Double array by duplicating values in Python-CodePudding

Is there any libraries functions (Scipy, Numpy, ?) that can double the array size by duplicating the values ? Like this :

  [[1, 2, 4],        [[1. 1. 2. 2. 4. 4.]
   [5, 6, 2],   ->    [1. 1. 2. 2. 4. 4.]
   [6, 7, 9]]         [5. 5. 6. 6. 2. 2.]
                      [5. 5. 6. 6. 2. 2.]
                      [6. 6. 7. 7. 9. 9.]
                      [6. 6. 7. 7. 9. 9.]]

Or is there a faster way to do this operation ? When the array is very large, it becomes a problem. This is what I have so far.

def double_array(array):
  nbr_rows, nbr_cols = array.shape

  array_new = np.zeros((nbr_rows*2, nbr_cols*2))
  for row in range(nbr_rows):
    for col in range(nbr_cols):
      array_new[row*2,col*2] = array[row,col]
      array_new[row*2 1,col*2] = array[row,col]
      array_new[row*2,col*2 1] = array[row,col]
      array_new[row*2 1,col*2 1] = array[row,col]

  return array_new

array = np.array([[1, 2, 4], [5, 6, 2], [6, 7, 9]])
array_new = double_array(array)
print(array_new)

CodePudding user response：

Based on the comment before, you can try np.repeat, for instance:

array = np.array([[1, 2, 4], [5, 6, 2], [6, 7, 9]])
new_arr = np.repeat(np.repeat(array,2,axis=1), [2]*len(array), axis=0)

print(new_arr.astype(np.float))

Output:

[[1. 1. 2. 2. 4. 4.]
 [1. 1. 2. 2. 4. 4.]
 [5. 5. 6. 6. 2. 2.]
 [5. 5. 6. 6. 2. 2.]
 [6. 6. 7. 7. 9. 9.]
 [6. 6. 7. 7. 9. 9.]]

CodePudding user response：

Using numpy.repeat numpy.tile:

N,M = 2,2 # rows,cols
out = np.repeat(np.tile(array, N), M).reshape(array.shape[0]*N, -1)

output:

array([[1, 1, 2, 2, 4, 4],
       [1, 1, 2, 2, 4, 4],
       [5, 5, 6, 6, 2, 2],
       [5, 5, 6, 6, 2, 2],
       [6, 6, 7, 7, 9, 9],
       [6, 6, 7, 7, 9, 9]])

CodePudding user response：

Just repeat one the second and then first axes:

N = 2
new_arr = a.repeat(N, axis=1).repeat(N, axis=0)

Output:

>>> new_arr
array([[1, 1, 2, 2, 4, 4],
       [1, 1, 2, 2, 4, 4],
       [5, 5, 6, 6, 2, 2],
       [5, 5, 6, 6, 2, 2],
       [6, 6, 7, 7, 9, 9],
       [6, 6, 7, 7, 9, 9]])