Github user holdenk commented on the issue:
https://github.com/apache/spark/pull/13571
So @rxin just commented with an explicit "Won't Fix" on the JIRA.
@zero323 certainly Jython can be slow, but as the benchmark shows it can be
much faster than our current Python UDF approach.
@davies Some people have been actively using it in a similar use case
(Python UDFs with Pig) (as @zero323 mentions).
I'd rather try and find a way to expose this as a experimental API - but if
the consesus is "Won't Fix" I'll put this on my back burner of things to
contribute as a Spark Package (although the cost of maintaining Spark Packages
is frustrating so I'll also take a look at adding it to one of the meta
packages like Bahir if there is interest there). In the meantime I'll do some
poking with arrow and also looking at pip install-ability since these seem to
be of importance to much of the PySpark community.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
