Adding back repeated or pre-calculated data on purpose, as warehouse dimension tables do, to speed up analytical queries. It reverses the trade-off made by normalised transactional schemas.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Denormalisation in context, with comparison tables and the common traps.
Terms in this definition
- Warehouse
Fabric item offering complete T-SQL support (DML, DDL, multi-table transactions) over Delta tables held in OneLake; lakehouse SQL analytics endpoints, by contrast, are read-only.
- Trade-off
Weighing competing goals, constraints and stakeholder interests to find the best result that can actually be achieved. Architects set out the options and what each means; the stakeholders choose.