What is Denormalization?
Deliberately duplicating data across tables or documents to avoid expensive joins and speed up read-heavy query patterns.
Definition
Denormalization intentionally copies data — storing a product's name inside each order line, or maintaining a comment_count column on posts — so common queries avoid joins and aggregations. It is a read-performance optimization purchased with write complexity: every duplicate must be kept in sync via triggers, application logic, or asynchronous jobs, and bugs there produce inconsistent data. Document stores like MongoDB and wide-column stores like Cassandra practically require it, since their query models are built around precomputed, query-shaped data. The pragmatic rule: normalize until performance data proves you need otherwise, then denormalize the specific hot paths.