Configuring Datasources

Setting up a connection to an external cloud data warehouse starts with defining a DataSource. Via this panel, accessible from the Developers menu while viewing any model. We'll use Databricks in this example.

Note this is a particularly powerful feature of the platform and as such, is enabled by your ClearFactr Administrator. We'll discuss all of this as we get you and your company onboarded.

The Details

When logged into your Databricks account, here's how to find the specific items you'll need for your DataSource:

1. Deployment Name: use the GUID at the far left of your browser URL, such as 'abc-123ab4de-1b9a'. You don't need the '.cloud.databricks.com'
2. Warehouse Name: use the GUID found under 'SQL Warehouses' in the 'Compute' section.
3. Depending on your chosen Auth Type, you'll need either the username/password of your Personal Access Token, or the OAuth ClientID and Client Secret for your OAuth-enabled connection.
   1. For PAT, look under User icon / Settings / User / Developer / Access Tokens
   2. For OAuth U2M, look under Organization menu / Manage account / Settings / App Connections
4. For OAuth M2M, look under Organization menu / Manage account / User Management / Service Principals
5. For all Auth Types, you'll use the 'Test' button to validate your credentials.
6. For OAuth U2M connections, please specify 'https://cfdemoapp.com/oauth/cdw_redirect' as the Redirect URL

Note: When testing, if you're already logged into Databricks in a browser window, you'll immediately see a JSON Access Token upon test completion. If you're not logged in, the testing process will have you authenticate yourself via Databricks' site.

Driving your Models...

Once your Databricks DataSource is configured, you'll be able to reference it in one or more CFSQL functions to bring data into your model on the fly. Click here to learn more about the Query Builder tool that facilitates this.

Please contact the ClearFactr team with any and all comments, suggestions and requests.