https://impala.apache.org/docs/build/html/topics/impala_hbase.html hasn't been updated for a while. The recommendations are a bit out of date - generally HBase is not the best format for analytic workloads yet that page seems to encourage using it.
If you have join queries that do aggregation operations on large fact tables and join the results against small dimension tables, consider using Impala for the fact tables and HBase for the dimension tables.
Assigning to myself to figure out what the best practice is, but I think we need to include:
- A statement Kudu offers significantly better performance for analytical workloads with mutable data
- A statement that HDFS tables are also preferable unless data is frequently mutated
- A pointer to the Kudu docs