In the previous article
[https://www.rittmanmead.com/blog/2016/12/etl-offload-with-spark-and-amazon-emr-part-1/] I gave the
background to a project we did for a client, exploring the benefits of
Spark-based ETL processing running on Amazon's Elastic Map Reduce (EMR) Hadoop
platform. The proof of concept we ran was on