luisquintanilla commented on PR #257: URL: https://github.com/apache/arrow-dotnet/pull/257#issuecomment-4775374238
Thanks for the offer to adapt this material, and for all the work here - the selection, comparison, casting and formatting kernels are a great start on a compute surface for arrow-dotnet. I've opened #379 with a focused first slice: SIMD-accelerated `Sum`/`Min`/`Max`/`Mean` aggregation kernels (generic `INumber<T>` + `TensorPrimitives` fast path on net8.0+, per-type scalar fallback on netstandard2.0/net462, validity-aware null handling, and BenchmarkDotNet coverage). It's intentionally small so it's easy to review. A suggestion for reconciling the two: keep #379 as the aggregation slice, and let the rest of your work land as its own focused slices (selection/filter, comparison, casting, formatting) rather than one large change - that seemed to be the main review concern here. Where I build on your API ideas (for example `ArgMin`/`ArgMax`, `IArrowArray` overloads, or an explicit null-handling option), I'll credit you. Does that split work for you? -- 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]
