Find data/ freeze selection appears to not work

Hi guys,
I appear to not be able to “freeze selection” on a find data. First, I wanted to know if I am doing something obviously wrong, if not, I will write up a bug. My user wants to do the following:

  • 5.8.0, Linux, builtin server.
  • Load can.exo. All vars on. Apply.
  • Point Data to Cell Data. Apply. We want cell data.
  • Last timestep.
    Now, we want to look at all cells that have a displacement over 10. We want to look at these cells over all time. (The actual goal is to look at high stress cells).
  • Find data.
    ** Find Cells from PointDataToCellData.
    ** DISPL(Magnitude) is >= 10
    ** Run Selection Query
    If you look at this selection, it appears to be correct.
    ** Freeze Selection
  • Extract Selection. (Copy Active Selection, or not, Doesn’t matter.) Apply.
    The selection has been changed. Further, if you go to time zero, you can see some cells were removed, but not according to DISPL!

Am I missing something, or is this a bug? Any ideas around it?

Thanks,

Alan

@cory.quammen @utkarsh.ayachit

Alan,

Freeze selection converts the query-based selection (DISPL(Magnitude) >= 10) to a cell ID-based query. Once you’ve frozen the selection, stepping through time will select the same cell IDs - the query is no longer considered because freezing the selection has converted it.

To get the behavior you expect to see, do not freeze the selection. The selection will remain a query-based selection, and you will see only cells that satisfy the query as you move through time.

  • Cory

I think I wasn’t clear. If I have 10 cells, and 5 fit some criterion, I want to extract these 5 cells, then watch how these 5 cells move and change over time. I do not want to see cells come and go, as they move in and out of the (DISPL(Magnitude) >= 10).

Ah, thanks for clarifying. I see the selection changing after “Freeze Selection” has been clicked. It looks like IDs from one block are being selected in the other block after “Freeze Selection”, despite them not passing the query. This is especially apparent when you set the criteria to DISPL(Magnitude) >= 19.8, timestep 43, where only two blocks are selected. So yeah, there is a bug here.

Thanks Cory, written up here: https://gitlab.kitware.com/paraview/paraview/issues/19711.