westonpace commented on code in PR #13959:
URL: https://github.com/apache/arrow/pull/13959#discussion_r959823613
##########
cpp/src/arrow/dataset/file_base.h:
##########
@@ -386,7 +389,7 @@ struct ARROW_DS_EXPORT FileSystemDatasetWriteOptions {
/// and only write the row groups to the disk when sufficient rows have
accumulated.
/// The final row group size may be less than this value and other options
such as
/// `max_open_files` or `max_rows_per_file` lead to smaller row group sizes.
- uint64_t min_rows_per_group = 0;
+ uint64_t min_rows_per_group = 10;
Review Comment:
It's fine for `max_rows_per_group` and `min_rows_per_group` to have the same
value I think. `1 << 18` is probably ok. The unit tests used `10` because I
needed to be able to test the various features without generating a whole bunch
of data (which would be time consuming) but `10` would have poor performance in
practice because that means we would need to write a big block of metadata
every 10 rows.
--
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]
