If Spark is enabled, the data context used by the Interpreter becomes *CalciteConnectionImpl.SlimDataContext* which has no *JavaTypeFactory* by design.
So when the *org.apache.calcite.interpreter.JoinNode* is initialized and tries to combine the rowtype of the Join's inputs, a NPE occurs because the type factory of the data context is called for utilization, but it's NULL as mentioned. Is this a missing implementation that I may try to work on ? Or should I set a configuration value ? What am I missing here please ? Thanks, Gelbana
