Skip to content
BENIBIX Learn

Denormalization

Data🧭 FoundationsData Modeling

Also known as Denormalisation, Denormalized Table

The same data is copied into one wide table. Reads get much faster.

Think of it like

It is like writing a friend's phone number in each book you own. It takes more ink, but you find it fast.

Example

An orders table also keeps the buyer's city and the product name. A report of sales by city then needs no joins.

Why it matters

Most analytics tables are denormalized. It is the main trick that makes big reports fast.

Definition

Denormalization puts copies of data back into one table on purpose. A normalized design keeps each fact in one place and joins tables to read it. Joins cost time on big data. So for reports, you copy the needed fields into one wide table. Reads get much faster, but each copy must be kept up to date.

Common mistake

Doing it in the app's own database. There, the copies drift apart and cause bugs.

Key takeaways

  1. 1
    Copy on purpose. The same fact sits in many rows.
  2. 2
    Fewer joins. One wide table answers the question.
  3. 3
    Harder to update. Every copy must change together.

Learn it properly