OK. I downloaded and looked at your data. I think the issue is a precision problem. The problem is that picking points on a 2D surface is finicky because you have to place the probe point exactly on the surface. Given that floating point numbers are themselves an imprecise representing of data, sometimes numbers don’t register as equal when they should.
Floating point numbers at 0 seem to compare better. If I edit your file to place the data at the x=0 plane, the probe filter seems to work. Here is your data with a minor modification to the file. (The change is on the 7th line of the file.)
2d-fixed.vtk (625.3 KB)