Frequently Asked Questions (FAQ) about CData MCP Server for Databricks

Q: What is the CData MCP Server for Databricks? A: It’s a read-only MCP server that allows LLMs (like Claude Desktop) to query live data from Databricks using natural language. It leverages CData JDBC Drivers to expose Databricks data as relational SQL models.

Q: What is MCP? A: MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs, enabling them to interact with external data sources and tools.

Q: Do I need to know SQL to use this server? A: No, LLMs can query Databricks data using natural language questions, eliminating the need for complex SQL queries.

Q: Is this server read-only? A: Yes, this particular MCP Server is read-only. For read/write capabilities, check out the free CData MCP Server for Databricks (beta) at https://www.cdata.com/solutions/mcp

Q: What data sources are supported? A: This server is specifically designed for Databricks. However, CData offers JDBC Drivers for hundreds of data sources (listed in the documentation), allowing you to connect to a wide range of data with similar MCP Servers.

Q: How do I set up the server? A: The setup involves cloning the repository, building the server, installing and licensing the CData JDBC Driver for Databricks, configuring the connection, creating a .prp file, and configuring your AI client (e.g., Claude Desktop).

Q: Where can I find the CData JDBC Driver for Databricks? A: You can download it from https://www.cdata.com/drivers/databricks/download/jdbc

Q: What is a .prp file? A: It’s a property file that contains the configuration settings for your JDBC connection, including the driver path, driver class, JDBC URL, and other parameters.

Q: How do I configure Claude Desktop to use the MCP server? A: You need to create or modify the claude_desktop_config.json file to add an entry for the MCP server, specifying the command and arguments to run the server.

Q: What if I can’t see the CData MCP Server in Claude Desktop? A: Make sure you have fully quit and restarted Claude Desktop. Sometimes, you need to use Task Manager (Windows) or Activity Monitor (Mac) to ensure it’s completely closed.

Q: What if Claude Desktop is unable to retrieve data? A: Verify that you have configured your connection properly using the Connection String builder. Double-check the JDBC URL in your .prp file.

Q: Where can I get help if I’m having trouble? A: For issues connecting to your data source, contact the CData Support Team. For issues using the MCP server, join the CData Community.

Q: Is the server licensed? A: This MCP server is licensed under the MIT License.

Q: Can I use this with UBOS Platform? A: Yes! CData MCP Server for Databricks seamlessly integrates with the UBOS platform, a full-stack AI Agent Development Platform. This integration allows you to connect this server and all your Databricks data for your AI Agents.

Q: What are the benefits of using this server with UBOS? A: This integration empowers your AI agents to deliver actionable insights, automate complex tasks, and ultimately, contribute to a more efficient and intelligent business operation.