Ready to get started?

Learn more about the Apache Spark MuleSoft Anypoint Connector or download a free trial:

Download Now

Access Spark Data in Anypoint Using SQL

Create a simple Mule Application that uses HTTP and SQL with the CData Mule Connector for Spark to create a JSON endpoint for Spark data.

The CData Mule Connector for Spark connects Spark data to Mule applications enabling read , write, update, and delete functionality with familiar SQL queries. The Connector allows users to easily create Mule Applications to backup, transform, report, and analyze Spark data.

This article demonstrates how to use the CData Mule Connector for Spark inside of a Mule project to create a Web interface for Spark data. The application created allows you to request Spark data using an HTTP request and have the results returned as JSON. The exact same procedure outlined below can be used with any CData Mule Connector to create a Web interface for the 170+ available data sources.

  1. Create a new Mule Project in Anypoint Studio.
  2. Add an HTTP Connector to the Message Flow.
  3. Configure the address for the HTTP Connector.
  4. Add a CData Spark Connector to the same flow, after the HTTP Connector.
  5. Create a new Connection (or edit an existing one) and configure the properties to connect to Spark (see below). Once the connection is configured, click Test Connection to ensure the connectivity to Spark.

    Set the Server, Database, User, and Password connection properties to connect to SparkSQL.

  6. Configure the CData Spark Connector.
    1. Set the Operation to 'Select with Streaming'.
    2. Set the Query type to Dynamic.
    3. Set the SQL query to SELECT * FROM #[message.inboundProperties.'http.query.params'.get('table')] to parse the URL parameter table and use it as the target of the SELECT query. You can customize the query further by referencing other potential URL parameters.
  7. Add a Transform Message Component to the flow.
    1. Map the Payload from the input to the Map in the output.
    2. Set the Output script to the following to convert the payload to JSON:
      %dw 1.0
      %output application/json
      ---
      payload
              
  8. To view your Spark data, navigate to the address you configured for the HTTP Connector (localhost:8081 by default) and pass a table name as the table URL parameter: http://localhost:8081?table=Customers
    The Customers data is available as JSON in your Web browser and any other tools capable of consuming JSON endpoints.

At this point, you have a simple Web interface for working with Spark data (as JSON data) in custom apps and a wide variety of BI, reporting, and ETL tools. Download a free, 30 day trial of the Mule Connector for Spark and see the CData difference in your Mule Applications today.