Skip to content

feat: Support ANSI mode avg expr (int inputs) - #2817

Merged
andygrove merged 18 commits into
apache:mainfrom
coderfender:add_support_ansi_mode_avg_expr
Dec 29, 2025
Merged

andygrove merged 18 commits into
apache:mainfrom
coderfender:add_support_ansi_mode_avg_expr

Conversation

@coderfender

@coderfender coderfender commented Nov 24, 2025 •

Copy link
Copy Markdown
Contributor

Which issue does this PR close?

Closes #531

Rationale for this change

Support ANSI eval mode (for integer types) for average expression

What changes are included in this PR ?

  1. Changes to avg implementation to support eval modes
  2. Given that Spark considers Inf as a valid output and doesnt necessarily throw an exception, we follow the same approach in comet as well (essentially making eval mode's redundant for numeric inputs)

How are these changes tested?

TODO : Support ANSI mode for avg expression but for decimal inputs (which throw exception in ANSI mode unlike their integer counterparts)

@coderfender coderfender changed the title Add support ansi mode avg expr feat: Support ANSI mode sum expr Nov 24, 2025
@coderfender coderfender changed the title feat: Support ANSI mode sum expr feat: Support ANSI mode avg expr Nov 24, 2025
@codecov-commenter

codecov-commenter commented Nov 24, 2025 •

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.55%. Comparing base (f09f8af) to head (b444cba).
⚠️ Report is 1153 commits behind head on main.

Additional details and impacted files
@@             Coverage Diff              @@
##               main    #2817      +/-   ##
============================================
+ Coverage     56.12%   59.55%   +3.42%     
- Complexity      976     1376     +400     
============================================
  Files           119      167      +48     
  Lines         11743    15486    +3743     
  Branches       2251     2564     +313     
============================================
+ Hits           6591     9223    +2632     
- Misses         4012     4969     +957     
- Partials       1140     1294     +154     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@andygrove

Copy link
Copy Markdown
Member

@coderfender You can merge latest from main to fix the current CI failures

@coderfender
coderfender force-pushed the add_support_ansi_mode_avg_expr branch from 5192933 to 33b0b2f Compare November 24, 2025 20:36
@coderfender

Copy link
Copy Markdown
Contributor Author

@andygrove , Thank you . I rebased my feature branch with main. Please kick off CI whenever you get a chance . Thank you

@coderfender

Copy link
Copy Markdown
Contributor Author

@andygrove , fixed issues with formatting and an unused import. Please kickoff the CI whenever you get a chance . Thank you

Comment thread native/core/src/execution/planner.rs Resolved
@coderfender coderfender changed the title feat: Support ANSI mode avg expr feat: Support ANSI mode avg expr (int inputs) Dec 1, 2025
@coderfender

coderfender commented Dec 20, 2025 •

Copy link
Copy Markdown
Contributor Author

@andygrove , @comphead . Please take a look whenever you get a chance . ANSI support for average is rather more relaxing compared to say 'Sum' because of the intermediate sum accumulation uses a float

Comment thread native/spark-expr/src/agg_funcs/avg.rs Resolved
Comment thread native/spark-expr/src/agg_funcs/avg.rs Resolved

@andygrove andygrove left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @coderfender

@coderfender

Copy link
Copy Markdown
Contributor Author

Thank you for the approval @andygrove

@coderfender

Copy link
Copy Markdown
Contributor Author

Fixing merge conflicts and test failures

@coderfender

coderfender commented Dec 23, 2025 •

Copy link
Copy Markdown
Contributor Author

@comphead , I made changes per review and it gives me great pleasure to remove Spark's EvalMode import from aggregates.scala :)

@coderfender

Copy link
Copy Markdown
Contributor Author

@comphead cpuld you please rerun the failed teats (due to OOM) ? Thank you

@coderfender

Copy link
Copy Markdown
Contributor Author

@comphead , thank you for rerunning the failed runs . Please review / merge whenever you get a chance

// without GROUP BY
withSQLConf(SQLConf.ANSI_ENABLED.key -> ansiMode.toString) {
val res = sql("SELECT avg(_1) FROM tbl")
checkSparkAnswerAndOperator(res)

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The last nit: can we remove res* variables here and refer to the sql text

@comphead comphead left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @coderfender for the PR

@coderfender

Copy link
Copy Markdown
Contributor Author

Thank you for the approval @comphead

@coderfender

Copy link
Copy Markdown
Contributor Author

@andygrove , @comphead please merge this whenever you get a chance

@andygrove
andygrove merged commit 2bf2835 into apache:main Dec 29, 2025
134 of 135 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add ANSI support in SUM and AVG

5 participants