sgg opened a new pull request, #1337: URL: https://github.com/apache/arrow-go/pull/1337
This diff adds support for sorting by nested fields. I noticed that the contributing guidelines suggest opening a thread on the mailing list for "significant" features; I wasn't sure if this qualified as significant but I'm happy start a thread there if the maintainers feel it's appropriate. ### Rationale for this change The primary motivation for this change is enabling support for nested field sorting in iceberg-go ### What changes are included in this PR? * `compute.SortKey` now has a `ColumnPath` field that can be used to fully address nested struct fields. `SortKey.ColumnIndex` is preserved for backwards compat/ergo. `ColumnPath` takes precedence over `ColumnIndex` if both are specified. * `sort_indices` now walks the `ColumnPath` if it is specified. * I've added some new benchmarks for `compute.SortRecordBatch`. * **Note** `maskedStructField` is basically a copy of `Struct.newStructFieldWithParentValidityMask`. I opted to copy this rather than exporting a increasing the surface area of the public API. ### Are these changes tested? * I've added some new unit tests and benchmarks. * I ran `ci/scripts/test.sh $(pwd)` after fetching the submodules. ### Are there any user-facing changes? * `compute.SortKey` has a new `ColumnPath` argument. -- 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]
