Question

Pointonareaoverlayer

  • 13 February 2013
  • 4 replies
  • 10 views

I'm trying to use the PointOnAreaOverlayer.  I have a point file, and an area file that I have to run thru an attribute prefixer first.  The output from the attributeprefixer is then used as input to the overlayer.  Looking at the log, the overlayer is actually starting before the prefixer starts, so my spatial join does not work.  

 

 

Any suggestions?  

4 replies

Hi,

 

 

maybe you can use the FeatureHolder to regulate the feature flow.

 

 

Best regards,

 

Stefan
Badge +14
The order of things isn't always indicative of process order, for a better idea try using the 'Run Translation with Inspection' option in workbench as this will highlight how features are moving through your process. If you add lots of inspection points between all your desired Transformers you can see well what is actually happening:

 

 

More details here: http://evangelism.safe.com/fmeevangelist74/

 

 

To fix the issue it may be as simple as just re-ordering your two input files in the Navigator window of Workbench as this determines which is read first and therefore which data starts to hit the Transformers in the process first.
Badge +1
Have you routed both the inputs through an inspector to make sure that they overlap as expected in FME?
Userlevel 4
Badge +13
Actually the PointOnAreaOverlayer waits for ALL input to arrive and does not start doing its work until it is sure nothing more is coming its way. So order of arrival won't matter -- something else is going on here.  I agree that putting inspectors or watching what is going on with Feature inspection (see Dave's answer here http://fmepedia.safe.com/AnswersQuestionDetail?id=906a0000000cf83AAA along with the video he links) would be helpful.

 

 

Note -- in FME 2013 SP1, we're introducing an "Areas Come First" mode into this transformer, and then order will matter.  If you can arrange for areas to come first, most usually by reading them first, then the points that come in will be shot out immediately after being compared with the areas.  Result -- much faster runtimes, much much larger data volumes possible. Watch for this in an SP1 beta coming soon -- www.safe.com/beta 

Reply