Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
0.6.0
-
None
-
None
Description
Most likely by deleting all new row versions by timestamp in case the Hive statement fails.
Note that this isn't a consistency guarantee (e.g. concurrent HBase queries may see the new rows appear and then disappear as phantoms), only atomicity with respect to the update effect.