I have a array which must keep its dtype fixed. However, after a append statement, its dtype changes. How can I append values without changing the dtype?
vertices = array([0.5, 0.0, 0.0, 1.0, 0.0, 0.0,
0.0, 0.5, 0.0, 0.0, 1.0, 0.0,
0.0, -0.5, 0.0, 0.0, 0.5, 0.0], dtype=np.float32)
print(vertices.dtype)
vertices = append(vertices, [-0.5, 0.0, 0.0, 0.0, 0.0, 1.0])
print(vertices.dtype)
Output: float32 float64
from numpy import *
import numpy as np
vertices = array([0.5, 0.0, 0.0, 1.0, 0.0, 0.0,
0.0, 0.5, 0.0, 0.0, 1.0, 0.0,
0.0, -0.5, 0.0, 0.0, 0.5, 0.0], dtype=np.float32)
print(vertices.dtype)
vertices = append(vertices, np.array([-0.5, 0.0, 0.0, 0.0, 0.0, 1.0], dtype=np.float32))
print(vertices.dtype)
random_arr = [-0.5, 0.0, 0.0, 0.0, 0.0, 1.0]
print(np.array(random_arr).dtype)
float32
float32
float64
By default, numpy assigns float64 datatype on your float array (check the last random_arr), so once you concatenate one float32 and one float64 array, obviously final array will be casted to float64. So, just specify the dtype when creating numpy arrays to be safe.
You can append the new data as a numpy array, passing in the type when you convert it as well.
import numpy as np
vertices = np.array([0.5, 0.0, 0.0, 1.0, 0.0, 0.0,
0.0, 0.5, 0.0, 0.0, 1.0, 0.0,
0.0, -0.5, 0.0, 0.0, 0.5, 0.0], dtype=np.float32)
print(vertices.dtype)
vertices = np.append(vertices, np.array([-0.5, 0.0, 0.0, 0.0, 0.0, 1.0], dtype=np.float32))
print(vertices.dtype)
np.appendis just a dumb front end tonp.concatenate. You don't need it (none of us do!). Just make sure all inputs have the desireddtype.