Encoding problem when retrieving result from QueryEngineHTTP

[Mark Hissink Muller <XM...@kombit.dk>]

Hi all,

I am trying to solve an encoding problem between a Java client, which is a Spring Boot server application that runs on Windows, and a rdf4j-server running on Tomcat 8 on Ubuntu 18.04 LTS.

I should retrieve correct data with Danish characters (e.g. "Fælles forståelse") from the server, but instead I get the ASCII-interpretation ("Fælles forstÃ¥else").

I've observed the following:
- The client application generates a Model with the correct UFT-8 encoding.
- The client application correctly inserts model in the server via the Atomgraph InsertDataBuilder and RDFConnectionFactory.connect(postUrl); insertConnection.update(insertRequest).
- The rdf4j-server then contains the correct context, as I can see in the rdf4f-workbench. (The presentation of the data in the workbench seems ASCII/ has garbled Danish characters, but when I double-click to edit the data and a form opens, I can see that the characters are represented correctly.)

But when I then query the repository from my client application again, I retrieve ASCII (or it seems), instead of the correct UTF-8 encoded data.

I use the following straightforward approach for the SELECT query:

Query query = QueryFactory.create(queryString);
QueryEngineHTTP qexec = QueryExecutionFactory.createServiceRequest(getEndpoint(), query);
    ResultSet resultSet = qexec.execSelect();
    while (resultSet.hasNext())
    {
        QuerySolution querySolution = resultSet.next();
        resultMapper.processQuerySolution(querySolution);
        ...
    }

I could not find a way to force the QueryExecutionFactory or QueryEngineHTTP to interpret the returned ResultSet as UTF-8.

Does anyone have a solution or idea how to address this...? Much appreciated.

Kind regards, Mark

Re: Encoding problem when retrieving result from QueryEngineHTTP

[Andy Seaborne <an...@apache.org>]

On 25/02/2020 15:08, Mark Hissink Muller wrote:
> Hi all,
> 
> I am trying to solve an encoding problem between a Java client, which is a Spring Boot server application that runs on Windows, and a rdf4j-server running on Tomcat 8 on Ubuntu 18.04 LTS.
> 
> I should retrieve correct data with Danish characters (e.g. "Fælles forståelse") from the server, but instead I get the ASCII-interpretation ("Fælles forstÃ¥else").

Looks like ISO-8859


> 
> I've observed the following:
> - The client application generates a Model with the correct UFT-8 encoding.
> - The client application correctly inserts model in the server via the Atomgraph InsertDataBuilder and RDFConnectionFactory.connect(postUrl); insertConnection.update(insertRequest).
> - The rdf4j-server then contains the correct context, as I can see in the rdf4f-workbench. (The presentation of the data in the workbench seems ASCII/ has garbled Danish characters, but when I double-click to edit the data and a form opens, I can see that the characters are represented correctly.)

Unclear.

> 
> But when I then query the repository from my client application again, I retrieve ASCII (or it seems), instead of the correct UTF-8 encoded data.

How are you displaying that in your application?
Are you running on MS Windows?

> 
> I use the following straightforward approach for the SELECT query:
> 
> Query query = QueryFactory.create(queryString);
> QueryEngineHTTP qexec = QueryExecutionFactory.createServiceRequest(getEndpoint(), query);
>      ResultSet resultSet = qexec.execSelect();
>      while (resultSet.hasNext())
>      {
>          QuerySolution querySolution = resultSet.next();
>          resultMapper.processQuerySolution(querySolution);
>          ...
>      }
> 
> I could not find a way to force the QueryExecutionFactory or QueryEngineHTTP to interpret the returned ResultSet as UTF-8.

Jena applies HTTP content negotation.

If you can provide a public URL, I'll look at the HTTP request.

Otherwise, please get the HTTP headers (use curl or wget)

> 
> Does anyone have a solution or idea how to address this...? Much appreciated.
> 
> Kind regards, Mark
>