[vtkusers] Data got duplicated when update()?

[Berk Geveci]

	In vtkFieldDataToAttributeDataFilter, the geometry and data are
	not duplicated. The points and the data are shared. If you
	can read all the data with a single reader and use
	vtkFieldDataToAttributeDataFilter afterwards, there is
	no duplication in any way. However, if you create two
	readers and read the same data file, the geometry will
	be duplicated. If  you need to do this, one solution might
	be reading the data once, extracting the data array you need,
	registering it and then deleting the reader. This way,
	you still have the data array but the geometry is deleted.
	Next time you read the same data file, you can add this
	data array to the reader's output's field data or attributes.
	Of course, this means that you read the geometry twice.
	Let me know if you need a more concrete example.

 > I also looked at what is done in the
 > graphics/examplesTcl/MultidimensionalSolution.tcl, it seems that this
 > is the right answer for it.  But I am wondering how the memory is
 > managed here?  When vtkFiedlDataToAttributeDataFilter is used twice to
 > create velocityDataSet and pressureGradientDataset, does the
 > underlaying data (10 million points) got duplicated in the memory for
 > each of them so that they can have independent pipelines?