Hive handles primitive types using the base Hadoop writable rather than
its own during joins, which means that in order to use the proper
DoubleWritable class from Hive you have to actually return an object of
the appropriate type in JsonSerDe.serializeField instead of a primitive.
Attempt #2.
Hive handles primitive types using the base Hadoop writable rather than its own during joins, which means that in order to use the proper DoubleWritable class from Hive you have to actually return an object of the appropriate type in JsonSerDe.serializeField instead of a primitive.