Set diagonal triangle in pandas DataFrame to NaN

Question

Given the data framework below:

import pandas as pd
import numpy as np
a = np.arange(16).reshape(4, 4)
df = pd.DataFrame(data=a, columns=['a','b','c','d'])

I would like to get the following result:

df([[ NaN,  1,  2,  3],
    [ NaN,  NaN,  6,  7],
    [ NaN,  NaN,  NaN, 11],
    [ NaN,  NaN,  NaN,  NaN]])

So far I tried to use it np.tril_indicies, but it only works with returning df to the numpy array and only works for whole purposes (not np.nan):

il1 = np.tril_indices(4)
a[il1] = 0

gives:

array([[ 0,  1,  2,  3],
       [ 0,  0,  6,  7],
       [ 0,  0,  0, 11],
       [ 0,  0,  0,  0]])

... this is almost what I'm looking for, but barfs when assigning NaN:

ValueError: cannot convert float NaN to integer

and

df[il1] = 0

gives:

TypeError: unhashable type: 'numpy.ndarray'

So, if I want to fill the bottom triangle of the DataFrame with NaN, then it 1) must be a numpy array, or can I do it using pandas directly? And 2) Is there a way to fill the bottom NaN triangle, and not use numpy.fill_diagonaland increase the line-by-line offset across the entire DataFrame?

: np , np.nan. NaN, !

Answer 1

float a, type of NaN float:

import numpy as np
a = np.arange(16).reshape(4, 4).astype(float)
print (a)
[[  0.   1.   2.   3.]
 [  4.   5.   6.   7.]
 [  8.   9.  10.  11.]
 [ 12.  13.  14.  15.]]


il1 = np.tril_indices(4)
a[il1] = np.nan
print (a)
[[ nan   1.   2.   3.]
 [ nan  nan   6.   7.]
 [ nan  nan  nan  11.]
 [ nan  nan  nan  nan]]

df = pd.DataFrame(data=a, columns=['a','b','c','d'])
print (df)
    a    b    c     d
0 NaN  1.0  2.0   3.0
1 NaN  NaN  6.0   7.0
2 NaN  NaN  NaN  11.0
3 NaN  NaN  NaN   NaN

Answer 2

np.where -

m,n = df.shape
df[:] = np.where(np.arange(m)[:,None] >= np.arange(n),np.nan,df)

-

In [93]: df
Out[93]: 
    a   b   c   d
0   0   1   2   3
1   4   5   6   7
2   8   9  10  11
3  12  13  14  15

In [94]: m,n = df.shape

In [95]: df[:] = np.where(np.arange(m)[:,None] >= np.arange(n),np.nan,df)

In [96]: df
Out[96]: 
    a    b    c     d
0 NaN  1.0  2.0   3.0
1 NaN  NaN  6.0   7.0
2 NaN  NaN  NaN  11.0
3 NaN  NaN  NaN   NaN