Skip to main content
Question

AreaOnAreaOverlayer collapses the geometry


juanmahere
Supporter
Forum|alt.badge.img+12

Hi guys,

I've experienced something that I couldn't expected from the AreaOnAreaOverlayer transformer.

The geometry used are these two polygon layers:

Block layer and Admin layer:

Trying to keep only with the geometry inside "Admin" layer, just after applying the AOAO transformer, I get this output:

As you see, the entire feature is deleted.

I've checked that the geometry ingested is not corrupted (GeometryValidator), or have holes, or any kind of geometry issues before entry to the transformer. I tried to "deaggregate" both layers, but same result.

I'm using FME Desktop 2016.0.1.0 - Build 16174 WIN64. I've also tried with FME versions: 2015 and 2014. Same results.

Shapefiles of dataset are also uploaded in case you want to test.

Thanks for your time,

Juanma,

9 replies

ebygomm
Influencer
Forum|alt.badge.img+39
  • Influencer
  • July 20, 2016

I have seen similar issues previously, using a coordinate rounder and rounding to 9 dp often solves these sorts of issues.

The clipper also fails to handle this geometry correctly without rounding. The clipper failing to handle particular geometries correctly was raised as PR58870 some time ago now.


Forum|alt.badge.img
  • July 20, 2016

Even if your large area is valid, it seems that FME have an issue working with it. I chopped it and rebuild the area before sending it to the AreaOnAreaOverlayer and the missing area is there.

I suggest that you file a support case here:

https://support.safe.com/knowledgeSubmitCase

Regards,

Larry


juanmahere
Supporter
Forum|alt.badge.img+12
  • Author
  • Supporter
  • July 20, 2016
ebygomm wrote:

I have seen similar issues previously, using a coordinate rounder and rounding to 9 dp often solves these sorts of issues.

The clipper also fails to handle this geometry correctly without rounding. The clipper failing to handle particular geometries correctly was raised as PR58870 some time ago now.

Thanks egomm, but in this case, rounding coordinates could lead into issues or unstable behavior for future steps. The only possibility that I've managed is to track lost features. Thanks also for point me to the clipper handle issue. I'll take into account.


juanmahere
Supporter
Forum|alt.badge.img+12
  • Author
  • Supporter
  • July 20, 2016
larry wrote:

Even if your large area is valid, it seems that FME have an issue working with it. I chopped it and rebuild the area before sending it to the AreaOnAreaOverlayer and the missing area is there.

I suggest that you file a support case here:

https://support.safe.com/knowledgeSubmitCase

Regards,

Larry

Many thanks Larry, sounds a nice workaround to the issue. I'm starting testing. I'll share with support, and let's see if they could handle this kind of issues.


ebygomm
Influencer
Forum|alt.badge.img+39
  • Influencer
  • July 20, 2016
juanmahere wrote:

Many thanks Larry, sounds a nice workaround to the issue. I'm starting testing. I'll share with support, and let's see if they could handle this kind of issues.

Chopping and rebuilding results in the same microscopic changes as rounding to 9dp. The geometry is sublety altered. This can be tested using the matcher, the rebuilt geometry will no longer match the original.


Forum|alt.badge.img
  • July 20, 2016
ebygomm wrote:

Chopping and rebuilding results in the same microscopic changes as rounding to 9dp. The geometry is sublety altered. This can be tested using the matcher, the rebuilt geometry will no longer match the original.

You need to deaggregate before using the Matcher (it seems that both agregates (original an rebuilded) differ even if all parts match).


juanmahere
Supporter
Forum|alt.badge.img+12
  • Author
  • Supporter
  • July 20, 2016
ebygomm wrote:

Chopping and rebuilding results in the same microscopic changes as rounding to 9dp. The geometry is sublety altered. This can be tested using the matcher, the rebuilt geometry will no longer match the original.

Furthermore, I've tested (I mean using the Chop+Rebuild) with the entire dataset (the one I uploaded was just the interested area) and still falls that feature. You're right egomm, for the same reason that I'm not fan of rounding, the same happens with chop+rebuild polygons.


gio
Contributor
Forum|alt.badge.img+15
  • Contributor
  • July 20, 2016

@juanmahere

Seems a capacity problem.

If you ddo following, then there is no issue.

Clipp the lareger area first with the boundingbox of the smaller one.

Then perform the AOA.

All works well.


gio
Contributor
Forum|alt.badge.img+15
  • Contributor
  • July 20, 2016

Amind area indeed does not match fully....


Cookie policy

We use cookies to enhance and personalize your experience. If you accept you agree to our full cookie policy. Learn more about our cookies.

 
Cookie settings