Search code examples
pythonstacknp

easy way to stack two 3D arrays

I have 2 arrays with the following shape:

array_1 (0,3,4)

and array_2 (0,1,4)

Whats the easiest way to stack or merge them together so I have a single array?

Solution

  • If you want an array of shape (0,4,4), you could append the arrays on the second axis:

    >>> a = np.zeros((0,3,4))
>>> b = np.zeros((0,1,4))
>>> np.append(a, b, axis=1)
array([], shape=(0, 4, 4), dtype=float64)

    After your comment, you want the first array to be in one row and the second to be in another. As they have different shapes, you will end with ragged nested sequences. This is currently deprecated but can be obtained by forcing an object dtype:

    >>> x = np.empty(2, dtype=object)
>>> x[0] = a
>>> x[1] = b

    After your comment to this question, you have:

    a1 = np.zeros((2,3))
a2 = np.zeros((4,2))
a3 = np.zeros((1,4))

    All you need is:

    x = np.array((a1, a2, a3), object)