All, We have an observation that doing a "Window Select" in MapGuide 2.1 (AJAX viewer) queries on the feature bounding boxes and not the feature geometry for Oracle (KingOra) connections. The following is my understanding and if anyone can confirm/replicate or advise that would be much appreciated.
Briefly: Within "ajaxmappane.templ" there is a function called 'QueryFeatureInfo(...)' that implements the MapGuide HTTP API call for OPERATION=QUERYMAPFEATURES. This is documented here: http://trac.osgeo.org/mapguide/wiki/HttpApi/RenderingService#QueryMapFeatures2.1 The current (default) value for the SELECTIONVARIANT value is "INTERSECTS" - in theory we can simply change this to one of the valid values: http://trac.osgeo.org/mapguide/wiki/HttpApi/RenderingService#SelectionVariant1.0 But the implication is that the default value of INTERSECTS is what is required for accurate selection (rather then ENVELOPEINTERSECTS). However, selected objects in an Oracle-based layer visually are outwith the selection rectangle. Turning on logging in the KingOracle provider shows that the query sent to the DB is the fast SDO_FILTER and not the accurate SDO_RELATE. However, the place where SDO_FILTER is built into KingOra is very clearly in the switch/case for FdoSpatialOperations_EnvelopeIntersects so this is how the query is processed! So, what is happening between INTERSECTS passed as a param to QUERYMAPFEATURES and it being received/interpreted as ENVELOPEINTERSECTS in KingOracle (other providers not checked)??? Crispin -- View this message in context: http://osgeo-org.1803224.n2.nabble.com/Window-Selection-Ajax-as-ENVELOPE-or-INTERESECTS-Confusion-tp5389467p5389467.html Sent from the MapGuide Users mailing list archive at Nabble.com. _______________________________________________ mapguide-users mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/mapguide-users
