[gdal-dev] GML / NAS code redundancy

[Even Rouault]

Martin,

gmlregistry.diff looks fine, but it would be good to document elementValue in 
drv_gml.html

As far as gmlhandler.diff is concerned, serveral issues :
- naming convention : "CPLString pszFilteredElementName" should be "CPLString 
osFilteredElementName"
- if ( pszElementName[m_poReader->GetState()->osPath.size()] == '|' ) : there 
should be an extra test to check that strlen(pszElementName) > m_poReader-
>GetState()->osPath.size()
- the casts to (char*) are useless

All in all, I think that the more simpler solution would be to drop that patch 
and rather modify the call to SetFilteredClassName() at line 130 of 
ogrgmllayer.cpp to something like this :

if (poDS->GetLayerCount() > 1 && poDS->GetReadMode() == STANDARD)
{
    const char* pszElementName = poFClass->GetElementName();
    const char* pszLastPipe = strrchr(pszElementName, '|');
    if( pszLastPipe != NULL )
         pszElementName = pszLastPipe + 1;
    poDS->GetReader()->SetFilteredClassName(pszElementName);
}

This is completely untested of course...

Even

> Dear Even,
> 
> 2014-03-19 11:17 GMT+01:00 Even Rouault <even.rouault at mines-paris.org>:
> > Just for Martin's enlightment, the GML driver manages to recognize
> > several layers in that case since the RekisteriyksikonTietoja is the top
> > feature present in the first bytes of the file, and the other feature
> > types are sub feature types, also defined in
> > RekisteriyksikonTietoja.xsd. In the case, the schema were simple enough
> > so that the XSD parser could understand them.
> > 
> > For Martin's case, I think that a .gfs will be necessary to extract
> > properties in nested XML elements.
> 
> thanks a lot for your notes, it helped a lot. When working on support
> for this format in OGR I found two problems.
> 
> First is related to GML registry. The input file contains different
> layers based on its internal type (TypSouboru). For each of these
> types are prepared separated GFS file.
> 
>     <!-- Czech VFR schame-->
>     <namespace prefix="vf"
>                uri="urn:cz:isvs:ruian:schemas:VymennyFormatTypy:v1
> ../ruian/xsd/vymenny_format/VymennyFormatTypy.xsd"
>                useGlobalSRSName="true">
>         <featureType elementName="TypSouboru"
>                      elementValue="OB_UKSH"
>                      gfsSchemaLocation="vfr_ob_uksh_v1.gfs"/>
>         <featureType elementName="TypSouboru"
>                      elementValue="ST_UKSH"
>                      gfsSchemaLocation="vfr_st_uksh_v1.gfs"/>
>         <featureType elementName="TypSouboru"
>                      elementValue="ST_UKSG"
>                      gfsSchemaLocation="vfr_st_uksg_v1.gfs"/>
>     </namespace>
> 
> I have added a new attribute to 'osElementValue' to
> 'GMLRegistryFeatureType' class (see attached 'gmlregistry.diff' file).
> Then the library also checks value of 'elementName' and if it's
> defined and than uses related GFS.
> 
> The second problem is related to class and tag names. The layer names
> are in plural form, features in singular form, eg.
> 
> <data>
>   <cities>
>    <city>
>    ...
>    </city>
>   </cities>
> </data>
> 
> To allow GML driver to create layer 'cities' I need to modify
> 'GMLHandler::startElementDefault()' method, see 'gmlhandler.diff'.
> 
> Thanks for any comments in advance, Martin

-- 
Geospatial professional services
http://even.rouault.free.fr/services.html