Here, I propose a table-parameter ("auto.purge") to set up tables to skip-trash when table/partition data is deleted, without needing to say "PURGE" on the Hive CLI. Apropos, on dropTable() and dropPartition(), table data is deleted directly (and not moved to trash) if the following hold true:
- The table is MANAGED.
- The deleteData parameter to the HMSC.drop*() methods is true.
- Either PURGE is explicitly specified on the command-line (or rather, "ifPurge" is set in the environment context, OR
- TBLPROPERTIES contains "auto.purge"="true"