How can I expand one axis to 2 axes in NumPy?-CodePudding

Let's say I have a string ndarray with any dimension. For example: [["abc", "def"], ["ghi", "jkl"]]. Now, I want to split each string into separate chars, such that basically an axis in the second dimension is added: [[['a', 'd'], ['b', 'e'], ['c', 'f']], [['g', 'j'], ['h', 'k'], ['i', 'l']]]. Or better said, it should behave like MATLAB converting a string array to a char array:

A = 
  2×2 string array
    "abc"    "def"
    "ghi"    "jkl"

should become:

  2×3×2 char array
ans(:,:,1) =
    'abc'
    'ghi'
ans(:,:,2) =
    'def'
    'jkl'

I tried functions like np.frompyfunc, np.apply_over_axis and np.apply_from_axis but so far nothing worked for me. Is there a clever trick to do this?

The reverse is actually pretty simple:

def row_to_string(row):
    return ''.join([chr(int(x)) for x in row])

return np.apply_along_axis(row_to_string, 1, np.asarray(x))

CodePudding user response：

Here you go:

In [1]: A = [["abc", "def"], ["ghi", "jkl"]]

In [2]: [[[*t] for t in zip(*a)] for a in A]
Out[2]: [[['a', 'd'], ['b', 'e'], ['c', 'f']], [['g', 'j'], ['h', 'k'], ['i', 'l']]]

CodePudding user response：

In [159]: alist = [["abc", "def"], ["ghi", "jkl"]]
In [160]: np.frompyfunc(list,1,1)(alist)
Out[160]: 
array([[list(['a', 'b', 'c']), list(['d', 'e', 'f'])],
       [list(['g', 'h', 'i']), list(['j', 'k', 'l'])]], dtype=object)
In [161]: np.array(np.frompyfunc(list,1,1)(alist).tolist())
Out[161]: 
array([[['a', 'b', 'c'],
        ['d', 'e', 'f']],

       [['g', 'h', 'i'],
        ['j', 'k', 'l']]], dtype='<U1')

From there you can transpose to the desired layout.

Splitting a string into characters is a python task. Numpy doesn't have special string code, just np.char functions that apply string methods to elements of an string dtype array.

frompyfunc is a convenience tool for applying a python function to elements of an array. It doesn't compile anything, but usually is comparable to list comprehensions in speed, and may be more convenient.