Github user rxin commented on the issue:
https://github.com/apache/spark/pull/15899
Thanks for the example. I didn't even know that was possible in earlier
versions. I just looked it up: looks like Scala 2.11 rewrites for
comprehensions into map, filter, and flatMap.
That said, I don't think it's a bad deal that this no longer works, given
it was never intended to work and there's been a deprecation warning.
I still maintain that it is risky to support this, because Scala users
learn for comprehension not just for a simple "for filter yield", but as a way
to chain multiple generators together, which is not really well supported by
Spark (even if it is, it's a really bad operation for users to shoot themselves
in the foot because it would be a cartesian product).
Rather than faking it as a local collection, users should know RDD is not.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
