# How to get the Cartesian product of multiple lists

## Question:

How can I get the Cartesian product (every possible combination of values) from a group of lists?

For example, given

``````somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]
``````

How do I get this?

``````[(1, 'a', 4), (1, 'a', 5), (1, 'b', 4), (1, 'b', 5), (2, 'a', 4), (2, 'a', 5), ...]
``````

One common application for this technique is to avoid deeply nested loops. See Avoiding nested for loops for a more specific duplicate. Similarly, this technique might be used to "explode" a dictionary with list values; see Combine Python Dictionary Permutations into List of Dictionaries .

If you want a Cartesian product of the same list with itself multiple times, `itertools.product` can handle that elegantly. See Operation on every pair of element in a list or How can I get "permutations with repetitions" from a list (Cartesian product of a list with itself)?.

Many people who already know about `itertools.product` struggle with the fact that it expects separate arguments for each input sequence, rather than e.g. a list of lists. The accepted answer shows how to handle this with `*`. However, the use of `*` here to unpack arguments is fundamentally not different from any other time it’s used in a function call. Please see Expanding tuples into arguments for this topic (and use that instead to close duplicate questions, as appropriate).

Use `itertools.product`, which has been available since Python 2.6.

``````import itertools

somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]
for element in itertools.product(*somelists):
print(element)
``````

This is the same as:

``````for element in itertools.product([1, 2, 3], ['a', 'b'], [4, 5]):
print(element)
``````
``````import itertools
>>> for i in itertools.product([1,2,3],['a','b'],[4,5]):
...         print i
...
(1, 'a', 4)
(1, 'a', 5)
(1, 'b', 4)
(1, 'b', 5)
(2, 'a', 4)
(2, 'a', 5)
(2, 'b', 4)
(2, 'b', 5)
(3, 'a', 4)
(3, 'a', 5)
(3, 'b', 4)
(3, 'b', 5)
>>>
``````

With itertools.product:

``````import itertools
result = list(itertools.product(*somelists))
``````

In Python 2.6 and above, you can use ‘itertools.product`. In older versions of Python you can use the following (almost — see documentation) equivalent code from the documentation, at least as a starting point:

``````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)
``````

The result of both is an iterator, so if you really need a list for further processing, use `list(result)`.

For Python 2.5 and older:

``````>>> [(a, b, c) for a in [1,2,3] for b in ['a','b'] for c in [4,5]]
[(1, 'a', 4), (1, 'a', 5), (1, 'b', 4), (1, 'b', 5), (2, 'a', 4),
(2, 'a', 5), (2, 'b', 4), (2, 'b', 5), (3, 'a', 4), (3, 'a', 5),
(3, 'b', 4), (3, 'b', 5)]
``````

Here’s a recursive version of `product()` (just an illustration):

``````def product(*args):
if not args:
return iter(((),)) # yield tuple()
return (items + (item,)
for items in product(*args[:-1]) for item in args[-1])
``````

Example:

``````>>> list(product([1,2,3], ['a','b'], [4,5]))
[(1, 'a', 4), (1, 'a', 5), (1, 'b', 4), (1, 'b', 5), (2, 'a', 4),
(2, 'a', 5), (2, 'b', 4), (2, 'b', 5), (3, 'a', 4), (3, 'a', 5),
(3, 'b', 4), (3, 'b', 5)]
>>> list(product([1,2,3]))
[(1,), (2,), (3,)]
>>> list(product([]))
[]
>>> list(product())
[()]
``````

Here is a recursive generator, which doesn’t store any temporary lists

``````def product(ar_list):
if not ar_list:
yield ()
else:
for a in ar_list:
for prod in product(ar_list[1:]):
yield (a,)+prod

print list(product([[1,2],[3,4],[5,6]]))
``````

Output:

``````[(1, 3, 5), (1, 3, 6), (1, 4, 5), (1, 4, 6), (2, 3, 5), (2, 3, 6), (2, 4, 5), (2, 4, 6)]
``````

Just to add a bit to what has already been said: if you use SymPy, you can use symbols rather than strings which makes them mathematically useful.

``````import itertools
import sympy

x, y = sympy.symbols('x y')

somelist = [[x,y], [1,2,3], [4,5]]
somelist2 = [[1,2], [1,2,3], [4,5]]

for element in itertools.product(*somelist):
print element
``````

I would use list comprehension:

``````somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]

cart_prod = [(a,b,c) for a in somelists for b in somelists for c in somelists]
``````

A minor modification to the above recursive generator solution in variadic flavor:

``````def product_args(*args):
if args:
for a in args:
for prod in product_args(*args[1:]) if args[1:] else ((),):
yield (a,) + prod
``````

And of course a wrapper which makes it work exactly the same as that solution:

``````def product2(ar_list):
"""
>>> list(product(()))
[()]
>>> list(product2(()))
[]
"""
return product_args(*ar_list)
``````

with one trade-off: it checks if recursion should break upon each outer loop, and one gain: no yield upon empty call, e.g.`product(())`, which I suppose would be semantically more correct (see the doctest).

Regarding list comprehension: the mathematical definition applies to an arbitrary number of arguments, while list comprehension could only deal with a known number of them.

Although there are many answers already, I would like to share some of my thoughts:

## Iterative approach

``````def cartesian_iterative(pools):
result = [[]]
for pool in pools:
result = [x+[y] for x in result for y in pool]
return result
``````

## Recursive Approach

``````def cartesian_recursive(pools):
if len(pools) > 2:
pools = product(pools, pools)
del pools
return cartesian_recursive(pools)
else:
pools = product(pools, pools)
del pools
return pools
def product(x, y):
return [xx + [yy] if isinstance(xx, list) else [xx] + [yy] for xx in x for yy in y]
``````

## Lambda Approach

``````def cartesian_reduct(pools):
return reduce(lambda x,y: product(x,y) , pools)
``````

Recursive Approach:

``````def rec_cart(start, array, partial, results):
if len(partial) == len(array):
results.append(partial)
return

for element in array[start]:
rec_cart(start+1, array, partial+[element], results)

rec_res = []
some_lists = [[1, 2, 3], ['a', 'b'], [4, 5]]
rec_cart(0, some_lists, [], rec_res)
print(rec_res)
``````

Iterative Approach:

``````def itr_cart(array):
results = [[]]
for i in range(len(array)):
temp = []
for res in results:
for element in array[i]:
temp.append(res+[element])
results = temp

return results

some_lists = [[1, 2, 3], ['a', 'b'], [4, 5]]
itr_res = itr_cart(some_lists)
print(itr_res)
``````

I believe this works:

``````def cartesian_product(L):
if L:
return {(a,) + b for a in L
for b in cartesian_product(L[1:])}
else:
return {()}
``````

You can use `itertools.product` in the standard library to get the Cartesian product. Other cool, related utilities in `itertools` include `permutations`, `combinations`, and `combinations_with_replacement`. Here is a link to a Python CodePen for the snippet below:

``````from itertools import product

somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]

result = list(product(*somelists))
print(result)
``````

The following code is a 95% copy from Using NumPy to build an array of all combinations of two arrays; all credits go there! This is said to be much faster since it is only in NumPy.

``````import numpy as np

def cartesian(arrays, dtype=None, out=None):
arrays = [np.asarray(x) for x in arrays]
if dtype is None:
dtype = arrays.dtype
n = np.prod([x.size for x in arrays])
if out is None:
out = np.zeros([n, len(arrays)], dtype=dtype)

m = int(n / arrays.size)
out[:,0] = np.repeat(arrays, m)
if arrays[1:]:
cartesian(arrays[1:], out=out[0:m, 1:])
for j in range(1, arrays.size):
out[j*m:(j+1)*m, 1:] = out[0:m, 1:]
return out
``````

You need to define the dtype as a parameter if you do not want to take the dtype from the first entry for all entries. Take dtype = ‘object’ if you have letters and numbers as items. Test:

``````somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]

[tuple(x) for x in cartesian(somelists, 'object')]
``````

Out:

``````[(1, 'a', 4),
(1, 'a', 5),
(1, 'b', 4),
(1, 'b', 5),
(2, 'a', 4),
(2, 'a', 5),
(2, 'b', 4),
(2, 'b', 5),
(3, 'a', 4),
(3, 'a', 5),
(3, 'b', 4),
(3, 'b', 5)]
``````

This can be done as

``````[(x, y) for x in range(10) for y in range(10)]
``````

another variable? No problem:

``````[(x, y, z) for x in range(10) for y in range(10) for z in range(10)]
``````

List comprehension is simple and clean:

``````import itertools

somelists = [
[1, 2, 3],
['a', 'b'],
[4, 5]
]
lst = [i for i in itertools.product(*somelists)]
``````

In 99% of cases you should use itertools.product. It is written in efficient C code, so it is probably going to be better than any custom implementation.

In the 1% of cases that you need a Python-only algorithm (for example, if you need to modify it somehow), you can use the code below.

``````def product(*args, repeat=1):
"""Find the Cartesian product of the arguments.

The interface is identical to itertools.product.
"""
# Initialize data structures and handle bad input
if len(args) == 0:
yield () # Match behavior of itertools.product
return
gears = [tuple(arg) for arg in args] * repeat
for gear in gears:
if len(gear) == 0:
return
tooth_numbers =  * len(gears)
result = [gear for gear in gears]

# Rotate through all gears
last_gear_number = len(gears) - 1
finished = False
while not finished:
yield tuple(result)

# Get next result
gear_number = last_gear_number
while gear_number >= 0:
gear = gears[gear_number]
tooth_number = tooth_numbers[gear_number] + 1
if tooth_number < len(gear):
# No gear change is necessary, so exit the loop
result[gear_number] = gear[tooth_number]
tooth_numbers[gear_number] = tooth_number
break
result[gear_number] = gear
tooth_numbers[gear_number] = 0
gear_number -= 1
else:
# We changed all the gears, so we are back at the beginning
finished = True
``````

The interface is the same as for itertools.product. For example:

``````>>> list(product((1, 2), "ab"))
[(1, 'a'), (1, 'b'), (2, 'a'), (2, 'b')]
``````

• It does not build up intermediate results in memory, keeping the memory footprint small.
• It uses iteration instead of recursion, meaning you will not get "maximum recursion depth exceeded" errors.
• It can accept any number of input iterables, making it more flexible than using nested for loops.

This code is based on the itertools.product algorithm from PyPy, which is released under the MIT licence.

If you want to reimplement it yourself, you can try with recursion. Something as simple as:

``````def product(cats, prefix = ()):
if not cats:
yield prefix
else: