Github user mgaido91 commented on the issue:
https://github.com/apache/spark/pull/22494
yes @cloud-fan, I see your point. I am neutral to this change honestly.
It is true that it avoids regressions form previous cases, but it is also
true that it doesn't make the behavior better: I mean in some (weird) cases
like this one it behaves better, but there are other cases when the opposite is
true since we were overestimating the precision.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
