Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/455#issuecomment-45420334
I personally think Converter is good to keep because it's the only way to
expose arbitrary InputFormats to Python. In Java and Scala, you can always call
SparkContext.hadoopRDD to get an RDD of Java objects of whatever type there was
in the InputFormat. That's why no such interface is needed there. But in
Python, there needs to be a way to turn the Java type into a Pickle-able
object, and this way seems as good as any.
We should just document this as being an advanced features and less
preferred than using Spark SQL when Spark SQL supports the given data source.
---
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.
---
