Hi,
 
 
This might be a solution.
 
-----
 
Add temporary ID (e.g. _tmp_id) to every polygon with the Counter.
 
Branch the data flow into two flows, connect one to the Requestor port, another to the Supplier port of the SpatialRelator.
 
Parameters of the SpatialRelator are:
 
Tests to Perform: WITHIN
 
Related Suppliers Count Attribute: _related_candidates
 
Attribute(s) that Must Differ: _tmp_id
 
 
The largest extent polygons should not be within any other polygons. i.e. _related_candidates of them should be equal to 0.
 
So, you can filter the output polygons with the Tester to get the required polygons.
 

 
 
Takashi
                
     
                                    
            Takashi
 
 
This works really well. 
 
I have found a few where the largest extent is intersected slightly by another polygon. In this instance the tester removes all polygons. Do you know if there is a way of finding these and just taking the largest area polygon in the group with the attributes? 
 
 

 
 
Many thanks 
 
Seb 
                
     
                                    
            The Dissolver keeps attributes from one of the dissolved features with largest area.
 
See help on the Dissolver.
 
So, try adding these transformers after the Tester.
 
-----
 
Dissolver
 
Keep Input Attributes: Yes
 
FeatureMerger
 
Join On: "_tmp_id"  to "_tmp_id"
 

 
                
     
                                    
            P.S. If the Dissolver works efficiently, the SpatialRelator and the Tester may not be necessary.
 
