The end goal will be to implement this in HBase, and consolidate as much into BaseDataStore, but the generalities will be complex without concrete implementations first.

Currently it runs a query, followed by a delete by data ID for the results. The BatchDeleter extends ScannerBase which is the same class used in AccumuloQuery for getting the query results. It should be most efficient to just use the BatchDeleter for deletion using the initial query constraints, rather than a 2 step process of querying using the query constraints and then deleting by data ID.