Mola-maker opened a new pull request, #55582: URL: https://github.com/apache/arrow/pull/55582
Fixes #55575. `DecodeBasicAggregate` mapped Substrait's `distribution` option backwards for `variance` and `std_dev`: `POPULATION` got `ddof = 1` and `SAMPLE` got `ddof = 0`. `VarianceOptions` divides by `N - ddof`, so population needs `ddof = 0` and sample needs `ddof = 1`. This swaps the two values; `stddev` shares the same code path. Tests: corrected `SubstraitRoundTrip.AggregateRelOptions` (it encoded the buggy behavior) and added `AggregateRelOptionsSample`, `AggregateRelVarianceSingleRowGroup`, and `AggregateRelVarianceSingleRowGroupSample` (single-row group: `0.0` for `POPULATION`, `NULL` for `SAMPLE`). Verification (fork CI, run 43289990482): `arrow-substrait-substrait-test` passed in the ASAN UBSAN and ARM64 jobs (113/113 in the latter). The run's overall failure is two unrelated issues: the Conda AVX2 job fails to link `flight-grpc-example` (mold: undefined symbol `absl::lts_20240722::Mutex::Dtor()`), and the ASAN job's only failing test is `arrow-azurefs-test` (Azurite emulator failed to start, Node `ERR_REQUIRE_ESM`). -- 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]
