Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/22494
yea, that's why I change the test to use `1e6 * 1000`, instead of `1000e6`.
The point is, we know there is a bug and it's hard to fix. What I'm trying
to do here is not fixing the bug, but to allow users to fully turn off #20023.
Thus we can avoid regressions and not make it worse.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
