[Python] Pandas (`pd.DataFrame.subtract()`) is generating a number that is both recognized as...

pd.DataFrame.subtract() is having a very weird buggy behavior on my machine:

df = final_df.cumsum(axis=1)

print((df["A"] == 0).all()) # evaluates to True

print((df.subtract(df["A"], axis=0) - df)["B"].iloc[3] != 0.) # evaluates to False
print((df.subtract(df["A"], axis=0) - df)["B"].iloc[3] == 0.) # evaluates to False
print((df.subtract(df["A"], axis=0) - df)["B"].iloc[3]) # evaluates to 0.032420000000000004
# but keeps changing to random numbers every time I run the cell


Any suggestion on how to debug something like this?

Continue reading...