Hi Jarad,
I've received your file; thanks!
I haven't had a chance to run tests yet, and I don't see anything at first glance that would definitely lead to this issue, but I did want to provide a couple thoughts and suggestions.
First, I noticed that while you have piped the name and type physical object values in the CSV, you didn't pipe the location - meaning that half of these entries are ending up without a location defined. AtoM has been known to behave strangely when some of the physical storage data is missing, so there is a small chance that this relates to your issue.
I also wanted to point out that, due to the current limitations of the AtoM physical storage module (in which we have no hierarchical organization), many of your file level containers will end up in the same big bucket. What I mean is, on the archival description you'll have 2 locations (a box and a file) linked, but because the container names are not individually unique, when you click on the location for the file, you will find many descriptions listed in from other file containers that share the same name.
For example, one row has OP01|6 for the Box|File values in the physicalObjectName column. On the description, with the data you have now, you'll see something like this:
However, in AtoM's physical storage module, there's no actual relationship between that box, and that folder (and unfortunately, no way to make one at this time) - they are just 2 different containers.
This means that as the import progresses, when AtoM finds the next value that has 6 as the File value, it will link that folder to the existing location - so when you browse the physical storage module, you'll end up seeing file 6 from not just OP01, but also P01, P03, P04, etc.
If that doesn't bother you (for example, if you are only going to use the information on the description view page, and don't intend to use the physical storage module directly), you can ignore this. But I have shared a few previous thoughts on strategies to work around this:
Finally: did you manually add File to the Physical Object Type taxonomy, or did you just use the import to create the new type on the fly? The default term is "Folder." Technically, it should be fine adding a new term to the taxonomy by importing it, but it's one more place where there could be buggy behavior, so just trying to figure out possible vectors to explore :)
I will try to modify your CSV a bit so I can import it locally soon!
Regards,