Till Grallert#MarcXML encapsulated as a string in #JSON is a pain to work with as I am finding out trying to query #HathiTrust APIs for digitised items with a list of #OCLC numbers obtained via #SPARQL from #Wikidata.
Things I wish for:
1. direct access to the MARCXML from HathiTrust
2. a publicly accessible API for #Worldcat / OCLC to find holdings