NumPy Array Conversion

Question

I'm new to python and I'm trying to convert a (m,n,1) multidimensional array to (m,n) in a fast way, how can I go about it?

Also given a (m,n,k) array how can I split it to k (m,n) arrays? (each of the k members belongs to a different array)

when you say "array" (the input one) what do you mean? can you show a code example — Nullman, Feb 26 '19 at 15:09
Why do you need to do the latter? What have you tried? This is pretty basic stuff, so can you at least show what documentation you have looked at? — Mad Physicist, Feb 26 '19 at 15:09
It would help if you gave some toy data and an example of what you already tried. — Bram Vanroy, Feb 26 '19 at 15:22

score 3 · Answer 1 · answered Feb 26 '19 at 15:25

3

To reshape array a you can use a.reshape(m,n).

To split array a along the depth dimension, you can use numpy.dsplit(a, a.shape[2]).

answered Feb 26 '19 at 15:25

nkaushik

I like using dsplit, however the returned list will contain arrays of shape (m, n, 1) – randomwalker Feb 26 '19 at 15:27

score 1 · Accepted Answer · answered Feb 26 '19 at 15:24

1

To reshape a NumPy Array arr with shape (m, n, 1) to the shape (m, n) simply use:

arr = arr.reshape(m, n)

You can get a list of (m, n)-shaped arrays out of a (m, n, k) shaped array arr_k by:

array_list = [arr_k[:, :, i] for i in range(arr_k.shape[2])]

answered Feb 26 '19 at 15:24

randomwalker

2 Answers2