Apache Spark continues to grow in popularity - due to advanced analytics/machine learning, high performance processing, real-time streaming and multiple language support. Big Data technology is adding more data processing options to an already long list of legacy databases and file systems. As a result, enterprises continue to look for effective and approachable ways to federate all these data sources to solve business information needs. One under-appreciated feature of Spark is its ability to help quickly and powerfully enable federated data access. This presentation will discuss and demonstrate using Spark to query/combine multiple disparate data sources. We will see how to access the various data sources from Spark, normalize to Spark RDDs and combine for processing. The demo will show combining sources such as HDFS, JSON files, HBase, Hive and PostgreSQL and write the result back to a Data Mart for analysis. Also we will show the use of SparkSQL to access federated data in Spark through the Spark Thrift Server using the the Tableau BI tool.