alamb commented on code in PR #5900:
URL: https://github.com/apache/arrow-datafusion/pull/5900#discussion_r1160617706
##########
datafusion/core/src/physical_plan/file_format/file_stream.rs:
##########
@@ -264,9 +264,9 @@ impl<F: FileOpener> FileStream<F> {
Ok(reader) => {
let partition_values = mem::take(partition_values);
- let next = self.start_next_file().transpose();
-
+ // include time needed to start opening in
`start_next_file`
Review Comment:
> return a Future and the call should take almost no time.
> Sorry that I do not get why making this change.
It was suggested by @thinkharderdev in
https://github.com/apache/arrow-datafusion/pull/5790#discussion_r1155734134:
> I think we want this call before self.start_next_file. In some cases (for
example, when reading from the LocalFileSystem store outside of a Tokio
context) this can block and we would not capture the time spent on disk IO.
--
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]
