You could consider using the KiteSDK. It uses Avro schemas to describe
datasets, can infer an avro schema from CSV headers, and includes tools for
importing from CSV to Avro for storage. There's a tool demo that can walk
you through getting things into avro and displaying a plain text version.
The current MapReduce support is a first pass, but is enough to start
playing on. Unfortunately, I don't think it has a demo walkthrough yet.
The next release is supposed to be more user friendly.
On Mon, Jul 7, 2014 at 4:13 PM, Bhuvana Bellala <
[EMAIL PROTECTED]> wrote: