search

marcelotto / spread2rdf

DSL-based converter for spreadsheets to RDF
http://rubygems.org/gems/spread2rdf
MIT License
4 stars 2 forks source link

Invalid RDF/XML output #6

Closed GitHubber1 closed 10 years ago

GitHubber1 commented 10 years ago

Hello Marcel,

I am trying to use your tool for transforming excel-based data to RDF/XML. The problem is that the output XML is not valid because of missing XML end tags. This is the case for property elements on a node element, e.g. skos:broader and skos:InScheme tags are not closed in the following snippet:

<skos:Concept rdf:about='http://www.example.com#Whatever'>
    <rdf:type rdf:resource='http://www.example.com#MyType'></rdf:type>
    <myprop:example>value</myprop:example>
    <skos:broader rdf:resource='http://www.example.com#Broader'>
    <skos:inScheme rdf:resource='http://www.example.com#inScheme'>
  </skos:Concept>

Using any other format (nt, n3, ttl, xml (TriX), json) validation is successful and I can add and query the data in a triple store.

Does anybody else have this problem?

Thank you, Andy

marcelotto commented 10 years ago

Hi Andy,

since the other formats are generated correctly, it might be a problem in RDF.rb's RDF/XML writer. To ensure it's not a problem with Spread2RDF, please try to convert one of the valid output files to RDF/XML with the following command:

$ rdf serialize your-file.ttl --output-format rdfxml -o your-file.rdf

If the RDF/XML produced with this command-line interface of RDF.rb is still invalid, I would encourage you to raise an issue on the RDF/XML writer. Otherwise I can't provide any further help without the input data and the mapping file.

Cheers, Marcel

GitHubber1 commented 10 years ago

Hello Marcel, thank you very much for your fast help. Converting the .ttl file with the command line also results in an invalid RDF/XML file. I raised an issue here. Best regards, Andy

marcelotto commented 10 years ago

To check wether the problem is caused by Spread2RDF and work on it, I must reproduce the problem. Can you provide the mapping and, if your data is sensitive, some dummy data exhibiting the problem, possibly by personal mail?