Github user davies commented on the pull request:
https://github.com/apache/spark/pull/13179#issuecomment-220177521
@hvanhovell I think this is not an behavior change since 1.6. Before 2.0,
the size estimation will never decrease, so autoBroadcastJoinThreshold + 1L
should be enough to make sure that the table will not be used as broadcast. But
Spark 2.0 could decrease the size estimation, which break the internal
assumption, this PR just fix that to maintain the same behavior as before. cc
@rxin
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
