Github user kiszk commented on the issue:
https://github.com/apache/spark/pull/19598
I see. It makes sense for debugability. On the other hand, the NULLable
property in schema is very important for optimizations to achieve better
performance.
One of ideas is to add conf for debug mode to enforce insertion of null
check at reading data from an iterator or column vector.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
