You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@hudi.apache.org by "Alexey Kudinkin (Jira)" <ji...@apache.org> on 2022/04/14 20:03:00 UTC

[jira] [Created] (HUDI-3882) Make sure Hudi Spark relations implementations provide similar file-scanning metrics

Alexey Kudinkin created HUDI-3882:
-------------------------------------

             Summary: Make sure Hudi Spark relations implementations provide similar file-scanning metrics
                 Key: HUDI-3882
                 URL: https://issues.apache.org/jira/browse/HUDI-3882
             Project: Apache Hudi
          Issue Type: Improvement
            Reporter: Alexey Kudinkin
            Assignee: Alexey Kudinkin
             Fix For: 0.12.0


After rebasing Hudi's COW table onto `BaseFileOnlyRelation` from `HadoopFsRelation`, we lost very helpful file-scanning metrics like # of files scanned, total size of files scanned, etc.

 

After some investigation, i've found out that this occurred b/c now instead of relying on `FileScan` node in Spark plans like we're before, it now uses `DataScan` node that doesn't provide such metrics.

This unwanted transition occurred b/c Spark internally predicates on `HadoopFsRelation` to decide whether it's a `FileScan` or `DataScan`, and since we stopped using `HadoopFsRelation` Hudi relations now fall into the latter bucket



--
This message was sent by Atlassian Jira
(v8.20.1#820001)