Denormalization

Denormalization is deliberately introducing redundancy into a normalized design — for example duplicating a column or pre-joining tables — to make reads faster by avoiding joins.

It is a performance trade-off, not a mistake. Data warehouses and reporting systems denormalize into star schemas so analysts run fewer joins. The cost is that duplicated data must be kept in sync on every write, usually via triggers, ETL, or application logic.

Denormalize only after measuring that joins are the real bottleneck — premature denormalization creates consistency bugs for no gain.