vectorize numpy append loop

Question

I am trying to vectorize the following loop in I am trying to append multiple arrays to an empty array.

# ff is a matrix of shape [100,1,96]
temp = np.array([]).reshape(0,96)
for kk in range(1,10,1):
   temp = np.append(tr,ff[kk],axis=0)
temp = temp.reshape(1,10,96)

Is it possible to vectorize the above loop using numpy? Any help is welcome!

Answer 1

You can use slicing to extract the data you need:

ff[:10,:,:]

That will yield an array of shape (10, 1, 96). To get rid of the empty dimension, you can run it through numpy.squeeze():

numpy.squeeze(ff[:10,:,:])

and get an array of shape (10, 96)