[Auscope-geosciml] RE : WFS filter woes

Boisvert, Eric Eric.Boisvert at RNCan-NRCan.gc.ca
Tue Mar 2 18:00:35 EST 2010


WFS 1.0.0 : no
WFS 2.0.0 : yes.
 
For GIN, we try to implement a solution using XPath predicates similar to this

________________________________

De: auscope-geosciml-bounces at lists.arcs.org.au de la part de Stephen M Richard
Date: mar. 2010-03-02 15:52
À: auscope-geosciml at lists.arcs.org.au
Objet : Re: [Auscope-geosciml] WFS filter woes


just a wild guess--
can the ogc:propertyName include an xpath filter ?

<ogc:Filter> 
        <ogc:And> 
        <ogc:PropertyIsEqualTo> 
                <ogc:PropertyName>subFeatureProperty/SubFeature[propB = '40']/propA</ogc:PropertyName> 
                <ogc:Literal>10</ogc:Literal> 
        </ogc:PropertyIsEqualTo>
</ogc:Filter> 

steve


On 3/2/2010 1:34 PM, Boisvert, Eric wrote: 

	This question is for the WFS guys that might be on this list.  I tried to find a answer or at least validation of what I think is right.

	I am struggling with an issue about boolean operator in OGC filter. We try to implement a use case that requires to select all WaterWell (Borehole) that has an interval with a certain porosity value and a certain conductivity value.  The problem is how the AND is scoped.  Let's see this example:


	<?xml version="1.0" encoding="UTF-8"?> 
	<Feature id="1"> 
	        <topProperty>100</topProperty> 
	        <subFeatureProperty> 
	                <SubFeature id="1000"> 
	                        <propA>10</propA> 
	                        <propB>20</propB> 
	                        <propB>30</propB> 
	                </SubFeature> 
	        </subFeatureProperty> 
	        <subFeatureProperty> 
	                <SubFeature> 
	                        <propA>11</propA> 
	                        <propB>0</propB> 
	                </SubFeature> 
	        </subFeatureProperty> 
	        <subFeatureProperty> 
	                <SubFeature> 
	                        <propA>16</propA> 
	                        <propB>40</propB> 
	                </SubFeature> 
	        </subFeatureProperty> 
	</Feature> 

	And this filter: 

	<wfs:Query typeName="Feature"> 
	<ogc:Filter> 
	        <ogc:And> 
	        <ogc:PropertyIsEqualTo> 
	                <ogc:PropertyName>subFeatureProperty/SubFeature/propA</ogc:PropertyName> 
	                <ogc:Literal>10</ogc:Literal> 
	        </ogc:PropertyIsEqualTo> 

	        <ogc:PropertyIsEqualTo> 
	                <ogc:PropertyName>subFeatureProperty/SubFeature/propB</ogc:PropertyName> 
	                <ogc:Literal>40</ogc:Literal> 
	        </ogc:PropertyIsEqualTo> 

	        </ogc:And> 
	</ogc:Filter> 
	</wfs:Query> 

	Am I right to consider this condition as being equivalent to 

	<xsl:copy-of select="Feature[subFeatureProperty/SubFeature/propA = '10' and subFeatureProperty/SubFeature/propB = '40'] "/> ?

	If so, the above example should match, because the scoping is the root element, so there is a path from the root element that satisfy both condition.

	My point here is that is no real way to say that I want the AND to be scoped in a given <SubFeature> (propA = 10 and probB = 40 in a given SubFeature).

	But the issue is that once this request will be converted in SQL, the AND will be scoped depending on where the propA and propB will be (in which table) and how the JOIN are done.  For instance, if the propA and propB are in some leaf table, it will scope to the last table prior in the join chain.

	The SQL would be equivalent to (scoped to SubFeature) 

	<xsl:copy-of select="Feature[subFeatureProperty/SubFeature/[propA = '10' and propB = '40']] "/>  (and this return an empty set as expected)

	While this matches the above example 
	<xsl:copy-of select="Feature[subFeatureProperty/SubFeature/[propA = '10' and propB = '20']] "/>  


	I'm not sure if : 

	A) there is a implied scoping rule in the way we encode Xpath in WFS that is different than "root node".  I don't think the "last class/data type" should be a scoping rule because, for example, all condition involving CGI_Value will scope to the content of CGI_Value.

	B) We should force the database implementation to follow the "root" scoping rule, otherwise, the result will depend of database implementation.

	WFS / Filter 2.0 does not solve the problem with the matchAction, because it does not specific the scoping either. 

	Eric 




	================================================================ 
	Eric Boisvert 
	Spécialiste TI-GI / IT-IM specialist 
	Eric.Boisvert at rncan.gc.ca, 418-654-3705, facsimile/télécopieur 
	418-654-2615 
	490, rue de la Couronne, Québec (Québec), G1K 9A9 
	490, rue de la Couronne, Quebec, Quebec, G1K 9A9 

	Laboratoire de cartographie numérique et de photogrammétrie (LCNP) 
	Digital Cartography and Photogrammetry Laboratory (DCPL) 
	Commission géologique du Canada (Québec) / Geological Survey of Canada (Quebec) 
	Ressources naturelles Canada / Natural Resources Canada 
	Gouvernement du Canada / Government of Canada 
	http://cgc.rncan.gc.ca/org/quebec <http://cgc.rncan.gc.ca/org/quebec>  
	http://cgc.rncan.gc.ca/dir/index_f.php?id=4186 <http://cgc.rncan.gc.ca/dir/index_f.php?id=4186>  / http://cgc.rncan.gc.ca/dir/index_e.php?id=4186 <http://cgc.rncan.gc.ca/dir/index_e.php?id=4186>  




	
	_______________________________________________
	Auscope-geosciml mailing list
	Auscope-geosciml at lists.arcs.org.au
	http://lists.arcs.org.au/cgi-bin/mailman/listinfo/auscope-geosciml
	  


-- 
Stephen M. Richard
Section Chief, Geoinformatics
Arizona Geological Survey
416 W. Congress St., #100
Tucson, Arizona, 85701 USA

Phone: 
Office: (520) 209-4127
Reception: (520) 770-3500 
FAX: (520) 770-3505

email: steve.richard at azgs.az.gov



More information about the GeoSciML mailing list