Skip to content
Quantoros

What is denormalization and when would you do it?

SQLLevel: IntermediateDatabase Design

What is denormalization and when would you do it?

Hints
  • Which functional dependencies hold between table attributes?
  • Check insert, update and delete anomalies; what changes when the table is decomposed?

Answer

Denormalization intentionally introduces redundancy to improve read performance. You do it when:

  • JOINs are too expensive for your query patterns
  • You're building a reporting / analytical system where read speed matters more than write simplicity
  • The data changes infrequently so keeping duplicates in sync is manageable

Common patterns: storing precomputed aggregates, copying frequently-joined columns to avoid a JOIN, using JSON columns for flexible attributes.

The tradeoff: you trade write complexity and risk of inconsistency for faster reads.

Source and licence: SQL Interview Questions — Arda Kocadoru · MIT

Related questions