Fix default pandas methods

[shantanuo]

Is it possible to "fix" pandas methods using pyjanitor? For e.g. I will like to validate the parameters for read_excel or read_csv. I have raised an issue but there is no progress.

pandas-dev/pandas#22189

If janitor accept the PR to override the default behavior of read methods, then it will be great.

[ericmjl]

Is it possible to "fix" pandas methods using pyjanitor? ... If janitor accept the PR to override the default behavior of read methods, then it will be great.

@shantanuo I'd be inclined to disagree with doing that. The purpose behind pyjanitor is to provide the convenience APIs that pandas did not provide. If there's a bug that's inside pandas, then that should be raised with the pandas devs, as you did already.

[Zsailer]

I agree with @ericmjl here.

pyjanitor aims to be a general purpose library that extends the Pandas API. Users should be able to add pyjanitor to their current workflow without breaking their existing code.

If pyjanitor starts patching pandas functions/methods, it could quickly cause trouble for users and prevent them from using pyjanitor.

Also, I don't think you need a sophisticated tool like pyjanitor to fix your problems. Because read_csv, read_excel are just functions, you could create a small module that enforces the signature you want:

import pandas as pd
from functools import wraps

@wraps(pd.read_excel)
def read_excel(parameters_that_you_want):
    # Only include the parameters you want in the function signature.
    return pd.read_excel(parameters_that_you_want)

Perhaps this would solve your problem until Pandas addresses the issue?