Python list of list from CSV files to create grouped bar chart

I have several CSV files in this format, named n-data.csv where n is the number of threads:

elapsed,label
120,Step 01
260,Step 02
113,Step 03
100,Step 01
200,Step 02
103,Step 03

and I would like to get the average of each step per thread (or n), and plot them as grouped bar charts much like this: http://matplotlib.sourceforge.net/examples/api/barchart_demo.开发者_运维技巧html

My code works fine with one file (thanks to Yet Another list csv file in Python Question):

#!/usr/bin/env python

import csv
import sys

import pylab as p

fig = p.figure()
ax = fig.add_subplot(1,1,1)

result = dict()
av = []
idx = []
for file in sys.argv[1:]:
  for row in csv.DictReader(open(file)):
    label = row['label']
    elapsed = row['elapsed']
    if label in result:
      result[label].append(elapsed)
    else:
      result[label] = [elapsed]    
  for i in sorted (result.iterkeys()):
    s = sum(int(v) for v in result[i])
    a = s/float(len(result[i]))
    av.append(a)
    idx.append(i)

  y = av
  N = len(y)
  ind = range(N)
  ax.bar(ind, y, facecolor='#56A5EC',
          align='center',label='1 Thread') 
  ax.set_ylabel('Average Response Time')
  ax.set_title('Counts, by group',fontstyle='italic')
  ax.legend()
  ax.set_xticks(ind)
  ax.grid(color='#000000', linestyle=':', linewidth=1)
  group_labels = idx
  ax.set_xticklabels(group_labels)
  fig.autofmt_xdate()  
  p.grid(True) 
  p.show()

However, when I run this script with create_bar.py *csv, all the data gets pushed into av (average) and idx (index), and it aggreates all the data into av and idx -- which works as it should.

How do I break out av and idx so I can create group charts? I tried:

    threads = file.split('-')[0]
    for count in threads
      for row in csv.DictReader(open(file)):
      ...
        av['threads'].append(a)
        idx['threads'].append(i)

to no avail.

Perhaps have a list of lists ...

av = []
idx = []

for file in sys.argv[1:]:
    next_av = []
    next_idx = []

    ...

    for i in sorted (result.iterkeys()):
        s = sum(int(v) for v in result[i])
        a = s/float(len(result[i]))
        next_av.append(a)
        next_idx.append(i)
    av.append(next_av)
    idx.append(next_idx)

for index in range(len(av)):
    # do something with the lists av[index] and idx[index]