Initializing 2D lists in Python: How to make deep copies of each row?

Question

Let\'s say I want to initialize a 2D Python list with all 0\'s, I\'d do something like:

test = [[0.0] * 10] * 10

Then I start modifying val

Answer 1

Doing [[0.0] * 10] * 10actually creates multiple copies of the same list so modifying one will affect all of them:

>>> test = [[0.0] * 10] * 10
>>> [id(x) for x in test]     #see all IDs are same
[3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L, 3065020524L]

Try this:

>>> test = [[0.0]*10 for _ in xrange(10)]
>>> test[0][0] = 1.0
>>> test
[[1.0, 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], [0.0, 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]]

Integers/floats are immutable, while lists are mutable:

>>> x = y = []
>>> x is y       # Both point to the same object
True 
>>> x.append(1)  # list.extend modifies a list in-place
>>> x,y          # both references to [] can see the modification
([1], [1])
>>> x = y = 1
>>> x is y      #both points to the same object
True
>>> x+=1        # only x gets modified, it now points to a new object 2
>>> x,y         # y still points to the same object 1
(2, 1)

Answer 2

The list test contains multiple iterations of the same list, hence a change in one (as you are making by reassigning the first element of test[0]) is reflected in all the others. Try this instead:

[[0.0]*10 for _ in xrange(10)]  # or `range` in Python 3.x

Of course, you wouldn't need to worry about this if all you had was [0.0] * 10, since this creates a list of ints, none of which can ever mutate. Lists, on the other hand, are indeed mutable.