We have a 2 dimensional list (for this example we have it populated with unique 6 nodes and 3 masks)

myList = [[node1, mask1], [node2, mask1], [node3, mask1], [node4, mask2], [node5, mask2], [node6, mask3]]

Now i need to test each other somehow and generate a new list to put each node that is connected to a mask in a separate [], so i can easily access it later, but also i need to filter nodes like "node6" because "node6" is only connected to one mask (in our case with "mask3" only)

Basically i want my new list to look like this:

newList = [[node1, node2, node3], [node3, node4]] 

This has been giving me headache for couple of hours now.. THANK YOU IN ADVANCE!

note: it would be nice to see also what is the most efficient way to do this

edit1: what i tried:

myList =[[node1, masks1][node2, mask1] etc..] #this is earlier dynamically populated with nodes/masks
newList= []
for i in range(len(myList)):for j in range(len(myList[i])):try:if myList[i][0] in newList:pass elif myList[i][1] == myList[j][1] and len(myList) > 1:newList.append([db[i][0]])breakexcept IndexError:#print 'passed error'pass

I know this dose not makes much sense for what i asked.. my previous attempts are not saved - in this example i tried to populate each node in a newlist that is connected to the same mask twice or more time.. but this is not working as expected.

itertools.groupby() provides an efficient way to gather items:

from itertools import groupbymy_list = [['node1', 'mask1'],['node2', 'mask1'],['node3', 'mask1'],['node4', 'mask2'],['node5', 'mask2'],['node6', 'mask3']]masks_to_keep = ('mask1', 'mask2')# create a dict keyed by mask with (node, mask) pairs
as_dict = {x[0]: list(x[1]) for x in groupby(my_list, lambda x: x[1])}# create a list, in masks_to_keep order, of lists of nodes per mask
nodes = [[x[0] for x in as_dict[mask]] for mask in masks_to_keep]