Extending R2RML to a Source-independent Mapping Language for RDF

Although reaching the fifth star of the Open Data deployment scheme demands the data to be represented in RDF and linked, a generic and standard mapping procedure to deploy raw data in RDF was not established so far. Only the R2RML mapping language was standardized but its applicability is limited to mappings from relational databases to RDF. We propose the extension of R2RML to also support mappings of data sources in other structured formats (indicatively CSV, TSV, XML, JSON). Broadening further its scope, the focus is put on the mappings and their optimal reuse. The language becomes source-agnostic, and resources are integrated and interlinked at a primary stage.