Looping through a list of lists

Question

I am new to programming and I have this basic problem I cannot fix. I simplified it as much as possible. In this simplified version, I am iterating through an empty list. I just want to store the indices in the "symmetric matrix":

n = 2
B = [[None] * n] * n
print B, "\n"
for i in range(n):
    for j in range(n):
        B[i][j] = [i, j]
        print B

Initially, the list looks like this:

[[None, None], [None, None]] 

After looping trough, I would expect the print-out to be:

[[[0, 0], None], [None, None]]
[[[0, 0], [0, 1]], [None, None]]
[[[1, 0], [0, 1]], [[1, 0], None]]
[[[1, 0], [1, 1]], [[1, 0], [1, 1]]]

Instead, I get this:

[[[0, 0], None], [[0, 0], None]]
[[[0, 0], [0, 1]], [[0, 0], [0, 1]]]
[[[1, 0], [0, 1]], [[1, 0], [0, 1]]]
[[[1, 0], [1, 1]], [[1, 0], [1, 1]]]

What am I missing? Thanks for helping out...

Answer 1

Your initialization creates two references to the same list:

n=2
B=[[None]*n]*n
for x in B:
     print id(x)

Output:

4534759576
4534759576

Answer 2

What you need instead of the current way you're defining B is n "fresh" instances of the sublist:

B = [[None] * n for _ in range(n)]

which is equivalent to but shorter and (for a Pythonista) more readable than:

B = []
for _ in range(n):
    B.append([None] * n)

This is because lists in Python are not values but objects; ie they are not copied by default. For example:

>>> a = []
>>> b = [a, a]  # list of 2 items, both pointing to the same list instance
>>> b[0].append(1)  # b[0] refers to the same list instance also referenced by a
>>> print b
[[1], [1]]

whereas:

>>> b = [[], []]  # list of 2 distinct sublists
>>> b[0].append(1)
>>> print b
[[1], []]

If you apply that knowledge to

B = [[None] * n] * n

you'll see there's an error in the construction—you're creating a list that contains another list of n None values; you're then "upsizing" the list to contain n of such lists; however, the same list instance will be used just like in the above, more trivial examples.

PS more accurately, everything in Python is an object not a value, except in cases where the object is immutable, such as int s and float s and str s etc, it is as if they are values and are always copied—this is because if an object is immutable, you might as well pretend a new reference to it is a new copy of it, because it looks the same and cannot be modified.

Answer 3

Here is how you could do, what you want:

    n=2
    B = [[None for i in range(n)] for i in range(n)]
    print B, "\n"
    for i in range(n):
        for j in range(n):
        B[i][j] = [i, j]
        print B

Answer 4

The key to understanding what is happening here is understanding that you are creating lists with shared references when you use the [[None]*n]*n method. Have a look at this visualisation of what your code does, it should make things more clear. This blog post also explains how names and values work in Python.