[vtkusers] vtkExtractEdges changes point IDs?

[David Gobbi]

On Fri, Apr 9, 2010 at 6:41 PM, Burnie West <west at ieee.org> wrote:
>
> The word "usually" troubles me. In what cases might I expect the output
> points to be ordered differently from the input points for a filter that
> traverses the points?

If you are asking if there is a specific superclass for filters that
keep the same point ordering, the answer is that there is not.  But
here is a brief sampling of filters that will guarantee the same
ordering:

Filters that create/modify cells while keeping all points:
vtkDelaunay2D, vtkDelaunay3D, vtkTriangleFilter, vtkStripper, vtkReverseSense.

Filters that change point positions without changing cells:
vtkTransformPolyDataFilter, vtkShrinkPolyData, vtkWarpTo, etc.

Filters that act on point scalars without changing geometry:
vtkProbeFilter etc.

And here is the exception that proves the rule:  The
vtkPolyDataNormals filter only provides this guarantee if its
Splitting option is turned off.

So, hopefully this is about what you would expect.  Note that this
thread was started with a question about vtkExtractEdges.  Any filter
that does geometry extraction, clipping, etc. will not maintain
ordering.

   - David