Thanks wesm for clearing this. My understanding (albeit limited) of Arrow was that it would complement Parquet. An example would be speeding up the in-memory re

Thanks wesm for clearing this. My understanding (albeit limited) of Arrow was that it would complement Parquet. An example would be speeding up the in-memory representation of Parquet files. I believe mitigating the serialization costs will also help projects like PySpark. Looking forward to your post.