JonSnow1807 commented on PR #55460: URL: https://github.com/apache/arrow/pull/55460#issuecomment-6755845049
Thanks for the guidance. I ran an adversarial review with two separate reviewers and compared the kernel against the existing ones (round, checked arithmetic, casts, `divide(duration, int64)`). From that, the kernel now uses the physical `Int64Type` like its neighbours, the test also covers the lower int64 bound, and the docs say the computation is done in float64. I switched to truncation toward zero, so Randolf’s example now gives exactly numpy’s output, and I added it as a Python test. I’ve gone through every line myself. Also, to correct my earlier comment: float32 and float16 are not cast in this PR. -- 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]
