The transformer class should methods for querying model's "native" data schema (all fields as defined in the PMML document) and "mapped" schema (as was configured during transformer building). The return type of such methods could be Apache Spark's standard StructType, which can then be inspected and processed using built-in tools.
The transformer class should methods for querying model's "native" data schema (all fields as defined in the PMML document) and "mapped" schema (as was configured during transformer building). The return type of such methods could be Apache Spark's standard
StructType
, which can then be inspected and processed using built-in tools.