GMLWriter output not valid?? about nettopologysuite HOT 11 CLOSED

GMLWriter writes a GML "representation" of a Geometry: a GML fragment and not a complete GML document.
To create a "valid" GML document, you (or we) should create a writer that, at least, takes a list of IFeature, then manage geometry AND attributes to generate A GML document, that at least contains initial xml declaration and namespace imports.

from nettopologysuite.

Hi, thanks for the feedback. I managed to use sharpkml to get a quick usable output, but it would be good to see an export to gml file option.

from nettopologysuite.

If I got it right, the problem is the wrapping <gml:GML></gml:GML> tag that is created using the writer. At least JTS does not write that in its xml.

from nettopologysuite.

Have you tried to read JTS GML output with QGIS?
I suspect you can obtain another error, jost becaus the GML fragment isn't a GML/XML valid document.
Anyway, remove gml namespace can be a valid strategy, at least to maintain JTS behavior.

from nettopologysuite.

Having said that, I think the API should change and the GMLWriter.Write functions should take an XmlWriter as argument, thus knowing which namespaces have already been defined to not pollute the file with namespace declarations.

from nettopologysuite.

As a comparison:
JTS:

<gml:Point>
  <gml:coordinates>
    52,-0.9 
  </gml:coordinates>
</gml:Point>`

NTS:

<gml:GML xmlns:gml="http://www.opengis.net/gml">
  <gml:Point>
    <gml:coord>
      <gml:X>52</gml:X>
      <gml:Y>-0.9</gml:Y>
    </gml:coord>
 </gml:Point>
</gml:GML>

from nettopologysuite.

I'm working on a pull request.
Actually, I don't remember the meaning of gml:GML tag, I remember that code was modified for MSSQLSpatial purposes...
Anyway, code fixes to made output like JTS one is trivial.
I don't actually know how many existing code this change can break.

from nettopologysuite.

+1 for change the API

from nettopologysuite.

@FObermaier @RobertSmart please check. Hope this fix the problem.

from nettopologysuite.

can I close this issue now?

from nettopologysuite.

I haven't had time to check, but if matches the JTS implementation, then it should be fine to close now.

Thanks, Rob

from nettopologysuite.