Bug #197

SPARQL results char encoding

Added by Gregg - almost 10 years ago. Updated almost 10 years ago.

Status:New Start date:
Priority:Urgent Due date:
Assignee:Paula Gearon % Done:


Target version:SPARQL Query Engine


The SPARQL Protocol definition says that for the HTTP binding "the whttp:outputSerialization is application/sparql-results+xml with UTF-8 encoding, application/rdf+xml with
UTF-8 encoding." (Section 2.2) That's for XML results; I haven't found the equivalent requirement for JSON output, but for my application in any case full utf-8 support for json and xml is essential.

As a general matter (principle of least surprise), I think the expected behavior would be "encoding-in equals encoding-out", so if I populate a graph with utf-8 data, query results should be utf-8, no matter the output serialization. Alternatively, one could argue that the standard HTTP 1.1 Accept-Charset header should govern; since it is an HTTP binding, HTTP rules should apply.

The SPARQL Protocol definition doesn't explicitly address character encoding for the SOAP binding, but since SOAP is an HTTP protocol it should probably do utf-8 or honor the Accept-Charset header.


Updated by Gregg - almost 10 years ago

Oops. I guess that should read something like "the charset parameter of the Accept header should govern"; e.g. "Accept: application/sparql-results+json; charset=UTF-16". Accept-Charset doesn't apply.

Also, RFC 4627, "The application/json Media Type for JavaScript Object Notation (JSON)", says (Section 3, Encoding) " JSON text SHALL be encoded in Unicode. The default encoding is UTF-8." So it's not the sparql definition but the json definition that requires unicode and, effectively, utf-8.

Also available in: Atom PDF