In this post we’ll be going over how to configure an ODBC connection to DataBricks within Composable so you can run spark queries from Composable resources (i.e. QueryViews and DataFlows)
The first thing you’ll need to do is install the 64 bit Simba Spark ODBC Driver. This can be downloaded here:
Next, you’ll need to create a new ODBC connection settings key in the Composable Key Vault.
Below are the settings that you’ll need to specify. If The Simba Spark ODBC Driver is not available in your list after the driver is installed, you may need to restart the Composable web app in IIS.
Host: This is the same host name you use for accessing databricks via the user interface. i.e. xyz.azuredatabricks.net
Password: To create an authentication token in Databricks, click on the profile icon in the upper right corner of Databricks. Under ‘User Settings’, click ‘Access Tokens’. And then click ‘Generate New Token’. Paste the token into this password field.
ThriftTransport: 2 (2 corresponds to SparkThriftServer (Spark 1.1 and later))
httpPath: This is the path to the cluster that will be used to execute the query. Click on your cluster in the Databricks UI. Then under ‘Configuration’, scroll down to ‘Advanced Options’. Then click on the tab ‘JDBC /ODBC’. There you will see a field called ‘HTTP Path’.
AuthMech: 3 (3 corresponds to UserName and Password)
If you receive the error “ERROR [HY000] [Simba][ThriftExtension] (14) Unexpected response from server during a HTTP connection: SSL_connect: certificate verify failed.”, then set the UseSystemTrustStore value to 1, otherwise you can ignore the setting UseSystemTrustStore.