[FIXED] how to read/write pandas dataframe acrossing multi-class in a shared manner

Issue

I got this problem when I was trying to access pandas dataframe in a shared manner.

Basically, there are 3 different classes maintain different aspects of a single dataframe:

• A class creates the dataframe
• B class adds and maintain feature B into the dataframe
• C adds some anaylsis and summerization information to the dataframe

So, over all my classes looks like:

    import pandas as pd
    
    class AClass:
        def __init__(self):
            self.df_original = pd.DataFrame({'A':[1, 2, 3]})
    
        def feature(self):
            """
            add new row to the dataframe every time when it called
            """
            i = len(self.df_original) + 1
            self.df_original = self.df_original.append(pd.DataFrame({'A':[i]}, index=[i]))
    
    class BClass:
        def __init__(self, df_in):
    #         self.df_featureB = df_in.copy(deep=False) doesn't work
            self.df_featureB = df_in
    
        def feature(self):
            # process 'A' column, out come to 'B'
            self.df_featureB['B'] = self.df_featureB['A'] + 1
    
    class CClass:
        def __init__(self, df_in):
    #         self.df_featureC = df_in.copy(deep=False) does not work
            self.df_featureC = df_in
        
        def feature(self):
            # process 'A' and 'B' column, out come to 'C'
            self.df_featureC['C'] = self.df_featureC['A'] + self.df_featureC['B']
    
    # declare a b and c objects from above defined classes
    # set data frame share between the 3 objects
    a = AClass()
    b = BClass(a.df_original)
    c = CClass(a.df_original)

Here is what I want do: To repeat call bellow lines as a function

    do_step_process():
        a.feature()
        b.feature()
        c.feature()

    debug_print():
        print(a.df_original)
        print(b.df_featureB)
        print(c.df_featureC)

# call above function 3 times
    do_step_process()
    do_step_process()
    do_step_process()
    debug_print()

I expect to see all the print are same after many times calling but I get bellow after 3 times running, what I have growed from AClass doesn't apear available from BClass and CClass:

       A
    0  1
    1  2
    2  3
    4  4
    5  5
    6  6
       A  B  C
    0  1  2  3
    1  2  3  5
    2  3  4  7
       A  B  C
    0  1  2  3
    1  2  3  5
    2  3  4  7

I have tried with copy(deep=False), it doesnot work

Solution

I just figure it out...

In my case I can use loc to append new items to my a.df_original, the problem was because 'append' creates new dataframe

SO, my solution is change AClass like bellow:

    class AClass:
        def __init__(self):
            self.df_original = pd.DataFrame({'A':[1, 2, 3]})
    
        def feature(self):
            i = len(self.df_original) + 1
            self.df_original.loc[i, 'A'] = i

This works by my side, this will prevent pandas from creating new dataSet

Answered By - yunfei