I have an array in numpy:
A=np.zeros((5,10))
and I want to change one value in each row with ones according to another array N=np.array([7, 2, 9, 4, 5])
like: A[:,N]=1;
0 0 0 0 0 1 0 0 0 0
0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 1
0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 1 0 0 0 0
How can I do that?
Since you want to set a single element per row, you need to fancy-index the first axis using arange(5). this can be thought of as setting indices (I0[0], N[0])=(0,7), (I0[1],N[1])=(1,2), ...
I0 = np.arange(A.shape[0])
A[I0, N] = 1
A
=>
array([[ 0., 0., 0., 0., 0., 0., 0., 1., 0., 0.],
[ 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],
[ 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.]])
A.nonzero()
=> (array([0, 1, 2, 3, 4]), array([7, 2, 9, 4, 5]))