Verify file ingestion (AVRO format)

Verified ingestion of Avro-formatted files.
Mixed case works. I think there was an earlier bug related to mixed case that you may have hit. It was closed some time back.

Added a fix to schema discovery for Avro files. https://github.com/Teradata/kylo/pull/17

Looks like there are issues choosing an Avro Sample file when defining a feed?

Looking at the code, it looks like we are generating a script to run on the spark shell service. https://github.com/Teradata/kylo/blob/f66903fe61e5f968856f8e159b50e190de4aa5ca/plugins/schema-discovery-default/src/main/java/com/thinkbiganalytics/discovery/parsers/hadoop/SparkFileSchemaParserService.java

The script:

Given that we are running sqlContext.read.avro and not importing any avro libraries this should fail, however no logs are being captured in SparkFileSchemaParserService.java making it difficult to determine.

Try field names with mixed case as part of your test, I think we saw failures with mixed case field names when processed by Spark (1.5.2).