In Python: Create A Table To Visualize Pairwise Data

I have performed a bunch of pairwise comparisons (241 x 241 to be exact). The resulting file looks like this: A,X,10 A,Y,20 X,Y,15 I would like to convert this to a table showing

Solution 1:

Store your paired data into a dict:

result_dict = {
    (A, X): 10,
    (A, Y): 20,
    (X, Y): 15,
}

and row/column labels:

cols = sorted(set(a for a, b in result_dict.iterkeys()))
rows = sorted(set(b for a, b in result_dict.iterkeys()))

then it's how you gonna print rows...

for b in rows:
    row = list(result_dict.get((a, b), None) for a in cols)
    # print the row

Solution 2:

I have a feeling there are more efficient ways, but you could give something like this a shot. It uses the csv module to load/parse your data and then writes it back to csv as well (assuming you want the output in a file, that is - if not, this can be adjusted):

import csv
from collections import defaultdict

# Open the input file and read it all into a defaultdict
with open('table.csv', 'rb') as f:
  reader = csv.reader(f)

  # Dictionary to hold the nested values
  # This will be keyed by the letter ID, and each line of
  # the file will be written twice (so A,X,10 will be represented
  # as {'A': {'X': 10}, 'X': {'A': 10}}
  d = defaultdict(dict)
  for row in reader:
    d[row[0]][row[1]] = row[2]
    d[row[1]][row[0]] = row[2]

# Now we open the output file and write out our defaultdict
with open('t_out.csv', 'wb') as o:
  # Here our fieldnames will start with the 'empty' first header
  # and then be comprised of the keys of the dictionary (which
  # should contain all possible values for the table)
  fieldnames = [' '] + d.keys()
  writer = csv.DictWriter(o, fieldnames=fieldnames)

  # In Python 2.7, you can use writer.writeheaders()
  writer.writerow(dict((h, h) for h in fieldnames))

  # Now iterate through our dictionary, creating a row
  # dictionary that will contain the information to be written
  for k, v in d.iteritems():
    # Here we are putting the key in the 'empty' first column
    v[' '] = k
    writer.writerow(v)