Handling Data Transformations in Virtual Knowledge Graphs with RML View Unfolding

Although declarative data transformation functions defined in RML-FNML were implemented in KG materialization systems, they have not yet been studied or implemented in virtual knowledge graph systems. In this work we propose to translate RML-FNML mapping to RML Views, which can be transparently used by virtual knowledge graph systems without modifying them. We implemented a research prototype of RML-FNML to RML Views unfolding and applied it to GTFS data to show the feasibility of the approach.