Skip to main content
Solved

TIN generator behaviour

  • November 3, 2025
  • 6 replies
  • 93 views
cwisse
Contributor
Forum|alt.badge.img+4

I'm having a set of polygons and for every polygon a raster with height values. I'm combining them in the TIN generator to create TIN's for every polygon. Now I see a big difference in behaviour of the TIN generator, depending on the amount of polygons I'm sending to it. For example: when I use one single polygon as input, as a result I get a TIN with 234 triangles.

If I use 459 polygons as input and proces them individually by using Group processing, the result of the same polygon is a TIN with 3602 triangles.

 

Can anyone explain why this is happening and how to prevent it?

Best answer by crystalwang

Hi ​@cwisse,

Thank you for the additional information! I was able to reproduce the same behavior in FME 2024.2. When group processing is enabled, the TINGenerator outputs 431 triangles, while processing a single input produces 221 triangles.

While I’m not entirely certain of the underlying cause, the good news is that this issue appears to be resolved in FME 2025.1 and later. To correct this issue, I would recommend upgrading to FME 2025.1 or newer. In FME Form 2025.1 b25606 for example, the number of triangles for the two features is 387 and 141 respectively, regardless of group processing or single input.

We apologize for the inconvenience this issue has caused. If you continue to experience unexpected results with the TINGenerator in FME 2025.1+, please don’t hesitate to contact our Support team. We’ll be happy to investigate further there if needed.

Thank you for your understanding!

danilo_fme

6 replies

crystalwang
Safer
Forum|alt.badge.img+14
  • crystalwangSafer
  • Safer
  • November 7, 2025

Hi ​@cwisse,

The difference in triangle counts between individual and group processing can happen for a few reasons. In your case, a couple of possibilities come to mind:

  • Rasters not clipped to polygons – If the rasters aren’t clipped to the polygons before reaching the TINGenerator, it may include extra points from overlapping raster areas, resulting in a denser triangulation.

  • Grouping setup – The Group By field might not be isolating a single polygon and raster pair. You can use a StatisticsCalculator before the TINGenerator to confirm how features are grouped.

If you’re still seeing inconsistent results, could you share a small sample workspace or dataset? That would help us better understand your workflow and test what’s happening. Hope this helps!

cwisse
Contributor
Forum|alt.badge.img+4
  • cwisseContributor
  • Author
  • Contributor
  • November 10, 2025

Hi ​@crystalwang,

The rasters are clipped based on the bounds of the cell and cells on the clipping boundary are output as outside. So I'm certain that only the cells that are completly inside are used. And I have also tested if there are duplicate ID's, but there aren't any.

 

Attached you will find (a part of) the workspace with the input contours and rasters of two locations. One of them is the same as descriped above. When I only take this one as input, I get the result of 221 triangles. When I take both as input with group processing, the result is 431 triangles. The other gets the same result as single input and with group processing.

1 Attachments
cwisse
Contributor
Forum|alt.badge.img+4
  • cwisseContributor
  • Author
  • Contributor
  • November 10, 2025

Meanwhile, I found out that the first location that arrives is staying the same and the second one has a different output.

crystalwang
Safer
Forum|alt.badge.img+14
  • crystalwangSafer
  • Safer
  • Best Answer
  • November 10, 2025

Hi ​@cwisse,

Thank you for the additional information! I was able to reproduce the same behavior in FME 2024.2. When group processing is enabled, the TINGenerator outputs 431 triangles, while processing a single input produces 221 triangles.

While I’m not entirely certain of the underlying cause, the good news is that this issue appears to be resolved in FME 2025.1 and later. To correct this issue, I would recommend upgrading to FME 2025.1 or newer. In FME Form 2025.1 b25606 for example, the number of triangles for the two features is 387 and 141 respectively, regardless of group processing or single input.

We apologize for the inconvenience this issue has caused. If you continue to experience unexpected results with the TINGenerator in FME 2025.1+, please don’t hesitate to contact our Support team. We’ll be happy to investigate further there if needed.

Thank you for your understanding!

cwisse
Contributor
Forum|alt.badge.img+4
  • cwisseContributor
  • Author
  • Contributor
  • November 11, 2025

Hi ​@crystalwang,

Too bad that there is no solution for 2024.2. We just got this version a month ago in our organization, so I think the upgrade will be at least next year.

I made a workaround now by creating a custom transformer with multiple TINGenerators in it. Within the custom transformer, every location is sent seperately to a TINGenerator. With use of parallel processing I can start 16 processes at the same time. This way I'm getting the same results as if I process them one by one, but the lead time is much shorter.

Thank you for you help and I'm looking forward to the new version.

crystalwang
Safer
Forum|alt.badge.img+14
  • crystalwangSafer
  • Safer
  • November 12, 2025

Hi ​@cwisse,

I’m glad you were able to find a temporary solution for 2024.2. Thank you for sharing your workaround on the Community as well, it’ll be helpful for anyone in the same situation!

Community Stats

32,127
Posts
121,960
Replies
39,777
Members

Latest FME

Terms & ConditionsCookie settingsAccessibility statement
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.