sohailshk opened a new pull request, #41284: URL: https://github.com/apache/beam/pull/41284
Fixes #41283 `Series.skew()` and `Series.kurtosis()` merge per-partition moments in `combine_moments`. The result used `combined_n`, which is only assigned when two non-empty partitions are merged, so when only one partition holds data (for example a small input processed in one bundle) the call raised `UnboundLocalError: cannot access local variable 'combined_n'`. After the merge loop `n` already holds the total count (it equals `combined_n` after every merge), so the result now uses `n`; results for several partitions are unchanged. Tests: `test_series_skew_kurtosis_single_partition` evaluates both on one partition (`distributed=False`); it fails on master with the `UnboundLocalError` and passes with the change. `pytest apache_beam/dataframe/frames_test.py`: 460 passed, 19 skipped. With `to_dataframe` on a `beam.Create` of 6 floats, both calls failed on master on the default runner and on the FnApiRunner; after the change they return 0.9191843544212499 and -0.25414322711619963, the same as pandas. -- 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]
