Removing excel sheet with openpyxl raise error when I open the file

Question

I tested the openpyxl .remove() function and it's working on multiple empty file.

Problem: I have a more complex Excel file with multiple sheet that I need to remove. If I remove one or two it works, when I try to remove three or more, Excel raise an error when I open the file.

Sorry, we have troubles getting info in file bla bla.....

logs talking about pictures troubles

logs about error105960_01.xml ?

The strange thing is that it's talking about pictures trouble but I don't have this error if I don't remove 3 or more sheet. And I don't try to remove sheet with images !

Even more strange, It's always about the number, every file can be deleted without trouble but if I remove 3 or more, Excel yell at me.

The thing is that, it's ok when Excel "repair" the "error" but sometimes, excel reinitialize the format of the sheets (size of cell, bold and length of the characters, etc...) and everything fail :(

bad visual that I want to avoid

If someone have an idea, i'm running out of creativity !

For the code, I only use basic functions (simplify here but it would be long to present more...).

INPUT_EXCEL_PATH = "my_excel.xlsx"
OUTPUT_EXCEL_PATH = "new_excel.xlsx"
wb = openpyxl.load_workbook(INPUT_EXCEL_PATH)
ws = wb["sheet1"]
wb.remove(ws)
ws = wb["sheet2"]
wb.remove(ws)
ws = wb["sheet3"]
wb.remove(ws)

wb.save(OUTPUT_EXCEL_PATH)

Answer 1

In my case it was some left over empty CalculationChainPart. I used DocxToSource to investigate the corrupted file. Excel will attempt to fix the file on load. Save this file and compare it's structure to the original file. To delete descendant parts you can use the DeletePart() method.

  using (SpreadsheetDocument doc = SpreadsheetDocument .Open(document, true)) {
     MainDocumentPart mainPart = doc.MainDocumentPart;
     if (mainPart.DocumentSettingsPart != null) {
        mainPart.DeletePart(mainPart.DocumentSettingsPart);
     }
  }

CalculationChainPart can be also removed anytime.

While calculation chain information can be loaded by a spreadsheet application, it is not required. A calculation chain can be constructed in memory at load-time (source)