I'm trying to write a function that accepts either a float OR an array of floats, and handles both of them using the same lines of code. For example, I want to return the float itself if it's a float, and the sum of the array of floats if it's an array. Something like this
def func(a):
return np.sum(a)
and have both func(1.2) return 1.2, and func(np.array([1.2,1.3,1.4]) return 3.9.
The usual way to make sure the input is a NumPy array is to use np.asarray():
import numpy as np
def func(a):
a = np.asarray(a)
return np.sum(a)
func(1.2)
# 1.2
func([1.2, 3.4])
# 4.6
func(np.array([1.2, 3.4]))
# 4.6
or, if you want to get len() of your array, make sure it is at least 1-dimensional, use np.atleast_1d():
def func(a):
a = np.atleast_1d(a)
return a.shape[0]
func(1.2)
# 1
func([1.2, 3.4])
# 2
func(np.array([1.2, 3.4]))
# 2
This already works, where's the problem?
import numpy as np
def func(a):
return np.sum(a)
print(func(np.array([1.2,2.3,3.2])))
print(func(1.2))
Output:
6.7
1.2
You can use argument flattening:
def func(*args):
# code to handle args
return sum(args)
Now the following have the same behaviour:
>>> func(3)
3
>>> func(3, 4, 5)
12
>>> func(*[3, 4, 5])
12
You can check if the input is a float, and then put it in a list before processing the sum:
def func(a):
if isinstance(a, float):
a = [a]
return np.sum(a)
int, or anything else that supports addition?