Relational database design: model data that stays correct

A database schema outlives most of the code written against it. Good design makes features easy to build and data easy to trust. Poor design leads to duplicated data, contradictory records and painful migrations years later. This course teaches the thinking behind solid relational design, with practical patterns for situations real applications face.

What you will learn

• Turning requirements into entities, attributes and relationships
• Drawing clear entity relationship diagrams
• Cardinality and optionality, and how they become keys and constraints
• Normalisation to third normal form, explained with practical examples
• When and how to denormalise deliberately
• Choosing natural or surrogate keys, and using constraints as documentation
• Patterns for many to many relationships, hierarchies, history, money and multi tenant data
• Evolving schemas safely with migrations and zero downtime changes

Case study

We design the database for a clinic booking platform: clinics, practitioners, services, patients, appointments, payments and reviews. Requirements arrive in stages, as they do in real projects, so you practise adapting a design without breaking it.

Tools

Examples use PostgreSQL syntax, but the design principles apply to any relational database, including MySQL, SQL Server and SQLite. For diagrams you can use any tool; free options such as dbdiagram.io or draw.io work well.

Who it is for

Developers who can write SQL and want to design schemas with confidence, technical leads reviewing data models and analysts who build their own databases. You should be comfortable with basic SQL, including creating tables and writing joins.