plot selection over time crash

Hello. Using results from openfoam. I define a scalar boundary for a species concentration. Calculate the volume, integrate variables and then plotselectionovertime. Over the time, that volume (defined by the scalar boundary) may reach zero (depending upon the assigned concentration). If this volume at any time step is zero, paraview crashes/exits. To prevent this occuring i have to eliminate the data files for all time steps when the volume is zero - then paraview doesn’t crash/exit. This is a rather laborious workaround, especially when the subsequent calculations use a different scalar (concentration) boundary and i have to restore/eliminate other time step data files accordingly. Is there an easier way to avoid this?

Welcome to the ParaView community, Daniel!

Could you please post a dataset that shows this behavior?

Thanks, Walter. Sure. How would i do that? (Please excuse my ignorance.) The openfoam files, or loaded into paraview and saved?

How big is the smallest dataset you can create that shows the issue? This would be the files that ParaView reads in. (I don’t know openfoam at all…)

About 120 MB

@mwestphal any advice?

sharing the OpenFOAM results is fine and can be opened in ParaView.

the question is how to get 120 MBytes where Kitware can get to it. That’s too big for Discourse, isn’t it?

We can use “wetransfer”

I should be able to put a link pn here and you can download it, or send the link to an email address

Yes, any file sharing website is fine.

You should be able to download it from here. There’s also a pdf with some steps.

Just in case, I have paraview version 5.6.0 64 bit

Daniel, could you try ParaView 5.9.1? Does that work correctly?

I installed 5.9.1 and repeated the exercise as described and unfortunately there was the same outcome (crashed/exited).

I’am unable to reproduce.

Please explain, step by step, how to reproduce your issue.

Sorry, Mathieu, for my vague instructions. (A lifetime habit, I’m afraid!)
Hopefully the following is sufficient:

Open the file xyz.foam in paraview
Press “Apply”
Add “clip” filter
Select “scalar” from the drop-down
From “scalars” choose “C3H8”
Unselect “invert”
Input a value of “0.024”
Press “apply”
Add “calculator” filter
Attribute type to “cell data”
In the calculator input (line below “result array name”) insert “1”
Press “apply”
Add “integrate variables” filter
Press “apply”
Select/highlight the values row in the (newly appeared) spreadsheet view box
Add “plot selection over time” filter
Press “apply”
Cross fingers…

If you repeat it with a much lower concentration (e.g., input value of say, 0.0001), so that there is always a volume for the duration of the time steps, then the crash doesn’t occur.

I reproduce. This is caused because CopyTuple is called on an array where CopyAllocate has not been called. This would require further investigation to understand why, probably the usecase where the “volume is zero” as @Danno stated.

An issue should be opened on

I tried replicating with g1s1. No crash.

So I should open an issue on ? And copy and paste all the text from above?

What is g1s1?

g1s1 is a dataset that Kitware has.