Using hive table over parquet in Pig
If you have files with 2 different schemas, the following seems to be sensible:
- Split up the files, based on which schema they have
- Make tables out of them
- If desirable, load the individual tables and store them into a supertable