How to "embed" a small numpy array into a predefined block of a large numpy array?

Question

I have a small NXN array "block" that I want to plug into a specified region (i.e., a diagonal region at "start") of a large array "wall". Is there an efficient method to archive this?

wall[start:start+N][start:start+N] = block[:][:] 

currently what I am doing is simply:

for i in xrange(N):     wall[start+i][start:start+N] = block[i][:] 

Answer 1

you can use multi dimension index:

import numpy as np  wall = np.zeros((10,10),dtype=np.int) block = np.arange(1,7).reshape(2,3)  x = 2 y = 3 wall[x:x+block.shape[0], y:y+block.shape[1]] = block 

the output is:

>>> wall array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 1, 2, 3, 0, 0, 0, 0],        [0, 0, 0, 4, 5, 6, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]) 

Answer 2

Here is a solution that works even if the position given goes off the edges of the wall numpy array, and works for any number of dimensions for your wall and block (note: the provided location loc needs to be a tuple, even in the 1D case).

def paste_slices(tup):   pos, w, max_w = tup   wall_min = max(pos, 0)   wall_max = min(pos+w, max_w)   block_min = -min(pos, 0)   block_max = max_w-max(pos+w, max_w)   block_max = block_max if block_max != 0 else None   return slice(wall_min, wall_max), slice(block_min, block_max)  def paste(wall, block, loc):   loc_zip = zip(loc, block.shape, wall.shape)   wall_slices, block_slices = zip(*map(paste_slices, loc_zip))   wall[wall_slices] = block[block_slices] 

---

Tests:

1D

>>> b = np.zeros([10]) >>> a = np.arange(1, 5) >>> b array([ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]) >>> a array([1, 2, 3, 4]) >>> paste(b, a, (8,)) >>> b array([ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  1.,  2.]) 

2D

>>> b = np.zeros([10, 10]) >>> a = np.arange(1,33).reshape(4,8) >>> b array([[ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],        [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]]) >>> a array([[ 1,  2,  3,  4,  5,  6,  7,  8],        [ 9, 10, 11, 12, 13, 14, 15, 16],        [17, 18, 19, 20, 21, 22, 23, 24],        [25, 26, 27, 28, 29, 30, 31, 32]]) >>> paste(b, a, (-1, -3)) >>> b array([[ 12.,  13.,  14.,  15.,  16.,   0.,   0.,   0.,   0.,   0.],        [ 20.,  21.,  22.,  23.,  24.,   0.,   0.,   0.,   0.,   0.],        [ 28.,  29.,  30.,  31.,  32.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.],        [  0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   0.]])