[BUG] Investigation read conflict in F# but not in python

[Lu98Be]

Describe the bug
The F# variant of arctrl can't open up this investigation file isa.investigation.xlsx, while the python variant seems to be able to.

To Reproduce
Read the file using f#.

Code used with F#:

Code used with Python:

Expected behavior
Both versions of arctrl to be able to read the file

OS and framework information (please complete the following information):

• OS: Windows 11
• OS Version: 24H2
• .Net core SDK version: 8.0.303
• python version: 3.11.9

Additional context
The Excel file was previously edited using python (pandas and openpyxl).
This seems to somehow create conflicts with arctrl (at least the f# variant).
This problem was discovered when the arcjson pipeline was not able to read any investigation file edited using arcmanager (which uses pandas and openpyxl).
I tried fixing this by editing the file using arctrl (python) but this did not fix the issue.

[HLWeil]

Hey @Lu98Be, thanks for bringing this up.

I will take a look into the file and see if I can find the issue.

[Halima-saker]

Is there any progress on this issue?

[HLWeil]

I found the source of this issue. Unfortunately, basically each implementation of xlsx file parsers user their own dialect. We do test our polyglot excel library FsSpreadsheet against various xlsx file sources, including the og Excel, but also .Net, JS and Python (openpyxl) libraries.

In the file you sent, I encountered a new pattern in the worksheet file to store cell content:

  <c r="A1" t="inlineStr">
      <is>
          <t>ONTOLOGY SOURCE REFERENCE</t>
      </is>
  </c>

These values here are stored as Inline Strings as opposed to being references to the actual values in the Shared String Table. I would need to check how deep the issue runs in the .NET and Javascript libraries we use to parse these files. For this I will open an issue in FsSpreadsheet.

Maybe you could try to identify a setting in Pandas that forces the parser to use the Shared String Table when writing the xlsx files?

@Halima-saker @Lu98Be