SwathGrid sampler no DE issue

[bena-nasa]

@atrayano reported that when trying @metdyn swath regridding he looked like what was the infamous no-DE issue, i.e. something got decomposed in a way that some processors had no DE's in the grid.

I was kind of able to reproduce via a hack (since we don't have any very coarse synthetic swath data this was the best I could do). I think I caught the places. Making an issue for a PR eventually.