Pandas replace is executing over two separate dataframes

I’m trying to make a temporary dataframe where NaN values are replaced by zeros without affecting the original dataframe. However, I noticed that when I replace the NaN’s of df_2 with 0s the corresponding column in df_1 is also changed. Have I done something wrong here when creating df_2?

Code

d = {'A':[np.nan, np.nan],'B':[1,2]}
df_1 = pd.DataFrame(data=d)
df_2 = df_1
print(df_1)
print(df_2)
df_2['A'] = df_2['A'].replace(np.nan,0)
print(df_1)
print(df_2)

Outputs

    A  B
0 NaN  1
1 NaN  2
    A  B
0 NaN  1
1 NaN  2

     A  B
0  0.0  1
1  0.0  2
     A  B
0  0.0  1
1  0.0  2

>Solution :

Use deep copy

df_2 = df_1.copy()

See more in pandas.DataFrame.copy