Numbers saved as doubles instead of strings in pipeline filters

[StopkaKris]

Hi Mike,

I am posting the issue we discussed in our email. You mentioned the issue is due to the numbers being saved as doubles instead of a string. My initial email is shown below:

I am using the attached pipeline to read in surface data from a .dream3d file to compute grain boundary surface area. In the first three filters, I 1) create a data container, 2) create an attribute matrix with the necessary number of tuples to match the data set which I will import using “Import HDF5 Dataset”, and 3) I import said data set. I do this twice in the attached pipeline, first for a data set of length 664,221 and then once more for a data set of length 1,371,757. Everything worked fine when I first created this pipeline. After I saved and reopened it, the following takes place: the number “1371757” that I previously entered into “Create Attribute Matrix” gets overwritten to “1.37176e+06” (it seems to just be a rounding setting). The preflight check then causes the pipeline to immediately crash (my though is because it looks into the .dream3d file for the data set of length 1371760 (incorrectly rounded from 1371757). However, if I first disable the filters and THEN save the pipeline, I can open it without it crashing. If I then enable the filters, I will get the explicit preflight error message that the data set in the .dream3d file does not match the dimensions specified in the previous folder.

It seems that the fix for this would be to just prevent “Create Attribute Matrix” from rounding the previously saved number when a pipeline is saved and reopened, which only appears to be an issue if the number is greater than 1,000,000. This happens in both 6.5.141 and 6.6.281 on windows. I can clarify further if needed.

Before saving:

After saving:

Best regards,
Kris