I have a numpy array A, which contains values between 0 and 1. I want to create another numpy array y, such that the value of y(i) = 1 if A(i) >= 0.5, and y(i) = 0 if A(i) < 0.5. I used the following python code:
f=lambda v: 1 if v>0.5 else 0
vf=np.vectorize(f)
Y=vf(A)
Is there a way to do this function in one line command instead of three lines?
Use a vectorized comparison and cast the result to int:
(A >= 0.5).astype(int)
A >= 0.5 produces an array of elementwise >= 0.5 comparison results, and astype(int) casts True to 1 and False to 0.
If you can live with single byte integers
(A >= 0.5).view(np.int8)
is a bit faster. Unlike astype view does not create new data. It reinterprets the data buffer of its operand
import numpy
A = numpy.random.rand(10)
print(A)
Array A:
[ 0.76702953 0.89697124 0.54573644 0.48079479 0.39556016 0.50646642
0.45998033 0.11159339 0.69824144 0.37451713]
Create another numpy array y, such that the value of y(i) = 1 if A(i) >= 0.5, and y(i) = 0 if A(i) < 0.5.
Y = (A >= 0.5).astype(int)
print(Y)
Array Y:
[1 1 1 0 0 1 0 0 1 0]
