How to make a flat list out of list of lists?


I wonder whether there is a shortcut to make a simple list out of list of lists in Python.

I can do that in a for loop, but maybe there is some cool "one-liner"? I tried it with reduce(), but I get an error.


l = [[1, 2, 3], [4, 5, 6], [7], [8, 9]]
reduce(lambda x, y: x.extend(y), l)

Error message

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 1, in <lambda>
AttributeError: 'NoneType' object has no attribute 'extend'
10/16/2019 9:13:02 PM

Accepted Answer

Given a list of lists l,

flat_list = [item for sublist in l for item in sublist]

which means:

flat_list = []
for sublist in l:
    for item in sublist:

is faster than the shortcuts posted so far. (l is the list to flatten.)

Here is the corresponding function:

flatten = lambda l: [item for sublist in l for item in sublist]

As evidence, you can use the timeit module in the standard library:

$ python -mtimeit -s'l=[[1,2,3],[4,5,6], [7], [8,9]]*99' '[item for sublist in l for item in sublist]'
10000 loops, best of 3: 143 usec per loop
$ python -mtimeit -s'l=[[1,2,3],[4,5,6], [7], [8,9]]*99' 'sum(l, [])'
1000 loops, best of 3: 969 usec per loop
$ python -mtimeit -s'l=[[1,2,3],[4,5,6], [7], [8,9]]*99' 'reduce(lambda x,y: x+y,l)'
1000 loops, best of 3: 1.1 msec per loop

Explanation: the shortcuts based on + (including the implied use in sum) are, of necessity, O(L**2) when there are L sublists -- as the intermediate result list keeps getting longer, at each step a new intermediate result list object gets allocated, and all the items in the previous intermediate result must be copied over (as well as a few new ones added at the end). So, for simplicity and without actual loss of generality, say you have L sublists of I items each: the first I items are copied back and forth L-1 times, the second I items L-2 times, and so on; total number of copies is I times the sum of x for x from 1 to L excluded, i.e., I * (L**2)/2.

The list comprehension just generates one list, once, and copies each item over (from its original place of residence to the result list) also exactly once.

3/29/2019 10:29:16 AM

You can use itertools.chain():

import itertools
list2d = [[1,2,3], [4,5,6], [7], [8,9]]
merged = list(itertools.chain(*list2d))

Or you can use itertools.chain.from_iterable() which doesn't require unpacking the list with the * operator:

import itertools
list2d = [[1,2,3], [4,5,6], [7], [8,9]]
merged = list(itertools.chain.from_iterable(list2d))

Note from the author: This is inefficient. But fun, because monoids are awesome. It's not appropriate for production Python code.

>>> sum(l, [])
[1, 2, 3, 4, 5, 6, 7, 8, 9]

This just sums the elements of iterable passed in the first argument, treating second argument as the initial value of the sum (if not given, 0 is used instead and this case will give you an error).

Because you are summing nested lists, you actually get [1,3]+[2,4] as a result of sum([[1,3],[2,4]],[]), which is equal to [1,3,2,4].

Note that only works on lists of lists. For lists of lists of lists, you'll need another solution.


I tested most suggested solutions with perfplot (a pet project of mine, essentially a wrapper around timeit), and found

functools.reduce(operator.iconcat, a, [])

to be the fastest solution, both when many small lists and few long lists are concatenated. (operator.iadd is equally fast.)

enter image description here

enter image description here

Code to reproduce the plot:

import functools
import itertools
import numpy
import operator
import perfplot

def forfor(a):
    return [item for sublist in a for item in sublist]

def sum_brackets(a):
    return sum(a, [])

def functools_reduce(a):
    return functools.reduce(operator.concat, a)

def functools_reduce_iconcat(a):
    return functools.reduce(operator.iconcat, a, [])

def itertools_chain(a):
    return list(itertools.chain.from_iterable(a))

def numpy_flat(a):
    return list(numpy.array(a).flat)

def numpy_concatenate(a):
    return list(numpy.concatenate(a))
    setup=lambda n: [list(range(10))] * n,
    # setup=lambda n: [list(range(n))] * 10,
    n_range=[2 ** k for k in range(16)],
    xlabel="num lists (of length 10)",
    # xlabel="len lists (10 lists total)"

from functools import reduce #python 3

>>> l = [[1,2,3],[4,5,6], [7], [8,9]]
>>> reduce(lambda x,y: x+y,l)
[1, 2, 3, 4, 5, 6, 7, 8, 9]

The extend() method in your example modifies x instead of returning a useful value (which reduce() expects).

A faster way to do the reduce version would be

>>> import operator
>>> l = [[1,2,3],[4,5,6], [7], [8,9]]
>>> reduce(operator.concat, l)
[1, 2, 3, 4, 5, 6, 7, 8, 9]

Don't reinvent the wheel if you use Django:

>>> from django.contrib.admin.utils import flatten
>>> l = [[1,2,3], [4,5], [6]]
>>> flatten(l)
>>> [1, 2, 3, 4, 5, 6]


>>> from pandas.core.common import flatten
>>> list(flatten(l))


>>> import itertools
>>> flatten = itertools.chain.from_iterable
>>> list(flatten(l))


>>> from matplotlib.cbook import flatten
>>> list(flatten(l))


>>> from unipath.path import flatten
>>> list(flatten(l))


>>> from setuptools.namespaces import flatten
>>> list(flatten(l))

Licensed under: CC-BY-SA with attribution
Not affiliated with: Stack Overflow
Email: [email protected]