Dear all, We are happy to announce the release of RumbleDB 1.14.0 "Acacia" beta.
RumbleDB is a JSONiq engine for querying massive amounts of JSON/text/CSV/Parquet/Avro/LibSVM/Root/... datasets either on your local drive (with the cores of your laptop) or on a data lake like S3 or HDFS (with the power of a big cluster). RumbleDB newly features the definition of user-defined object types with the JSound Compact syntax (www.jsound-spec.org) co-designed with Dana Florescu, Cezar Andrei, Jonathan Robie and Pavel Velikhov. Just like JSONiq is largely inspired by XQuery, JSound is largely inspired by XML Schema and brings the best XML has to offer into the JSON and DataFrames world. https://rumble.readthedocs.io/en/latest/Types.html RumbleDB also now supports Machine Learning at scale and allows training and prediction on large amounts of data -- leveraging function items as standardized in XQuery 3.0. It is more stable and also supports generic pipelines. https://rumble.readthedocs.io/en/latest/RumbleML.html As a consequence, RumbleDB allows you to discover, clean up, prepare, normalize, validate and annotate your data to then perform training and prediction via your favorite Machine Learning pipelines, from the one and same high-level query language and in the popular Jupyter notebooks. It is free and open source as always (coming from the academic world), and ideal for teaching: https://systems.ethz.ch/education/rumble-for-teaching.html Enjoy! Kind regards, Ghislain Fourny _______________________________________________ talk@x-query.com http://x-query.com/mailman/listinfo/talk
