Question

I have a simple function called get_gradient which takes a numpy array of [[x,y,Vx,Vy]] and returns (should return) an array of [[Vx,Vy,Ax,Ay]]. I want to update my k[0][0:2] to equal state[0][2:4] but when I print them out k never changes value. I created a copy of the initial array in the hope that it would sort it out but it still doesn’t work! any help would be much appreciated

Y          = numpy.array([[0, 0, 0, 0]]) # [x,y,Vx,Vy]
k          = numpy.array([[0, 0, 0, 0]]) # [x,y,Vx,Vy]

def get_gradient(Y):
    state = Y.copy()
    spring_force   = spring.force(state[0], feet_pos[0])
    brownian_force = mass.brownian(dt)
    drag_force     = mass.drag(state[0][2:4])
    total_force    = spring_force + brownian_force + drag_force     #2d array [Fx,Fy]

    k[0][0:2] = state[0][2:4]            ##### DOESN'T WORK!!!
    k[0][2:4] = total_force/mass.mass    ##### but this works :s

    print k[0][0:2]
    print state[0][2:4]

    return k

Answer 1

It’s not that k isn’t changing, it’s that it has an integer dtype, and so when you try to assign a float to it, you might not get what you expect:

>>> k
array([[   0,    0, 4863, 1464]])
>>> k.dtype
dtype('int64')
>>> k[0][0] = 10
>>> k
array([[  10,    0, 4863, 1464]])
>>> k[0][0] = 1e-3
>>> k
array([[   0,    0, 4863, 1464]])

or

>>> k
array([[  0,    0, 4863, 1464]])
>>> k[0][0] = 10.2
>>> k
array([[  10,    0, 4863, 1464]])

but:

>>> f = k.astype(float)
>>> f
array([[    0.,     0.,  4863.,  1464.]])
>>> f[0][0] = 1e-3
>>> f
array([[  1.00000000e-03,   0.00000000e+00,   4.86300000e+03,
          1.46400000e+03]])

Notice the . after the zeros.

I would probably use

>>> k = numpy.zeros((1,4))
>>> k
array([[ 0.,  0.,  0.,  0.]])

instead; it defaults to float dtype.