Skip to main content
Question

Centerline Replacer (Skeleton) fails on simple polygon

  • November 2, 2015
  • 9 replies
  • 28 views

 

Using quite a simple polygon in Geometry Replacer, this transformer fails to convert to skeleton centerline. In this question example we isolated only 1 polygon that makes the centerline replacer to crash as an example; it’s something we have observed over time with different datasets, etc..

 

 

The question is - why does centerline replacer using skeleton option fails even if this geometry from shapefile seems rather simple?

 

 

Tested FME 2014 SP and FME 2015 SP1. The centerline replacer failure using skeleton option was there in previous versions also.

 

 

In the example we are sending, we identified 1 coordinate that seems to be offending the polygon and causing this crash. We isolate this coordinate, and by moving it very little, the centerline skeleton already works, still, it was not a coordinate that was near or on top of another etc.

 

 

Attached you will find the polygon (1 polygon only) and the WS with some tests.

 

 

Assumptions:

 

 

  • -We do know that using generalizer helps to prevent crashes using Centerline Replacer, but we don’t want to generalize as we want to maintain this exact polygon coordinates to maintain topology to neighboring polys; Being already a simple poly we don’t want to generalize further;
  • -Tested for duplicate coordinates and there are no duplicate coordinates on the polygon boundary.

 

Restating the question, why does centerline replacer using skeleton option fails even if this geometry from shapefile seems rather simple? By trying to understand why this 1 polygon fails, might help to modify the skeleton replacer and prevent more features to crash.

 

 

Thanks in advance,

 

Ricardo

 

 

PS: Topics in not directly related - CenterLineReplacer not available as a Topic

 

This post is closed to further activity.
It may be an old question, an answered question, an implemented idea, or a notification-only post.
Please check post dates before relying on any information in a question or answer.
For follow-up or related questions, please post a new question or idea.
If there is a genuine update to be made, please contact us and request that the post is reopened.

9 replies

fmelizard
Safer
Forum|alt.badge.img+20
  • Safer
  • 3719 replies
  • November 2, 2015
Did you try the GeometryValidator before the Centerlinereplacer @ricardometrogos ?

Did you try the GeometryValidator before the Centerlinereplacer @ricardometrogos ?

Did before pretty sure, and did it just now to make sure; All geometry is valid checking for all types of issues. Thanks for your answer.


fmelizard
Safer
Forum|alt.badge.img+20
  • Safer
  • 3719 replies
  • November 3, 2015
Hi @ricardometrogos I've made this question into case C109030 and filed PR65409 about it. Besides the GeometryExtractor and GeometryReplacer, I was also able to work around the problem with a Reprojector (to a local coordinate system) or with a CoordinateRounder - even to very high precision. I'll let you konw when this has been fixed.

 


Hi @ricardometrogos I've made this question into case C109030 and filed PR65409 about it. Besides the GeometryExtractor and GeometryReplacer, I was also able to work around the problem with a Reprojector (to a local coordinate system) or with a CoordinateRounder - even to very high precision. I'll let you konw when this has been fixed.

 

Hi. Thanks for follow up. Will be interesting to minimize the crash of skeleton using CenterlineReplacer since here and there it does stop. We identified also a few workarounds, but sometimes it's just puzzling why it stops even if the geometry is simple as the one I attached. In certain processes I do refrain from generalizing or rounding to avoid missing what we had before as a correct topology with some other features. Rounding, even if minimal could cause random snapping to a "grid" and these features vs neighboring features, could end up causing small slivers. Thansk in advance do keep us posted please.


Hi @ricardometrogos I've made this question into case C109030 and filed PR65409 about it. Besides the GeometryExtractor and GeometryReplacer, I was also able to work around the problem with a Reprojector (to a local coordinate system) or with a CoordinateRounder - even to very high precision. I'll let you konw when this has been fixed.

 

Hi Dan, Do we have any update on thie C109030 (PR65409)? about the Centerline Replacer (Skeleton mode) failures.. Thanks in advance for follow up.


fmelizard
Safer
Forum|alt.badge.img+20
  • Safer
  • 3719 replies
  • July 20, 2016

No update yet, sorry.


  • 1 reply
  • January 24, 2017

I'm running up on this issue too. Interestingly, FME2013 worked fine but I no longer have access to the older version. Also, the same features cause PostGIS/SFCGAL to bail as well: https://github.com/Oslandia/SFCGAL/issues/133


paalped
Contributor
Forum|alt.badge.img+5
  • Contributor
  • 130 replies
  • March 16, 2018

 

Same issue with the new release of FME 2018. Hope they can fix this bug.

fmelizard
Safer
Forum|alt.badge.img+20
  • Safer
  • 3719 replies
  • March 16, 2018

 

Same issue with the new release of FME 2018. Hope they can fix this bug.
Hi @paalped I've just tested the CenterLineReplacer and the original problem now works for me with FME 2018.0. Please send us your data and workspace via www.safe.com/support Thanks!