SherbendGeneralizer performance issues

Hi,

 

 

The SherbendGeneralizer seems to consume long time when the "Spatial Constraints" parameter is specified. Especially, when there are duplicate lines and also the parameter setting includes "Line-Line Intersection", it will take very long time and generate unexpected results. If input features include administrative boundaries,  there could be duplicate lines. If so, try transforming the boundaries to non-duplicate lines using the TologyBuilder beforehand.

 

 

Takashi

Hi Takashi,

 

 

Thank you very much for your reply and sorry for late responce as I want to use Topology Builder instead of Intersector.

 

 

So I just want to inform you that even after I have used Topologybuilder there is no change. Still running running and running.

 

 

I am really impressed with the SherbendGeneralizer because it is exactly what I need to generalize the boundaries which are really small, just consider the postal code boundaries in London. They are really small polygons and with other generaliser I am getting really bad results but with SherbendGeneralizer I am really satisfied.

 

 

Other than processing time what I am missing from this transformer(SherbendGeneralizer) is there is no message for the user and it is running without giving user any output message like 10% finished or any other messages which can give user little hint about the process stage.

 

 

Any thoughts…..

Hi,

 

 

In my testing, the SherbendGenerlizer doesn't log any messages while processing even if it finished successfully. I think it should be improved.

 

  Anyway, another check point might be whether there are self-intersections in the input lines. Did you check the data using GeometyValidator? If the SherbendGeneralizer still doesn't work for clean data, it's difficult to find out the reason for the problem. Maybe it will be a Safe support matter ...

 

 

Takashi

No I am not performing any GeometyValidator as source data is my final product and what I am trying to do is produce the generalized version from the final product. But I will check the geometry before starting this transform and let you know the results.

 

 

Rest for logging messages I will pass this request to FME support.