Spark connector for Azure Cosmos DB will enable real-time data science, machine learning, advanced analytics, and exploration over globally distributed data in Azure Cosmos DB by connecting it to Apache Spark. The connector will efficiently exploit the native Azure Cosmos DB managed indexes and enable updateable columns when performing analytics. It will also use push-down predicate filtering against fast-changing, globally-distributed data addressing a diverse set of IoT, data science, and analytics scenarios. Spark structured stream support using Cosmos DB change feed, query performance improvements, and support for the latest Spark version will also be included.
Source: Azure Roadmap