On-the-fly data set combinations with RNTuple

With the expected data volume increase for HL-LHC and the even more complex computing challenges set by future colliders, the need for efficient data storage and processing becomes more pressing. ROOT’s next-generation data format and I/O subsystem, RNTT uple , is designed to address these challenges. RNTT uple already demonstrates a clear improvement in storage and I/O efficiency, as well as overall stability and robustness with respect to its predecessor, TTT ree . These improvements provide a solid baseline to introduce novel extensions to common high-energy and nuclear physics (HENP) workflows. Notably, many workflows could benefit from the ability to arbitrarily join and chain data set samples at runtime, which could reduce overall storage requirements and improve application runtime and ergonomics. In this paper, we present the RNT uple P rocessor , which enables HENP data set combinations with RNT uple . We will discuss the main design considerations, present the interfaces to support data set combinations and show how they integrate in typical workflows.