Python multi-dimensional array initialization without a loop

Question

Is there a way in Python to initialize a multi-dimensional array / list without using a loop?

Answer 1

The following does not use any special library, nor eval:

arr = [[0]*5 for x in range(6)]

and it doesn't create duplicated references:

>>> arr[1][1] = 2
>>> arr
[[0, 0, 0, 0, 0],
 [0, 2, 0, 0, 0],
 [0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0]]

Answer 2

It depends on what you what to initialize the array to, but sure. You can use a list comprehension to create a 5×3 array, for instance:

>>> [[0 for x in range(3)] for y in range(5)]
[[0, 0, 0], [0, 0, 0], [0, 0, 0], [0, 0, 0], [0, 0, 0]]

>>> [[3*y+x for x in range(3)] for y in range(5)]
[[0, 1, 2], [3, 4, 5], [6, 7, 8], [9, 10, 11], [12, 13, 14]]

Yes, I suppose this still has loops—but it's all done in one line, which I presume is the intended meaning of your question?

Answer 3

Depending on your real needs, the de facto "standard" package Numpy might provide you with exactly what you need.

You can for instance create a multi-dimensional array with

numpy.empty((10, 4, 100))  # 3D array

(initialized with arbitrary values) or create the same arrays with zeros everywhere with

numpy.zeros((10, 4, 100))

Numpy is very fast, for array operations.

Answer 4

Sure there is a way

arr = eval(`[[0]*5]*10`)

or

arr = eval(("[[0]*5]+"*10)[:-1])

but it's horrible and wasteful, so everyone uses loops (usually list comprehensions) or numpy

Answer 5

I don't believe it's possible.

You can do something like this:

>>> a = [[0] * 5] * 5

to create a 5x5 matrix, but it is repeated objects (which you don't want). For example:

>>> a[1][2] = 1
[[0, 0, 1, 0, 0], [0, 0, 1, 0, 0], [0, 0, 1, 0, 0], [0, 0, 1, 0, 0], [0, 0, 1, 0, 0]]

You almost certainly need to use some kind of loop as in:

[[0 for y in range(5)] for x in range(5)]

Answer 6

If you're doing numerical work using Numpy, something like

x = numpy.zeros ((m,n))
x = numpy.ones ((m,n))