Python: Can a function return an array and a variable?

Is there a simple way to get a function to return a np.array and a variable?

eg:

my_array = np.zeros(3) my_variable = 0. def my_function(): my_array = np.array([1.,2.,3.]) my_variable = 99. return my_array,my_variable my_function() 

so that the values calculated in the function can be used later in the code? The above ignores the values calculated in the function.

I tried returning a tuple {my_array, my_variable} but got the unhashable type message for np.array

DN

4 Answers

Your function is correct. When you write return my_array,my_variable, your function is actually returning a tuple (my_array, my_variable).

You can first assign the return value of my_function() to a variable, which would be this tuple I describe:

result = my_function() 

Next, since you know how many items are in the tuple ahead of time, you can unpack the tuple into two distinct values:

result_array, result_variable = result 

Or you can do it in one line:

result_array, result_variable = my_function() 

Other notes related to returning tuples, and tuple unpacking:

I sometimes keep the two steps separate, if my function can return None in a non-exceptional failure or empty case:

result = my_function() if result == None: print 'No results' return a,b = result # ... 

Instead of unpacking, alternatively you can access specified items from the tuple, using their index:

result = my_function() result_array = result[0] result_variable = result[1] 

If for whatever reason you have a 1-item tuple:

return (my_variable,) 

You can unpack it with the same (slightly awkward) one-comma syntax:

my_variable, = my_function() 

It's not ignoring the values returned, you aren't assigning them to variables.

my_array, my_variable = my_function() 

easy answer

my_array, my_variable = my_function() 

After the definition of my_function, use my_function = np.vectorize(my_function).
For example,

def jinc(x): if x == 0.0: return 1 return 2*j1(x)/x jinc = np.vectorize(jinc)