[PYTHON] Apply functions to rows or columns of numpy.array without using list comprehensions

Thing you want to do

>>> def dobule(x):
...     print(x)
...     return x * 2
...
>>> A = np.arange(6).reshape((3,2))
>>> A
array([[0, 1],
       [2, 3],
       [4, 5]])

Consider how to apply your own function in rows or columns of ʻA`.

>>> np.array([double(a) for a in A])
[0 1]
[2 3]
[4 5]
array([[ 0,  2],
       [ 4,  6],
       [ 8, 10]])

This can be achieved by using list comprehension as above, I want to think of a way to avoid using a for statement.

numpy.vectorize

https://docs.scipy.org/doc/numpy/reference/generated/numpy.vectorize.html

>>> import numpy as np
>>> np.vectorize(double)(A)
0
0
1
2
3
4
5
array([[ 0,  2],
       [ 4,  6],
       [ 8, 10]])

(Why is it printed 7 times even though the number of elements is 6?)

With numpy.vectorize, the function will be applied element by element, It cannot be applied on a row or column basis.

numpy.apply_along_axis

https://docs.scipy.org/doc/numpy/reference/generated/numpy.apply_along_axis.html

>>> np.apply_along_axis(double, 0, A)
[0 2 4]
[1 3 5]
array([[ 0,  2],
       [ 4,  6],
       [ 8, 10]])
>>> np.apply_along_axis(double, 1, A)
[0 1]
[2 3]
[4 5]
array([[ 0,  2],
       [ 4,  6],
       [ 8, 10]])

You can use numpy.apply_along_axis to apply a function on a row or column basis.