Transposing selected MultiIndex levels in Pandas DataFrame

Question

I have a MultiIndexed DataFrame:

import pandas as pd
import numpy as np

l0, l1 = ['A', 'B'],['a', 'b']
c0 = ['c1', 'c2', 'c3']
data = np.arange(12).reshape(4,3)
df = pd.DataFrame(data=data, 
                  index=pd.MultiIndex.from_product([l0,l1]),
                  columns=c0)

>>>
     c1  c2  c3
A a   0   1   2
  b   3   4   5
B a   6   7   8
  b   9  10  11

I want to transpose a level of the MultiIndex and of the columns so that I result in:

df2 = pd.DataFrame(index=pd.MultiIndex.from_product([l0, c0]),
                   columns=l1)

>>>
    a    b
A c1  NaN  NaN
  c2  NaN  NaN
  c3  NaN  NaN
B c1  NaN  NaN
  c2  NaN  NaN
  c3  NaN  NaN

And obviously I want to populate the right values. My solution is currently to use map with an iterator but it feels like Pandas would have some native way of doing this. Am I right, is there a better (faster) way?

from itertools import product
def f(df, df2, idx_1, col_0):
    df2.loc[(slice(None), col_0), idx_1] = \
        df.loc[(slice(None), idx_1), col_0].values
m = map(lambda k: f(df, df2, k[0], k[1]), product(l1, c0))
list(m) # <- to execute

>>> df2
>>>
      a   b
A c1  0   3
  c2  1   4
  c3  2   5
B c1  6   9
  c2  7  10
  c3  8  11

Answer 1

First stack the columns and then unstack the level that you want to become new columns:

df.stack().unstack(level=1)
Out: 
      a   b
A c1  0   3
  c2  1   4
  c3  2   5
B c1  6   9
  c2  7  10
  c3  8  11