CODEWITHSUNDEEP
pythonpandas
Amin Shn
Amin Shn

Reputation: 598

The copy() method in Python does not work properly

I have a pandas dataframe that I would like to make a duplicate of and do some operations on the duplicated version without affecting the original one. I use ".copy()" method but for some reason it doesn't work! Here is my code:

import pandas as pd
import numpy as np

x = np.array([1,2])
df = pd.DataFrame({'A': [x, x, x], 'B': [4, 5, 6]})

duplicate = df.copy()
duplicate['A'].values[0][[0,1]] = 0

print(duplicate)
print(df)

        A  B
0  [0, 0]  4
1  [0, 0]  5
2  [0, 0]  6
        A  B
0  [0, 0]  4
1  [0, 0]  5
2  [0, 0]  6

As you can see "df" (the original dataset) gets affected as well. Does anyone know why, and how this should be done correctly?

Upvotes: 0

Views: 490

Answers (1)

Yilun Zhang
Yilun Zhang

Reputation: 9018

The problem is actually in the list value rather than the df itself. When you are copying the dataframe, even if it's by default a deep copy, it's not doing deepcopy on the value itself, so if the value is a list, the reference is copied over, you can tell this by the fact that even though you only tried to modify the first row, but all values of A in your duplicate are modified.

The proper way is probably:

import pandas as pd
import numpy as np
from copy import deepcopy # <- **

x = np.array([1,2])
df = pd.DataFrame({'A': [x, x, x], 'B': [4, 5, 6]})

duplicate = df.copy()
duplicate['A'] = duplicate["A"].apply(deepcopy)  # <- **

duplicate['A'].values[0][[0,1]] = 0

print(duplicate)
print(df)

        A  B
0  [0, 0]  4
1  [1, 2]  5
2  [1, 2]  6

        A  B
0  [1, 2]  4
1  [1, 2]  5
2  [1, 2]  6

Upvotes: 3

Related Questions