If I read the document correctly, there are 2 different ways to define XML attributes in the resulting JSON. Depending if there is  a text node or not. 

I would very much argue against this concept but add an object name for text objects.



    CommentAdd your comment...

    2 answers


      I've thought about your question and reopened the solved remarks:
      since we only have one mixed element (i.e comment) it is probably easier to make a specific mapping rather than a complicated generic mapping.

      The other 2 text only elements would also become simple arrays of string rather than nested arrays

      1. Peter Kleinheider

        As we are working on some projects where we use XJDF/JSON, I will show some examples on how the resulting JSON will look like. 

      2. Rainer Prosi

        I've also added a comparison table to the mapping page.

        feel free to chime in

      CommentAdd your comment...

      Discussing in Jira certainly makes sense and your proposal was discussed at the interop.

      Here is a copy of my comment in JDF-716

      You cal also create detailed issues and link them - see: JDF-839

      I agree, and in the end it is the choice of swallowing a frog or a toad...
      We had discussed that solution using "#text" but there are multiple issues:
      1.) you cant really use any key that maps to an invalid code variable name because of automagic json - code mappers.
      2.) there is no way to actually know whether "text" is text body or @text.

      obviously, these are not hard blockers, e.g. using "_text_" of some other certainly legal but improbable variable name.

      In the end, it boils down to flipping a coin, which solution is preferrable.
      Luckily, we are talking about 3 elements (Comment, AdressLine and OrganizationalUnit)

        CommentAdd your comment...