Add user-friendly names for datasets?

URNs are good for canonical and unique naming, but less friendly for visualization since not everyone thinks about their datasets as URNs (though machines will happily consume them). The URN can be globally unique, but dataset name need only be unique per data source.

For example, given a redshift table foo.bar, the URN might look like :redshift:foo.bar:. However, for visualization, we may want to present the dataset as <datasource, dataset_name>. In other words:

URN
:redshift:foo.bar:
:redshift:foo.baz:
:hive:baz.baz:

vs.

Data Source	Dataset
Redshift	foo.bar
Redshift	foo.baz
Hive	baz.baz

An added benefit of breaking data source as a column is a more natural workflow for filtering datasets by data source.

[sshah-wework]

Would this replace, or compliment the urn column in the datasets table?

It wouldn’t be a replacement to URN, just an additional field.

As an alternative, given some URN, the DatasetAPI could provide a more friendlier response i.e. datasource, dataset_name, latest_version, etc? Right now, we only respond back with the urn, but that’s not all that helpful

Yes exactly, but I don’t think we should do that by parsing URN… if we had dataset_name as a separate field we could just return it separately.

[wslulciuc]

Fixed #627