Load Databricks’ Delta tables into Power BI
The Delta format is a powerful one. As described by Databricks, Delta “is the foundation of a cost-effective, highly scalable lakehouse”. In other words, delta tables will be the end result of your lakehouse and how you’ll consume data for analysis or machine learning.
Focusing on the analysis part, I want to show you how the Databricks connector in Power BI can be used to load data from Delta tables.
Overall you need three, at most four, pieces of information to connect to the tables:
- The URL to your Databricks workspace
- The server hostname, in other words, the SQL endpoint for connecting to the cluster
- An authentication method, preferably via Azure Active Directory or via a Personal Access Token
- And, optionally, a database name to restrict the connection
Both the server hostname and the HTTP path are found under the advanced options of the cluster you’re connecting to. Please refer to the following screenshots on how to get there.
The database is completely optional, it restricts whether Power BI should connect to a single database or all databases in the cluster.
The last piece, authentication, is up to you. You can use Azure Active Directory for passthrough authentication, or a Personal Access Token to act as a “password” for user authentication. Just don’t forget you’ll need to manage permissions in both cases (ensure user has access to databases, tables, etc.).
