Best way to fill numpy array-CodePudding

 a = []
 for j in range(1000):
     a.append(sample_function(x))
 a = np.array(a)

What is the proper and elegant way to replace this code? I don't work working with python list is optimal.

CodePudding user response：

If we assume sample_function(x) is a scaler, you can use:

a = np.full(shape=1000, sample_function(x))  # or smth like --> np.full((1, 1000), sample_function(x))

but if you have an array that you would to apply a function on all of the elements:

a = np.apply_along_axis(sample_function, axis, arr)  # or --> np.apply_over_axes

CodePudding user response：

You can do:

a = example_function(np.arange(1000))

where example_function is arbitrary, i.e.:

def example_function(x):
    return x**2

CodePudding user response：

It looks like you want to create a numpy array using values from a function.

You could try using np.fromfunction like this:

import random  # for example sample function

def sample_function(_):
    return random.random() * 42

a = np.fromfunction(np.vectorize(sample_function), (1000, ))

If you want sample_function to calculate values depending on the coordinates then this is also straight forward:

def sample_function(i):
    return random.random() * i

CodePudding user response：

just use:

a=np.arange(0,1000,1)