Randomly selecting rows where the columns should switch values in a DataFrame

Question

Given a DataFrame like this:

import pandas as pd
import numpy as np

df = pd.DataFrame({'A': np.ones(5), 'B': np.zeros(5), 'C': np.ones(5), 'D': np.zeros(5)})

I want to be able to randomly select a number of rows where the A and B value along with their corresponding C and D values - BUT NOT column E and F are - switched so that the B column value is put in the A column and vice versa.

So it's not the whole row, but only certain columns that should switch (A, B, C , D) while E and F keep their values.

Does anybody have any ideas on how to accomplish this?

Answer 1

I think there definitely could be a more efficient way than taking copies of the Series here:

import pandas as pd
import numpy as np

df = pd.DataFrame({'A': np.ones(5), 'B': np.zeros(5), 'C': np.ones(5), 'D': np.zeros(5)})

rows_to_swap = np.random.choice(len(df), size=3, replace=False)

a_column = df['A'].copy()
b_column = df['B'].copy()

df.loc[rows_to_swap, 'A'] = b_column[rows_to_swap]
df.loc[rows_to_swap, 'B'] = a_column[rows_to_swap]