Question

I am a recent immigrant to the world of Python. I need to figure out how to copy a matrix to a larger matrix in Python. Let me illustrate this with an example in Matlab:

A = randn(4,4);
B = eye(2,2);
A(1:2,1:2) = B

gives

A = 

 1.0000          0     3.5784     0.7254
      0     1.0000     2.7694   - 0.0631
-2.2588   - 0.4336   - 1.3499     0.7147
 0.8622     0.3426     3.0349   - 0.2050  

I am trying a similar thing with Python using NumPy in the following fashion.

A = np.random.randn(4,4)
B = np.eye(2,2)
A[0:1,0:1] = B

gives

ValueError: output operand requires a reduction, but reduction is not enabled.

Of course, the simplest way to avoid this is to use a loop but I would want to keep it vectorized.

Can someone please point me to a way of doing this without using for loops?

Answer 1

Here’s the problem:

A[0:1,0:1] = B

You want:

A[0:2,0:2] = B

Why? Because Python uses half-open ranges. So the slice [0:1] is the half-open range [0, 1), meaning just the index 0; the slice [0:2] is the half-open range [0, 2), meaning the indices 0 and 1.