Error in Chapter 4 to 6/ 2 Process X_Y Learning Data fixNansInX function about build_your_own_ai_investor_2021

[rhinoV1]

OK, we managed to find what the problem was. I am running a newer version of python and pandas and one of them (probably pandas) has a problem with this part of the code:

def fixNansInX(x):
    '''
    Takes in x DataFrame, edits it so that important keys
    are 0 instead of NaN.
    '''
    keyCheckNullList = ["Short Term Debt" ,\
            "Long Term Debt" ,\
            "Interest Expense, Net",\
            "Income Tax (Expense) Benefit, Net",\
            "Cash, Cash Equivalents & Short Term Investments",\
            "Property, Plant & Equipment, Net",\
            "Revenue",\
            "Gross Profit",\
            "Total Current Liabilities",\
            "Property, Plant & Equipment, Net"]
    x[keyCheckNullList]=x[keyCheckNullList].fillna(0)

The problem is that "Property, Plant & Equipment, Net" appears twice in this list which causes the fixNansInX(x) function to crash. The solution is simple, just delete the last item in the keyCheckNullList and that is it.

And now everything works for me.

[Damonlee92]

Can you try running in this order?

fixNansInX(x_)
addColsToX(x_)
X=getXRatios(x_)
fixXRatios(X)

[rhinoV1]

Tried it, didn't help, same error. It seems as if the pandas library changed how fillna(0) works or something, I can't figure it out.

I extracted the code from the function fixNansInX(x) and tried running it step by step. Applying boolean masks works but when I need to replace the data it doesn't work and the error description doesn't make sense.

[Damonlee92]

Strange, it works for me. Email me on the protonmail address I'll give you a sample of the data that worked.

[Damonlee92]

ai_investor(at)protonmail.com