Introduction
The Databricks integration enables you to ask questions in natural language, which are automatically translated into SQL to analyze your data directly from Databricks. See example result below for the query “give me progression in temperature for Havana and plot it” for a user who connected Sana to their database with weather data.
Integration capabilities
With this integration, you can perform real-time read operations on your Databricks data within Sana Agents. There’s no need for regular data syncing as queries are executed live. The integration allows you to select and describe which Databricks views you want to make accessible in Sana.
Operation type | Supported |
Read | ✅ |
Write | ❌ |
Type of integration
Private: Set up per user and mirroring the underlying individual access rights
Availability
Integration type | Free tier | Team tier | Enterprise tier |
Private | ❌ | ❌ | ✅ |
Shared | ❌ | ❌ | ❌ |
Centralized | ❌ | ❌ | ❌ |
Integration set-up
Pre-requisite for workspace admins before any integration can be set-up:
Generate a Databricks Oauth Client ID and secret by following step 1) in this guide: https://docs.databricks.com/aws/en/dev-tools/auth/oauth-m2m. When prompted, use https://sana.ai/x-api/integration/databricks-oauth as the redirect URL.
Follow this guide to add your Databricks Oauth app in Sana: https://support.sana.ai/en/articles/265266-custom-oauth-app-for-integrations-guide
If these steps are not followed, you will receive the following error when trying to connect:
1. Find and click on “Databricks” integration inside of the list of available integrations.
2. Click “Connect just for me”
3. Add your Databricks credentials and click “Continue”.
4. Now that you have added your credentials you need to sign in by clicking “Connect just for me”.
5. Select the views you want to be able to access.
Click “Add view,” then choose “Browse views” to import existing ones or “Create view” to define a new view from scratch.
Browse views: Select the views you want, and click “Import views.”
Create view manually: Enter the required fields and your query, click “Sync,” and once successful, click “Create view.”
Click “Continue” after all your views have been successfully synced.
Tips and best practices when creating views:
To edit or delete a view, click the three dots next to its status badge.
Add descriptive titles, descriptions, and column details to help the agent generate accurate SQL queries and make data discovery easier.
5. You are now good to go!
Known limitations
Sana Agents does not support write operations to Databricks (i.e., creating and updating data)
FAQ
Q: What is a view?
A: A view can be described as a saved SQL query in Databricks used for simplify sharing of common data that you want to make easily accessible by others. Inside of Sana you can then choose to import a view to make it part of your integration to Databricks.
Q: How often is the data synced?
A: The data is synced in real-time.
Q: Can I create, update or delete data in Databricks from Sana?
A: No, this integration currently only supports read operations.
Data handling & privacy
Sana AI is fully committed to data security and privacy. All data accessed by Sana AI is encrypted both in transit and at rest. Sana does not train any underlying language models on your data, ensuring the privacy of your information. Sana AI is ISO 27001 certified; and SOC 2 and GDPR compliant, and adheres to the highest standards of data security.