daniel-adam-tfs commented on PR #654: URL: https://github.com/apache/arrow-go/pull/654#issuecomment-4224069093
Couple of thoughts: 1. if we could guarantee that []parquet.FixedLenByteArray (=[][]byte) is contiguous (ie if in memory we have a single backing []byte array) we could use SIMD implementation for this case. The output buffer is given to the Decode function by the user, so we cannot guarantee it. I guess we could change []parquet.FixedLenByteArray to a struct with private []byte field and which we would use for all FLBA items, but that would mean changing the public API. 2. I wanted to implement also encoding using SIMD, but I don't know when I'll have the time for it, so lets do it outside of this PR. 3) We should add some big endian architecture in CI builds. I've used emulation in Docker to emulate s390x and ran all unit tests, they all pass for me currently. Got nothing else, so review away and I'll fix anything necessary, so we can this merged. -- 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]
