+4Hi, I tried it with a mod counter, Inputs seem to work but no output at all comes out of the custom transformer. I am passing all attributes (not many anyway) into the custom transformer. It makes no difference I am trying to add a Z value to each point (frorm vertices). With the TIN I get a much cleaner result than with a raster, but it slows down the entire process x4.
Thanks in advance for any suggestion.
Best answer by redgeographics
Are you sure the _modCnt0 values of both your inputs actually match eachother? What happens with parallel processing is that the features get spread out over multiple cores, based on that grouping attribute, and they can’t interact with eachother. So if your tin points with modCnt 1 are not actually geographically overlapping your drape features with modCnt 1 you’re not getting results.
I would also recommend geographically dividing this, rather than by modulo counting, to make sure you’re not dealing with several almost-but-not-quite-exactly the same surfaces
+50How do you have your custom transformer set up? Can you share some screenshots?
+39And also make sure you’re not using Feature Caching, this will disable any parallel processing.
+4Hi, It has a simple setup. See screen shots. The mod count is already being passed on by the Counter. Within the custom transformer, all attributes are enabled to be passed on. It automatically enables them in the TIN input even though the TIN itself has no attribute.
Yes, no feature cashing, still not getting any output.
Thanks for looking at this!
+50Are you sure the _modCnt0 values of both your inputs actually match eachother? What happens with parallel processing is that the features get spread out over multiple cores, based on that grouping attribute, and they can’t interact with eachother. So if your tin points with modCnt 1 are not actually geographically overlapping your drape features with modCnt 1 you’re not getting results.
I would also recommend geographically dividing this, rather than by modulo counting, to make sure you’re not dealing with several almost-but-not-quite-exactly the same surfaces
+4Hi RedG,
No, not sure at all, I assumed as seen on the screenshot. All points overlap the TIN surface which is closed so I figure they will all catch a value in the end. Since you asked, I linked a new mod count into both inputs so the SurfacDraper. Now it runs in parallel and I get output points (z) YEEEY! 😁
I discarded grouping by region since the data is not orderd by these and they vary very very much in size, and I have the custom xformer setup to wait for all features, but correct me if I am wrong on this, as well. Much appreciated.
+50Glad you got it to work. In this case I think it is correct to wait until all features have been received
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
