public class BucketVersionPopulator
extends Transform
This class analyzes and sets the bucketing versions.
A set of data values can be distributed into N buckets differently depending on the used hashing algorithm.
Hive right now supports multiple hashing algorithms - the actual algo is identified by "bucketingVersion".
Bucketing version can be re-select after every Reduce Sink; because a full shuffle can re-distribute the data according to a new hash algo as well.
Depending on the table Hive might need to write it's data in some specific bucketing version.
In case a bucketed table is read from the table location; the data should be threated as described by the table's bucketing_version property.