Subtuples of a tuple

Question

How do I generate

[(0,), (1,), (2,), (0,1), (0,2), (1,2), (0,1,2)]

programmatically (that is, without writing everything out by hand)? That is, a list of all nonempty subtuples of the tuple (0,1,2).

(Note this is not asking for subsets but subtuples.)

It *is* asking for subsets, specifically non-empty ones that are ordered as a subsequence. — Shashank, May 01 '15 at 19:17

score 5 · Accepted Answer · edited May 01 '15 at 18:56

5

>>> from itertools import combinations
>>> t = (0, 1, 2)
>>> print [subset for r in range(1,4) for subset in itertools.combinations(t,r)]
[(0,), (1,), (2,), (0, 1), (0, 2), (1, 2), (0, 1, 2)]

Itertools is a powerful resource. You should check out the documentation

edited May 01 '15 at 18:56

chepner

497,756
71
530
681

answered May 01 '15 at 18:48

David Greydanus

2,551
1
23
42

score 4 · Answer 2 · answered May 01 '15 at 18:49

4

You can use the powerset() recipe and remove the empty set:

from itertools import chain, combinations

def powerset(iterable):
    "powerset([1,2,3]) --> () (1,) (2,) (3,) (1,2) (1,3) (2,3) (1,2,3)"
    s = list(iterable)
    return chain.from_iterable(combinations(s, r) for r in range(len(s)+1))

as follows:

In [3]: [ss for ss in powerset([0,1,2]) if ss]
Out[3]: [(0,), (1,), (2,), (0, 1), (0, 2), (1, 2), (0, 1, 2)]

answered May 01 '15 at 18:49

xnx

24,509
11
70
109

Aside from any details regarding order, I need usually the powerset without the empty set and without the full set - so I use the `recipe` above but use `range(1,len(s))` - not sure it is bulletproof... – nate Feb 02 '18 at 22:36
1

@nate I think this is safe: letting `r` range between `1` and `n-1` inclusive omits the empty set and the full set from the call to `combinations`. – xnx Feb 03 '18 at 09:36

Ryan Haining · Answer 3 · 2015-05-01T19:08:22.360

What you want is basically powerset but without the empty set. By modifying the recipe from the python itertools page to start with a combinations size of 1:

from itertools import combinations, chain

def almost_powerset(seq):
    return list(chain.from_iterable(combinations(seq, r) 
                for r in range(1, len(seq)+1)))

and then just pass in your sequence

lst = [0, 1, 2]
res = almost_powerset(lst)

This generates all the combinations of size 1, then size 2, etc, up to the total length of your sequence, then uses chain to connect them.

Subtuples of a tuple

3 Answers3