Chinese tuples data will lead to messy code，such as "%E5%9E%8B%E7%B3%96%E5" about d2rq HOT 1 OPEN

There are two different encodings here. \u7259\u89E6\u75DB and %E7%89%99%E8%A7%A6%E7%97%9B.

D2RQ outputs an older version of the N-Triples format where non-ASCII characters had to be escaped as \uXXXX Unicode escapes. This is why we see \u7259\u89E6\u75DB. That should not be a problem. When the N-Triples file is read into another RDF software, these characters will be unescaped and turned back into the original Chinese characters. (In newer versions of Turtle, this escaping is no longer necessary, but it shouldn't be a problem.)

The %E7%89%99%E8%A7%A6%E7%97%9B encoding is only used when D2RQ generates URIs with non-ASCII characters in them. This is arguably a bug in D2RQ. URIs can contain only US-ASCII characters, and therefore D2RQ applies percent-encoding to characters outside the US-ASCII range. But URIs in RDF (and hence N-Triples) are actually IRIs, which support the full Unicode range. So D2RQ shouldn't do the %-encoding.

from d2rq.