I’ve added two nex examples to the VTKExamples Project. They illustrate the difference between vtkTableBasedClipDataSet and vtkClipDataSet.
The former is faster and produces many fewer cells because it retains the original cells outside the clip boundary. See the difference in these images:
If you have been using vtkClipDataSet you may wish to switch to vtkTableBasedClipDataSet.
Thanks to Bane Sullivan for bringing up the original problem and for helping to track down a solution.
Also, he provided a great unstructured grid test dataset.