Comparison
DenormalisationvsStar schema
Denormalisation
you copy the customer's name into the orders table so the list page stops doing a join on every render.
Deliberately duplicating data to avoid joins or aggregate computation on the read path. It is the standard fix once a hot query's join cost dominates, and it is how most read models are built. The duplicated field is now yours to keep in sync forever, so it only pays where reads vastly outnumber writes.
Full entry →Star schema
every dimension is one join away from the fact, so no query needs more than a single hop to get a label.
A central fact table surrounded by fully denormalised dimensions, each reachable in one join. It is the default shape for analytical modelling because it is predictable to write, cheap to plan and easy for a BI tool to reason about. The redundancy inside the dimensions is the deliberate trade — you accept repeated country names in exchange for never chaining joins in a report.
Full entry →