Databricks: Python SDK

(Last Updated On: )

This post is how to use the Databricks Python SDK.

Install the Package

pip install databricks-sdk

Update Package

pip install databricks-sdk --upgrade

Check Package Version

pip show databricks-sdk | grep -oP '(?<=Version: )\S+'

Setup WorkspaceClient

from databricks.sdk import WorkspaceClient

secret = dbutils.secrets.get(scope = "<SCOPE>", key = "<KEY>")

w = WorkspaceClient(
  host = 'https://<URL>/'
  azure_workspace_resource_id = '<RESOURCE_ID_OF_DATABRICKS>',
  azure_tenant_id = '<TENANT_ID>',
  azure_client_id = '<CLIENT_ID>',
  azure_client_secret = secret
)

Setup AccountClient

You can get the account_id from the databricks account portal. By your id in the top right hand corner.

from databricks.sdk import AccountClient

secret = dbutils.secrets.get(scope = "<SCOPE>", key = "<KEY>")

a = AccountClient(
  host = 'https://accounts.azuredatabricks.net'
  account_id = '<ACCOUNT_ID>'
  azure_tenant_id = '<TENANT_ID>',
  azure_client_id = '<CLIENT_ID>',
  azure_client_secret = secret
)

List Workspace Groups

NOTE: You must also setup the workspaceclient to do this.

w.groups.list()

List Account Groups

NOTE: You must also setup the accountclient to do this. You must also be account admin.

a.groups.list()

Create Storage Credential

NOTE: Your SPN must be account admin to do this. You must also setup the workspaceclient to do this.

from databricks.sdk.service.catalog import AzureManagedIdentity

storage_credential_name = '<CREDENTIAL_NAME>'
comment = '<COMMENT>'
connector_id = '<DATABRICKS_ACCESS_CONNECTOR>'
az_mi = AzureManagedIdentity(access_connector_id = connector_id)

w.storage_credenditals.create(
  name = storage_credential_name,
  azure_managed_identity = az_mi
  comment = comment
)