Github user falaki commented on the issue:
https://github.com/apache/spark/pull/15328
A user ran into this limit when trying to parallelize a fairly large R
data.frame. The user has extensive logic implemented in R on that data.frame
and migrating it over to SparkDataFrame API is a lot of work (the fact that
SparkR has subtle API incompatibilities with base does not help here).
User wants to run an embarrassingly parallel function on many workers. So
after feature engineering on R data.frame it needs to be converted to
SparkDataFrame.
In general I think optimizing this code path will help with `spark.lapply`
and `dapply` use cases where there is a lot of data traffic between R and JVM.
---
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]
