Details
-
Sub-task
-
Status: Closed
-
Minor
-
Resolution: Fixed
-
0.14.0
-
None
Description
JOIN operator does not update the column stats when the number of rows changes. All other operators scales up/down the column statistics when the number of rows changes. Same should be done for JOIN operator as well. Because of this dataSize might become negative as numNulls can get bigger than numRows (if scaling down of column stats is not done).
Attachments
Attachments
Issue Links
- is related to
-
HIVE-5579 Update statistics rules for different types of joins
- Open