home / mcp / azure data lake storage mcp server

Azure Data Lake Storage MCP Server

Provides a read-only MCP server to query live Azure Data Lake Storage data via the CData JDBC Driver.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "cdatasoftware-azure-data-lake-storage-mcp-server-by-cdata": {
      "command": "java",
      "args": [
        "-jar",
        "/PATH/TO/CDataMCP-jar-with-dependencies.jar",
        "/PATH/TO/azure-data-lake-storage.prp"
      ]
    }
  }
}

You run a local, read-only MCP server that exposes live Azure Data Lake Storage data via the CData JDBC Driver. This lets large language models query your data in natural language without writing SQL, while keeping data access controlled and view-only on your setup.

How to use

Use an MCP client to connect to your local server. Start the server with the MCP configuration file you created, then load the server into your client so the AI can issue read queries and discover available tables and columns. You can ask natural language questions like what is the trend in a specific dataset, how many records meet a condition, or which datasets contain certain values. The server exposes a set of tools to list tables, list columns for a table, and run read-only queries against Azure Data Lake Storage.

How to install

Prerequisites you need before installation.

# Prerequisites
- Java 11+ (for Maven build and running the MCP server)
- Maven (for building the MCP JAR)
- Access to the CData JDBC Driver for Azure Data Lake Storage (download and license steps below)

Clone the project, build the MCP server, and prepare the JDBC driver and connection string as described.

git clone https://github.com/cdatasoftware/azure-data-lake-storage-mcp-server-by-cdata.git
cd azure-data-lake-storage-mcp-server-by-cdata
mvn clean install

After building, you will have the MCP JAR ready for use in the next steps.

Download and license the CData JDBC Driver for Azure Data Lake Storage, then configure a connection string and create a .prp file for the MCP server.

# Example steps (adjust paths to your environment)
# 1) Install the JDBC Driver from the official source (instructions include licensing)
# 2) Run the driver connection utility to create a connection string
java -jar cdata.jdbc.adls.jar
# Follow the prompts to test and copy the connection string

Create a properties file for the MCP server with your JDBC connection details and table scope. This example shows the required fields and structure.

Prefix=adls
ServerName=CDataADLS
ServerVersion=1.0
DriverPath=/path/to/cdata.jdbc.adls.jar
DriverClass=cdata.jdbc.adls.ADLSDriver
JdbcUrl=jdbc:adls:InitiateOAuth=GETANDREFRESH;
Tables=

Additional setup notes

The MCP server you build is local and read-only, designed to work with Claude Desktop or similar clients that support MCP. You run the server on the same machine as the client, and the client communicates via standard input/output (stdio) with the Java process.

Configuration and usage details

To connect Claude Desktop or another MCP-capable client, create a Claude config entry that launches the MCP server using the built JAR and your .prp file. You configure the client to run the following command, which starts the MCP server with the specified configuration.

java -jar CDataMCP-jar-with-dependencies.jar /PATH/TO/azure-data-lake-storage.prp

Troubleshooting

If you cannot see the MCP server in your client, ensure you have fully closed and relaunched the client so it rescans available servers. If data cannot be retrieved, verify your connection string is correct and that the .prp file points to the right driver and URL. If issues persist, contact CData Support or join the CData Community for guidance.

License

This MCP server is licensed under the MIT License. You may use, modify, and distribute the software in accordance with the terms of the MIT License.

All supported sources

This MCP server configuration is designed to query Azure Data Lake Storage via the CData JDBC Driver. It exposes the Azure Data Lake Storage data model as relational constructs through the MCP interface.

Available tools

adls_get_tables

Retrieves a list of tables available in the data source. The output is returned in CSV format with a header row.

adls_get_columns

Retrieves a list of columns for a specified table. The output is returned in CSV format with a header row.

adls_run_query

Executes a SQL SELECT query against the configured data source and returns the result set.