Question

Let’s say I have the following code:

a_list = [[0]*10]*10

This generates the following list:

[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]

Then I want to modify the first element in the first list:

a_list[0][0] = 23

I expected only the first element of the list to be modified, but actually the first element of each list was changed:

[[23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [23, 0, 0, 0, 0, 0, 0, 0, 0, 0]]

I managed to find another way to represent my data to avoid this but why is this happening? Why isn’t just the first list changed? When I do the second *10, does Python actually copy the first list’s address instead of allocating a new memory block?

Answer 1

Your hunch about copying addresses is correct. Think about it like this:

sub_list = [0] * 10
a_list = [sub_list] * 10

This code is actually equivalent to the code you have posted above. What this means is that you are actually changing the same list sub_list whenever you change any element of a_list. You can even make sure of it by typing:

a_list = [[0] * 10] * 10
for n in a_list:
    print id(n)

And it will show up the same for every element. To remedy this, you should use:

a_list = [[0] * 10 for _ in range(10)]

In order to create a new sublist for every element of a_list.