I use rdflib to connect to a Blazegraph triplestore using the SPARQLUpdateStore.

When I add data containing non-ASCII characters using addN() it results in garbled data in the triplestore (UTF-8 interpreted as ISO-8859).

The problem seems to be that SPARQLConnector.update() does not set "charset=UTF-8" in the Content-Type header:

but it does use UTF-8 in data=query.encode() in line 183.

The problem goes away if I patch the Content-Type to be "application/sparql-update; charset=UTF-8".

I can provide a PR with this one-line change if you like.