According to Greg, it is necessary to visit Data Wrangler before attempting to define a feed with a Hive datasource.
I witnessed the following error:
```2018-10-02 18:41:56 INFO launcher-proc-2:SparkShellApp:61 - java.lang.ClassNotFoundException: java.lang.NoClassDefFoundError: org/apache/hadoop/hive/shims/ShimLoader when creating Hive client using classpath: file:/opt/kylo/kylo-services/plugin/app/kylo-spark-shell-client-v1-0.9.2-SNAPSHOT.jar```
which did not come up, if I had visited DW first.
1. Click on + for new feed
2. Click on Data Ingest for template
3. Choose "Source" to go to source step.
4. Choose Hive, choose a table and click Next.
The dataset will be fetched. Click on it takes you to a Preview.
If Data Wrangler had not been visited first which is loading something needed in the Spark-Shell session.