In February, with the release of MariaDB ColumnStore 1.1.3, we introduced a new Apache Spark connector (Beta) that exports data from Spark into MariaDB ColumnStore. The Spark connector is available as part of our MariaDB AX analytics solution and complements our suite of rapid-paced data ingestion tools such as a Kafka data adapter and MaxScale CDC data adapter. The connector empowers users to directly export machine learning results stored in Spark DataFrames to ColumnStore for high performance analytics. Internally, it utilizes ColumnStore's Bulk Data Adapters to inject data directly into MariaDB ColumnStore's WriteEngine.
In this blog, we'll explain how to export the results of a simple machine learning pipeline on the classification example of the well-known mnist handwritten digits dataset. Feel free to start your own copy of our lab environment by typing:
No comments:
Post a Comment