convert from table to matrix in python

Question

I have the following data in form of pandas data frame (df):

Each column name has starting and ending nodes separated by '_'. I.e, 'A_B' means a starting node of A and ending node of B

A_A   A_B   A_C   B_B   B_A   B_C   C_C   C_A   C_B
12    23    33    43    19    98    44    77    35

and I want to convert it into a a 2D matrix that gets output to a csv so that it looks like:

   A    B    C
A  12   23   33
B  19   43   98
C  77   35   44

To do this, assume that we have

from itertools import groupby, combinations_with_replacement
col_names = ['A','B','C']

for i in combinations_with_replacement(col_names, 2):
    get_val = df[i[0]+'_'+i[1]]

Is there a way in pandas to output this as a matrix and also print out row and column names?

Answer 1

I'm not sure if this was all done as elegantly as possible, but:

df = pd.read_clipboard()
df
Out[3]: 
   A_A  A_B  A_C  B_B  B_A  B_C  C_C  C_A  C_B
0   12   23   33   43   19   98   44   77   35

dfT = df.transpose().reset_index()

dfT
Out[8]: 
  index   0
0   A_A  12
1   A_B  23
2   A_C  33
3   B_B  43
4   B_A  19
5   B_C  98
6   C_C  44
7   C_A  77
8   C_B  35

dfT['col_name'] = dfT['index'].str.split('_').map(lambda x: x[1])
dfT['row_name'] = dfT['index'].str.split('_').map(lambda x: x[0])
dfT.pivot(index='row_name', columns='col_name', values=0)
Out[17]: 
col_name   A   B   C
row_name            
A         12  23  33
B         19  43  98
C         77  35  44

And to drop the index names:

final_mat = dfT.pivot(index='row_name', columns='col_name', values=0)
final_mat
Out[19]: 
col_name   A   B   C
row_name            
A         12  23  33
B         19  43  98
C         77  35  44

final_mat.index.name = None
final_mat.columns.name = None
final_mat
Out[22]: 
    A   B   C
A  12  23  33
B  19  43  98
C  77  35  44