I have a 2d array, A, with shape (n x m), where each element of the array at position (i,j) holds a third value k. I want to increment a 3d array with dimensions nxmxl at position (k,i,j) based on the 2d array value and position.
So for example if
A = [[0,1],[3,3]] -> I would want B to be [[[1,0], [0,0]], [0,1], [0,0]], [0,0], [0,1]], [0,0], [0,2]]]
How do you do this in numpy efficiently?
Advertisement
Answer
The question is somewhat ambiguous, but if the intent is to increment some unknown array B
at indices (0,0,0)
, (1,0,1)
, (3,1,0)
, and (3,1,1)
, then the following should be fine:
B[(A.ravel(), ) + np.unravel_index(np.arange(np.prod(A.shape)), A.shape)] += increment
For example:
A = np.array([[0,1],[3,3]]) B = np.zeros((4,2,2), dtype=int) increment = 1 B[(A.ravel(), ) + np.unravel_index(np.arange(np.prod(A.shape)), A.shape)] += increment >>> B array([[[1, 0], [0, 0]], [[0, 1], [0, 0]], [[0, 0], [0, 0]], [[0, 0], [1, 1]]])
Another way of doing the same thing is:
w, h = A.shape indices = (A.ravel(),) + tuple(np.mgrid[:w, :h].reshape(2, -1)) # then B[indices] += increment