This error,
The processing instruction target matching \"[xX][mM][lL]\" is not allowed
occurs whenever I run an XSLT page that b
I had a similar issue with 50,000 rdf/xml files in 5,000 directories (the Project Gutenberg catalog file). I solved it with riot (in the jena distribution)
the directory is cache/epub/NN/nn.rdf (where NN is a number)
in the directory above the directory where all the files are, i.e. in cache
riot epub/*/*.rdf --output=turtle > allTurtle.ttl
This produces possibly many warnings but the result is in a format which can be loaded into jena (using the fuseki web interface).
surprisingly simple (at least in this case).