1 Reply by RobLarge

  • RobLarge
  • RobLarge
  • Veteran
  • Offline
  • From: Wiltshire
  • Registered: 15-07-2008
  • Posts: 812

Re: Polygon reports XML

Would it be possible to obtain any further information about how the polygon reports actually work.

I have looked at the XML code and it is certainly not clear from here what the underlying process is doing. I should proabably point out that we have not yet imported any polygon data so I haven't yet tried running this report to see what results are produced

With particular reference to e.g. P5 (locations in polygon with partial overlaps). What I would like to understand is; does the query perform a full gis comparison of the selected polygon and any other polygons to detect overlaps and return location keys, or is it some other method (such as a simple bounding-box comparison or a comparison of grid squares associated with locations).

I am hoping to produce a series of more detailed xml reports tailored more closely to our needs, which would, for example return ALL records within a polygon (both those with a spatial ref in the polygon and those associated with any location polygons which overlap the specified polygon). In order to do this however it would help if I understood the process.

Rob Large
Wildlife Sites Officer
Wiltshire & Swindon Biological Records Centre

RobLarge's Website

2 Reply by MikeWeideli

  • MikeWeideli
  • Administrator
  • Offline
  • From: Littlefield Consultancy
  • Registered: 05-07-2006
  • Posts: 2,627

Re: Polygon reports XML

Rob

As far as I know the Locations are extracted based on the grid reference of the Location and the grid squares are not involved. Report 5 will, therefore,  pick up the Location where any part of the grid square as defined by the grid reference of the Location falls within the Polygon. This is based on an algorithm which is close to that used by a GIS system.  As Location Grid references  are likley to be site centroid then it is possible that Location overlapping the boundary will not be included.  Using Sample Grid Ref is probably a better approach as Samples  Grid Refs should not generally be site centroids.


Mike   





.

Mike Weideli

MikeWeideli's Website

3 Reply by RobLarge

  • RobLarge
  • RobLarge
  • Veteran
  • Offline
  • From: Wiltshire
  • Registered: 15-07-2008
  • Posts: 812

Re: Polygon reports XML

Thanks Mike, that's pretty much what I thought. It just confirms that I can't achieve what I want from within Recorder. Looks like it's back to the GIS for me.

I don't agree with your suggestion about the Sample grid refs though. Where the sample is associated with a location it will normally have the same grid ref as the location, unless a supplementary reference has been added for the sample.

It seems to me that this is a major flaw in Recorder to date. In as much as it is possible, indeed entirely normal, to have occurrences associated with a location (implying that the taxon occurs somewhere, or possibly everywhere within the boundary of the site) and it is also possible to associate the location with a boundary in the map window and yet there is no mechanism for doing the kind of polygon overlap query which would return such occurences in a situation where the nominal spatial reference of the location falls outside the search area.

In effect this gives an importance to the location spatial reference (or site centroid) which is undeserved. As you know, it is possible in certain circumstances to generate a centroid which actually falls outside the site boundary (indeed with linear sites such as small rivers, this is almost inevitable). In this situation the whole location and all its occurrences may be missed by the query.

Rob Large
Wildlife Sites Officer
Wiltshire & Swindon Biological Records Centre

RobLarge's Website