mdepero commented on PR #13277: URL: https://github.com/apache/arrow/pull/13277#issuecomment-1149122241
@zeroshade My understanding of `BytesToBools` would still require separately allocated memory for bytes and bools, and it just copies the values from byte encoding to []bool's? Assuming `make([]bool, n)` only allocates 1 byte per 8 bool values, is it better to do two allocations each of size (n/8) bytes or one allocation of size (n) bytes? I've pushed a new version that retains the scratch space and only allocates once, but that one allocation is now n bytes since the cast from `[]byte` to `[]bool` creates 1 bool per 1 bytes. I do agree it's a bit hacky to simply ignore the buf passed into the skipValues readFn, which approach do you recommend? -- 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]
