The problem is that ReadaheadSpooler will make a copy anyway when reading from a memory-mapped file (because of padding).
More generally, I don't think it is really relevant here. Parsing CSV data will always be much slower than the speed of raw memory copies (which should be on the order of 30 GB/s from main memory). [ Full content available at: https://github.com/apache/arrow/pull/2576 ] This message was relayed via gitbox.apache.org for [email protected]
