Details
-
Sub-task
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
This is more likely to be a calcite enhancement but am logging it here to track it any way.
Currently queries like
select count (distinct dim) from table
is pushed partially to druid as group by dim followed by a count executed by hive QE. This can be enhanced by using the nested (eg chained execution) group by query such as the first (inner) GB query does group by key and the second (outer) does the count.
Attachments
Issue Links
- Dependency
-
CALCITE-1822 Push Aggregate that follows Aggregate down to Druid
- Open