Github user BryanCutler commented on the issue:
https://github.com/apache/spark/pull/22913
I'm a little against adding this because the Arrow Java Vectors used so far
were done to match the internal data of Spark, to keep things simple and avoid
lots of conversions on the Java side. Conversions to supported types are being
done before reading the data in Java.
For instance, there are lots of timestamp types for other time units, but
we only accept microseconds with tz in Java (to match Spark) and do any
necessary conversions in Python before writing the Arrow data. Can this also be
done in R?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
