Hi all,

I'm a newcomer to FME from the geospatial side of things and have inherited a somewhat large model of complex geometric operations (clips, intersects, bounding boxes etc...). On trying to run the model (with data and a workbench that has, supposedly, succeeded in 2018) I always encounter this "data shattering" warning leading to a fatal error.

Hit data-shattering case as 55 new nodes were created after phase 7, compared to 2 nodes created after phase 1

To resolve the above issue, try setting a tolerance if your transformer has a Tolerance parameter

Error running translation

I believe I've narrowed the problem down to a NeighborFinder operator, and have attached screenshots of the processes leading into it, as well as its parameters.

From this prior answer, and linked blog post, it seems that "data shattering" relates to problems of over accuracy, so to speak (?). Though, from from my (limited) understanding, there were no answers to my question there, hence my question here.

Ideally solutions that don't require too much tampering with the wider model would be best, though I recognise this may not be likely!

Other bits of info that may (or may not) be useful:

FME(R) 2019.2.2.0 (20200103 - Build 19817 - WIN64)

Windows 10

i7-7700K

64 GB RAM

Back end of the log is also attached.

Any help would be much appreciated!