Unique elements inside lists of list - list

Question

If I have a nested list like:
l = [['AB','BCD','TGH'], ['UTY','AB','WEQ'],['XZY','LIY']]
In this example, 'AB' is common to the first two nested lists. How can I remove 'AB' in both lists while keeping the other elements as is? In general how can I remove a element from every nested list that occurs in two or more nested lists so that each nested list is unique?
l = [['BCD','TGH'],['UTY','WEQ'],['XZY','LIY']]
Is it possible to do this with a for loop?
Thanks

Answer 1

from collections import Counter
from itertools import chain
counts = Counter(chain(*ls)) # find counts
result = [[e for e in l if counts[e] == 1] for l in ls] # take uniqs

Answer 2

One option is to do something like this:
from collections import Counter
counts = Counter([b for a in l for b in a])
for a in l:
for b in a:
if counts[b] > 1:
a.remove(b)
Edit: If you want to avoid the (awfully useful standard library) collections module (cf. the comment), you could replace counts above by the following custom counter:
counts = {}
for a in l:
for b in a:
if b in counts:
counts[b] += 1
else:
counts[b] = 1

Answer 3

A somewhat short solution without imports would be to create a reduced version of the original list first, then iterate through the original list and remove elements with counts greater than 1:
lst = lst = [['AB','BCD','TGH'], ['UTY','AB','WEQ'],['XZY','LIY']]
reduced_lst = [y for x in lst for y in x]
output_lst = []
for chunk in lst:
chunk_copy = chunk[:]
for elm in chunk:
if reduced_lst.count(elm)>1:
chunk_copy.remove(elm)
output_lst.append(chunk_copy)
print(output_lst)
Should print:
[['BCD', 'TGH'], ['UTY', 'WEQ'], ['XZY', 'LIY']]
I hope this proves useful.