Github user mgaido91 commented on the issue:
https://github.com/apache/spark/pull/22713
@gatorsmile thanks for your comment. Honestly I don't see much value in
holding, as we are (nearly always) using semantic equality throughout the
Analyzer/Optimizer. So different expressions with the same ID are already
causing troubles. Here we are changing/introducing no news about that. Usually,
instead, not using semantic equality results in a bug (that's also the reason
why we have `AttributeSet` for instance).
I agree on enforcing that exprIds are unique on the plan, otherwise we do
can have issues, but we would already have regardless of this PR.
cc also @cloud-fan for his thoughts.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
