Yida Wu created IMPALA-13081:
--------------------------------
Summary: Improve the roll up function for group aggregator
Key: IMPALA-13081
URL: https://issues.apache.org/jira/browse/IMPALA-13081
Project: IMPALA
Issue Type: Improvement
Components: Backend
Reporter: Yida Wu
Assignee: Yida Wu
Currently ROLLUP function in the pre aggregator would treat each aggregator
separately using repetitive same raw input from the child nodes, it is possible
to optimize the process by using the aggregator's aggregated result as the next
aggregator's input during the ROLLUP.
In simpler terms, the ROLLUP function can be smarter about reusing the results
from previous aggregations instead of starting from scratch each time. This
optimization could significantly improve performance when dealing with large
amount of data.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
