BUG: Pandas doesn't release the lock when a corrupted file is fed in. pandas==1.2.1 worked fine but pandas==1.2.4 has this issue.

[SAH-UJA]

Kindly refer the following code snippet.

import pandas as pd
import os

# Creating a corrupted file
with open('a.xlsx', 'w') as f:
pass

# Reading using pd.ExcelFile
data = pd.ExcelFile('a.xlsx', engine='openpyxl')

# Deleting the file
os.remove('a.xlsx')

Here, I create a .xlsx file which is not the correct way to make xlsx file but this will express my concern. Then I use pd.ExcelFile to read the xlsx file but it gives me zipfile.BadZipFile: File is not a zip file which is fine but later when I try to delete that file I get PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'a.xlsx'. This doesn't happen in pandas==1.2.1 and lower version. The current version i.e. pandas==1.2.4 has this bug.

[twoertwein]

thank you for your report! Pandas 1.2 encourages to use a context manager with pd.ExcelFile to handle closing resources in case of errors.

Does the following work or you:

from pathlib import Path
import pandas as pd

file = Path("a.xlsx")
file.touch()

with pd.ExcelFile(file, engine='openpyxl') as _:
    pass

file.unlink()

[SAH-UJA]

@twoertwein I tried the snippet which you suggested but file.unlink() again gives the same exception that PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'a.xlsx'. According to my belief, with context manager tries to call .close() but the problem is that the object is not getting returned because of the exception so .close() could not be called and therefore I feel that lock should have been released during exception handling internally by pandas. This is just my way of looking at this problem. I would love to get your feedback.