomaly. Insert and delete anomalies are similar. Normalisation splits data into tables so each fact lives in one place, referenced by keys.
Where Modelling fits in the SQL roadmap
Modelling belongs to the Multiple Tables stage (beginner level), which takes about 18 hours. Relate data across tables.