rdf_query works incorretly on UTF-8 characters

[ostroganov]

See attached example:
test.txt

rdf <- rdf_parse("test.txt", format = "nquads")
rdf

the results looks good:

Total of 1 triples, stored in hashes
-------------------------------
<http://subject> <http://predicate> "Error. Ошибка. 错误"@en .

dataframe produced by rdf_query however contains incorrect characters:

rdf_query(rdf,"SELECT ?s ?p ?o { ?s ?p ?o }")

# A tibble: 1 x 3
  s              p                o                          
  <chr>          <chr>            <chr>                      
1 http://subject http://predicate Error. Ошибка. 错误

[cboettig]

@gothub Any idea if this is a limitation of the underlying redland C libs? I believe we're requesting UTF-8 in the rdf_query() call to redland already,

https://github.com/ropensci/rdflib/blob/master/R/rdf_query.R#L57-L62

[gothub]

@cboettig how are the RDF graphs being constructed? redland can use the RDF language tag to specify encoding, for example, this test

[cboettig]

Construction is with redland::parseFileIntoModel, see https://github.com/ropensci/rdflib/blob/master/R/rdf_parse.R#L68. as OP shows, this part seems fine.

query is done by redland::librdf_query_results_to_string2, as shown in link above, and not with getNextResult, since the iterator had very poor performance on large graphs.

[Tesla2509]

Is there any solution to get correct characters?