Incorporating Functions in Mappings to Facilitate the Uplift of CSV Files into RDF

Many solutions have been developed to convert non-RDF data to RDF. A common task during this conversion is applying data manipulation functions to obtain the desired output. Depending on the data format of the source to be transformed, one can rely on the underlying technology, such as RDBMS for relational databases or XQuery for XML, to manipulate data to a certain extent while generating RDF. For CSV files, however, there is no such underlying technology. Instead, one has to resort to more elaborate Extract, Transform and Load (ETL) processes, which can render the generation of RDF more complex (in terms of number of steps), and therefore also less traceable and transparent. One solution to this problem is the declaration and inclusion of functions in mappings of non-RDF data to RDF. In this paper, we propose a method to incorporate functions into mapping languages and demonstrate its viability in Digital Humanities use case.