feiniaofeiafei opened a new pull request, #68936:
URL: https://github.com/apache/doris/pull/68936
### What problem does this PR solve?
Problem Summary:
`DecomposeRepeatWithPreAggregation` previously supported only SUM, SUM0,
MIN, MAX, ANY_VALUE, and COUNT. Queries containing AVG or other aggregates
outside this set could not benefit from pre-aggregation before Repeat, even
when their intermediate states could be merged.
This PR extends the rewrite to eligible built-in aggregates using AGG_STATE.
For example:
SELECT a, b, c, SUM(d), COUNT(*), AVG(d)
FROM t
GROUP BY ROLLUP(a, b, c);
The rewritten plan pre-aggregates the largest grouping set and shares its
output through a CTE:
- Use `avg_combine` to produce intermediate states.
- Use `avg_merge` to aggregate these states for the remaining grouping sets.
- Use `avg_finalize` to produce results for the largest grouping set without
another aggregation.
SUM, SUM0, MIN, MAX, ANY_VALUE, and COUNT retain their existing result-based
rollup path, including SUM over COUNT results. Both paths share the same CTE
producer and Repeat.
The rewrite also preserves input nullability when constructing aggregate
states and restores grouping-key nullability above Repeat.
The existing grouping-set requirements remain unchanged. DISTINCT aggregates
and unsupported functions, including UDAFs, existing aggregate combinators, and
aggregates with ORDER BY arguments, remain excluded.
### Release note
Extend pre-aggregation optimization for ROLLUP, CUBE, and GROUPING SETS to
eligible aggregate functions such as AVG, variance, standard deviation, and NDV.
### Check List (For Author)
- Test <!-- At least one of them must be included. -->
- [ ] Regression test
- [ ] Unit Test
- [ ] Manual test (add detailed scripts or steps below)
- [ ] No need to test or manual test. Explain why:
- [ ] This is a refactor/code format and no logic has been changed.
- [ ] Previous test can cover this change.
- [ ] No code files have been changed.
- [ ] Other reason <!-- Add your reason? -->
- Behavior changed:
- [ ] No.
- [ ] Yes. <!-- Explain the behavior change -->
- Does this need documentation?
- [ ] No.
- [ ] Yes. <!-- Add document PR link here. eg:
https://github.com/apache/doris-website/pull/1214 -->
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label <!-- Add branch pick label that this PR should
merge into -->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
