All combinations of a list of lists [duplicate]

Question

I'm basically looking for a python version of Combination of List<List<int>>

Given a list of lists, I need a new list that gives all the possible combinations of items between the lists.

[[1,2,3],[4,5,6],[7,8,9,10]] -> [[1,4,7],[1,4,8],...,[3,6,10]]

The number of lists is unknown, so I need something that works for all cases. Bonus points for elegance!

Moreno

Answer 1

you need itertools.product:

>>> import itertools
>>> a = [[1,2,3],[4,5,6],[7,8,9,10]]
>>> list(itertools.product(*a))
[(1, 4, 7), (1, 4, 8), (1, 4, 9), (1, 4, 10), (1, 5, 7), (1, 5, 8), (1, 5, 9), (1, 5, 10), (1, 6, 7), (1, 6, 8), (1, 6, 9), (1, 6, 10), (2, 4, 7), (2, 4, 8), (2, 4, 9), (2, 4, 10), (2, 5, 7), (2, 5, 8), (2, 5, 9), (2, 5, 10), (2, 6, 7), (2, 6, 8), (2, 6, 9), (2, 6, 10), (3, 4, 7), (3, 4, 8), (3, 4, 9), (3, 4, 10), (3, 5, 7), (3, 5, 8), (3, 5, 9), (3, 5, 10), (3, 6, 7), (3, 6, 8), (3, 6, 9), (3, 6, 10)]

Prehuman

Could someone explain the meaning of the asterisk in *a? – Bigner

*a means these are arguments being passed to the function or method. def fn(a,b,c): would respond to fn(*[1,2,3]) reference – Rale

@mjallday, would it be possible to add also these combinations: (7,4,1),(8,4,1),(9,4,1),(10,4,1),(7,5,1),(8,5,1),(9,5,1),(10,5,1) etc? – Selda

@Selda It's not entirely clear what you want to get but if it is, for example, also the reverse of each tuple you can use a a wrapper function that takes a as input, iterates over itertools.product(*a) and yields both the tuple produced by itertools and a reverse version (e.g. create a list, reverse() it and convert back to tuple). Best ask a new question. – Kellsie

if you are familiar with javascript [*a] would be the same the js spread operator [...a]. This is also true for dicts – Brushoff

Answer 2

Simply use itertools.product:

listOLists = [[1,2,3],[4,5,6],[7,8,9,10]]
for l in itertools.product(*listOLists):
    print(l)

Bibbye

Answer 3

The most elegant solution is to use itertools.product in python 2.6.

If you aren't using Python 2.6, the docs for itertools.product actually show an equivalent function to do the product the "manual" way:

def product(*args, **kwds):
    # product('ABCD', 'xy') --> Ax Ay Bx By Cx Cy Dx Dy
    # product(range(2), repeat=3) --> 000 001 010 011 100 101 110 111
    pools = map(tuple, args) * kwds.get('repeat', 1)
    result = [[]]
    for pool in pools:
        result = [x+[y] for x in result for y in pool]
    for prod in result:
        yield tuple(prod)

Adal

Answer 4

Nothing wrong with straight up recursion for this task, no need for external dependencies, and if you need a version that works with strings, this might fit your needs:

combinations = []

def combine(terms, accum):
    last = (len(terms) == 1)
    n = len(terms[0])
    for i in range(n):
        item = accum + terms[0][i]
        if last:
            combinations.append(item)
        else:
            combine(terms[1:], item)


>>> a = [['ab','cd','ef'],['12','34','56']]
>>> combine(a, '')
>>> print(combinations)
['ab12', 'ab34', 'ab56', 'cd12', 'cd34', 'cd56', 'ef12', 'ef34', 'ef56']

Newsstand

Answer 5

Numpy can do it:

 >>> import numpy
 >>> a = [[1,2,3],[4,5,6],[7,8,9,10]]
 >>> [list(x) for x in numpy.array(numpy.meshgrid(*a)).T.reshape(-1,len(a))]
[[ 1, 4, 7], [1, 5, 7], [1, 6, 7], ....]

Beaut

Could someone explain this? – Ceuta

This does not work if the things in a are arrays. a=[[[array A], [array B], [array C]], [...] – Involucel

this feels like asking 'how do i boil water' and instead of saying 'use a kettle', you get 'lower the atmospheric pressure around the water'. Both work! – Goolsby

Answer 6

One can use base python for this. The code needs a function to flatten lists of lists:

def flatten(B):    # function needed for code below;
    A = []
    for i in B:
        if type(i) == list: A.extend(i)
        else: A.append(i)
    return A

Then one can run:

L = [[1,2,3],[4,5,6],[7,8,9,10]]

outlist =[]; templist =[[]]
for sublist in L:
    outlist = templist; templist = [[]]
    for sitem in sublist:
        for oitem in outlist:
            newitem = [oitem]
            if newitem == [[]]: newitem = [sitem]
            else: newitem = [newitem[0], sitem]
            templist.append(flatten(newitem))

outlist = list(filter(lambda x: len(x)==len(L), templist))  # remove some partial lists that also creep in;
print(outlist)

Output:

[[1, 4, 7], [2, 4, 7], [3, 4, 7], 
[1, 5, 7], [2, 5, 7], [3, 5, 7], 
[1, 6, 7], [2, 6, 7], [3, 6, 7], 
[1, 4, 8], [2, 4, 8], [3, 4, 8], 
[1, 5, 8], [2, 5, 8], [3, 5, 8], 
[1, 6, 8], [2, 6, 8], [3, 6, 8], 
[1, 4, 9], [2, 4, 9], [3, 4, 9], 
[1, 5, 9], [2, 5, 9], [3, 5, 9], 
[1, 6, 9], [2, 6, 9], [3, 6, 9], 
[1, 4, 10], [2, 4, 10], [3, 4, 10], 
[1, 5, 10], [2, 5, 10], [3, 5, 10], 
[1, 6, 10], [2, 6, 10], [3, 6, 10]]

Rodl

Answer 7

This mostly mimics solutions like Answer by Jarret Hardie using itertools.product, but has these distinctions:

• this passes parameters to itertools.product in-line, instead of via variable a - so no *args syntax needed on the inline parameters
• if your mypy type-linter acts like mine, and you can get your code to otherwise "work" with the *args syntax with inline product parameters (like product(*[[1,2,3],[4,5,6],[7,8,9,10]])), mypy might still fail it (with something like error: No overload variant of "product" matches argument type "List[object]")
• So solution to that mypy, is to not use *args syntax, like this:

    >>> import itertools
    >>> list(itertools.product([1,2,3],[4,5,6],[7,8,9,10]))
    [(1, 4, 7), (1, 4, 8), (1, 4, 9), (1, 4, 10), (1, 5, 7), (1, 5, 8), (1, 5, 9), (1, 5, 10), (1, 6, 7), (1, 6, 8), (1, 6, 9), (1, 6, 10), (2, 4, 7), (2, 4, 8), (2, 4, 9), (2, 4, 10), (2, 5, 7), (2, 5, 8), (2, 5, 9), (2, 5, 10), (2, 6, 7), (2, 6, 8), (2, 6, 9), (2, 6, 10), (3, 4, 7), (3, 4, 8), (3, 4, 9), (3, 4, 10), (3, 5, 7), (3, 5, 8), (3, 5, 9), (3, 5, 10), (3, 6, 7), (3, 6, 8), (3, 6, 9), (3, 6, 10)]

Withstand

Answer 8

This answer isn't as clean as using itertools but the ideas could be useful.

Drawing inspiration from the construction of zip() here, we could do the following.

>>> a = iter([[1,2,3],[4,5,6],[7,8,9,10]])
>>> sentinel = object()
>>> result = [[]]
>>> while True:
>>>     l = next(a,sentinel)
>>>     if l == sentinel:
>>>         break
>>>     result = [ r + [digit] for r in result for digit in l]
>>> print(result)
[[1, 4, 7], [1, 4, 8], [1, 4, 9], [1, 4, 10], [1, 5, 7], [1, 5, 8], [1, 5, 9], [1, 5, 10], [1, 6, 7], [1, 6, 8], [1, 6, 9], [1, 6, 10], [2, 4, 7], [2, 4, 8], [2, 4, 9], [2, 4, 10], [2, 5, 7], [2, 5, 8], [2, 5, 9], [2, 5, 10], [2, 6, 7], [2, 6, 8], [2, 6, 9], [2, 6, 10], [3, 4, 7], [3, 4, 8], [3, 4, 9], [3, 4, 10], [3, 5, 7], [3, 5, 8], [3, 5, 9], [3, 5, 10], [3, 6, 7], [3, 6, 8], [3, 6, 9], [3, 6, 10]]

We use a as an iterator in order to successively get the next item of it without needing to know how many there are a priori. The next command will output sentinel (which is an object created solely to make this comparison, see here for some explanation) when we run out of lists in a, causing the if statement to trigger so we break out of the loop.

Bushwhacker

Indeed this is more or less what itertools.product does, -- see here -- except the use of a sentinel while itertools.product uses yield. – Bushwhacker

Answer 9

from itertools import product 
list_vals = [['Brand Acronym:CBIQ', 'Brand Acronym :KMEFIC'],['Brand Country:DXB','Brand Country:BH']]
list(product(*list_vals))

Output:

[('Brand Acronym:CBIQ', 'Brand Country :DXB'),
('Brand Acronym:CBIQ', 'Brand Country:BH'),
('Brand Acronym :KMEFIC', 'Brand Country :DXB'),
('Brand Acronym :KMEFIC', 'Brand Country:BH')]

Corded

This answer should be accepted, since it's the only one using a built-in function, while highlighting that it also works for any and also heterogenous types. – Irrelevancy

How is this answer any different than the one provided many years ago? – Weepy

The types of the parameters here are homogenous, not hetrogenous. – Withstand