[gdal-dev] Fwd: Gdal - multipolygon - geometry errors instead holes

[Tobias Wendorff]

Hi Marco,

Am 30.10.2019 um 09:46 schrieb Marco:
>
> Indeed Buffer 0 won't do any good on self intersections. Thats just not
> the way to repair them.

But many users act like this:

1. Get a warning about invalid geometry,
2. run Buffer 0,
3. Warning about invalid geometry is gone.

Again, this is not constructed, it's reality.

> Hence I wrote "Buffer 0 on INVALID geometries IMHO can lead to good
> results depending on the /nature of the invalidity/. To my experience
> /repeating points,//wrong coordinate ordering and holes touching outer
> shells in one point///are being repaired well without data loss.".

Of course, in this case it's correct. But a user needs to read and
understand the problem first. Most don't know what's going on at all,
run "Buffer 0 " and the problem seems to be gone.

> Of course self intersections can be easily overlooked. For that we have
> validity checkers.

No. That's what we've got "MakeValid", "Fix Geometry" etc. for. These
are the correct functions in every case, even if they're slower.

Sorry, I'm just talking from experience... might be different if users
know how to interpet the output of a validity checker. But please, never
recommend them to use Buffer 0 ... just recommend them to use the real ones.

Best regards,
Tobias