Github user kiszk commented on the issue:
https://github.com/apache/spark/pull/20224
I totally agree to add unique ID. This is because all of the generated code
by whole-stage codegen has the same class name. It makes us hard to debug in a
production environment.
On the other hand, IIUC, the current implementation disables caching
mechanism regarding the same query. To add an unique ID generated different
string for Java code.
I am thinking about adding an ID related to a task into a comment or other
parts.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
