We are getting varieties of JSONs/XMLs as input whose schema is always evolving. I want to process them using ORC or Parquet format in Hadoop/Hive environment for performance gain.
I know below common style of achieving same objective : Use JSONSerde or XMLSerde library, first create hive table using these serde. Later select * fields query will be fired on each xml/json hive table to save as orc or save as parquet into another table. Once done successful I can drop these Serde Table and XML/JSON data.
What would be another good ways of doing same ?