westonpace commented on pull request #11588: URL: https://github.com/apache/arrow/pull/11588#issuecomment-962370775
> I think the default should still be to prefetch automatically, though. That seems to be the dominant use case (if you load a batch, you probably want to use it in its entirety or a large part of it). That seems reasonable. What if wel add a new optional parameter to `arrow::io::MemoryMappedFile::Open` which is `bool auto_willneed = true`. Then whether the IPC reader does madvise or not will depend on how the file is created (and this is completely independent to whether the IPC reader uses a ReadRangeCache). -- 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]
