Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/5608#issuecomment-94958006
@shenh062326 can you explain why this particular change is likely to give
the right result? I understand the problem with shared objects (it was a known
problem when we designed SizeEstimator), but I'm not sure why sampling twice
and using the min and max sizes this way is likely to be good. I also don't see
why that particular formula at the end makes sense (for example, isn't it going
to be wrong now if there are no shared objects?).
---
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]
