Databricks connection & capability registration in EDGE

Has anyone registered databricks JDBC connection via EDGE in DGC?

Since there is no schemas concept in databricks, does that mean I should register as many JDBC connections as the number of databases in databricks? If so, so far so good.

But the metadata ingestion capability seems to be having one-to-one link with the JDBC connection. So, does that mean again as many capabilities as JDBC connections…?

noor

Posted 4 years ago · Edited 1 year ago·Last reply 4 years ago

3 comments

knutge_hofseth

·4 years ago · Edited

I think so. The token has access to a cluster. So there might be something you can do with that. If you can restrict what databases a cluster can access.

But what you see is about what I see, but I connect to an Azure Databrick so there might be small differences.

knutge_hofseth

·4 years ago · Edited

I have.

What we found was something that Edge interpreted as a database and schemas. It looked just like what I would expect to see if I connected to a SQL database. It called the database Spark as the automatic naming, and discovered as schemas what we found as databases in the Data tab in databricks.

noor

OP4 years ago · Edited

Thank you.

So, on my side, the connection brings “hive_metastore” as the top-level database and all other “real” databases as schemas underneath “hive_metastore”. is this what you see also?

Is there a way to restrict the databricks token to list only “certain” databricks databases