Mapserver WFS ignores PropertyName outside Filter

[bartvde@…]

When a performing a GetFeature like the following:

<GetFeature service="WFS" version="1.0.0"
  maxFeatures="20" 
  xmlns="http://www.opengis.net/wfs" 
  xmlns:gml="http://www.opengis.net/gml" 
  xmlns:ogc="http://www.opengis.net/ogc" 
  xmlns:xlink="http://www.w3.org/1999/xlink" 
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
  xsi:schemaLocation="http://schemas.opengis.net/wfs/1.0.0 WFS-basic.xsd">
  <Query typeName="Gemeentes">
    <PropertyName>GEMCODE</PropertyName>
    <ogc:Filter>
      <ogc:BBOX>
      	<PropertyName>MS_GEOMETRY</PropertyName>
      	<gml:Box><gml:coordinates>121000,469000 
140000,470000</gml:coordinates></gml:Box>
      </ogc:BBOX>
    </ogc:Filter>
  </Query>
</GetFeature>

Mapserver should return only the GEMCODE attribute, it returns the full set.

This is what the WFS 1.0 spec says about the behaviour:

The <PropertyName> element is used to enumerate the feature properties that 
should be selected during a query and whose values should be included in the 
response to a GetFeature request. A client application can determine the 
properties of a feature by making a DescribeFeatureType request before 
composing a GetFeature request. The DescribeFeatureType operation [sec. 8] will 
generate a GML application schema defining the schema of the feature type. The 
client can then select the properties to be fetched. In addition, the client 
can determine which feature properties are mandatory and must be fetched in 
order for the WFS to be able to generate an instance of the feature type that 
will validate againt the generated GML application schema. In the event that a 
WFS encounters a query that does not select all mandatory properties of a 
feature, the WFS will internally augment the property name list to include all 
necessary property names. A WFS client must thus be prepared to deal with a 
situation where it receives more property values than it requests.
If no <PropertyName> elements are specified, then all feature properties should 
be fetched.

[assefa]

Setting this as a Future enhancement. Tasks to be done :

  - extact the <Query typeName tga(s) and build a list of attributes
  - use the list of attributes in the msGMLWriteWFSQuery function to limit
    the output.

[bartvde@…]

The same applies to the HTTP GET interface.

E.g.:

http://www.someserver.com/wfs.cgi&SERVICE=WFS&VERSION=1.0.0&REQUEST=GetFeature&PROPERTYNAME=INWATERA_1M/WKB_GEOM,INWATERA_1M/TILE_ID&TYPENAME=INWATERA_1M

[jdoyon@…]

I'm going to vote for this :)

Also I don't know if the spec says anything specific about this, but it'd be
nice if the geometry could also be treated equally.

In my use case, I'm interested only in the bounding box of a feature, so I could
suppress everything else easily.  And considering how the actualy geometry can
easily be biggest component of a FeatureMember, this could help improve
performance quite a bit.

I wonder if I can gml_exclude "msGeometry"?? I'm going to go try that, it not,
that'd be nice too :)

[bartvde@…]

Assefa, how much work would be involved in this?

Also, could it be possible to only query the database for the columns needed?

[dmorissette]

Reassigned to Assefa. Please update status and/or milestone.

[assefa]

won't be done for 5.0.

[bartvde]

I am working on a patch for this issue. Hopefully I can cook up something soon.

[bartvde]

patch for HTTP POST against 5.2beta4

[assefa]

Bart,

Are the patch submitted in this bug complete? I seem to only see modifications done in mapows.h.

I am going to work on this bug (support of PropertyName and also the possibility to execlude geometries from the result) I am also wondering if the the geometry exclusion would be better done using a specific tag like gml_exclude_geometry.

Thx

[bartvde]

Hey Assefa, it is complete, but Trac only shows the first file. Please download the patch instead in its original format [1]

[1] ​http://trac.osgeo.org/mapserver/attachment/ticket/675/mapserver-5.2.0-675.patch

I think we need to decide on how METADATA relates to the WFS protocol definition, I think the WFS protocol should always override METADATA.

[bartvde]

I just found out this patch currently does not incorporate alias fields.

[bartvde]

new patch incorporating alias fields

[assefa]

committed in trunk. Added msautotest

[dmorissette]

(Changing target milestone since trunk is currently 5.6)

[bartvde]

I've tested this and it works fine.