Performance Results of FlexRML in the KGCW Challenge 2024

The Knowledge Graph Construction Workshop introduced a challenge to evaluate the performance metrics of different RML interpreters using a set of standardized benchmarks. We participated in the challenge’s performance track with our RML interpreter, FlexRML, and report the median execution time and peak memory consumption over five runs on the provided virtual machine using the challenge tool. Through this challenge, we were able to identify weaknesses in FlexRML, such as its current support for CSV data only, lack of support for the latest RML vocabulary, and a crash that occurs when the system executing the mapping runs out of memory. These are issues that we plan to address in future releases of FlexRML.