How do you model data in Cassandra? Partition key vs clustering key.

Answer preview

Cassandra modelling is query-first: you list the queries the application must serve, then create one denormalized table per query. Joins do not exist, so duplicating data across tables is the intended design, not a smell.…