Skip to main content

2022-02-18_13-29-27Hi!

Using FME Desktop 20202.0, I cannot explain to myself why GeometryValidator transformer can miss a geo-issue when using ShapeFile format while can catch it for the same dataset in FileGDB format. Attaching screenshot.

 

Any ideas?

Thanks,

Juanma,

Generally speaking, the FGDB format implements a snapping grid (tolerance and precision in the feature class definition) that might modify the geometry very slightly when written. If the feature class tolerance and/or precision is not right for the data, it can lead to issues where geometries with very tight tolerances are rendered invalid (e.g. self-intersections) by the snapping grid.

Shape files do not implement such a snapping grid, data is saved as-is in full precision.


Thanks @david_r​ , but that "hole" that you can see on the screenshot is located on both formats, so I expected the GeometryValidator could find for both datasets as well. Also MapINFO (miTAB) was used and GeometryValidator detected it.


@juanmahere​ Under the covers, MapInfo is also an interger based coordinate system (Stores coordinates natively as 32-bit integers. Yields 10 digits of precision) . Extended TAB is 64bit I believe.


Reply