I see, that is a fair reason. There are parallel efforts to add Parquet support to the old BucketingSink, but I see the point.
Before going into a deep review, can you update the description with how exactly the legacy truncater should be working: what copy and rename steps it does and how it behaves under failure / repeated calls. Also, I would suggest to name it `Truncater` rather than `TruncateManager`. Too many managers all around already ;-) [ Full content available at: https://github.com/apache/flink/pull/6608 ] This message was relayed via gitbox.apache.org for [email protected]
