I started looking into moving our build to use Spark 2.1.1 and there seems to be a regression on the Akka Structured Streaming data source with some of the tests writing to parquet. More details available at : https://issues.apache.org/jira/browse/BAHIR-120
Anyone available to take a quick look at this? -- Luciano Resende http://twitter.com/lresende1975 http://lresende.blogspot.com/
