With a long history of data processing and interfacing with third-party databases, the J system now becomes complete with a data engine of its own. The new JDB addon provides a very efficient relational database written completely in J.
The rich set of J's structural and grouping operations is unleashed on relational data in a compact yet expressive SQL-like syntax inspired by the best of breed query notation. Relational concepts, such as unique keys and natural joins, are present in a straightforward way that is familiar to experts and simple for beginners.
The solid engine core with journaled updates provide atomicity and transactional integrity, while client/server interface delivers elements of concurrency. Efficiency is by design of physical storage where native J types are laid out in inverted tables accessed through mapped files. In addition, there is a mechanism for low-level bulk operations.
Adoption of JDB is meant to be easy: simple API for metadata definition and discovery, convenient data access routines and query syntax have been tested when creating many included sample databases. J native types and non-terse open source of JDB make for a flexible boundary between the internals and user code, allowing for simple extensibility.
Though already quite powerful in its initial phase, the released JDB beta can still be improved and enriched with additional features, so user feedback is crucial at this stage. Please enjoy and share your experience and ideas.