Python: Using two function returns simultaneously

Question

I'm searching for a (probably obvious) way of reducing the time my code takes to run.

At the moment my code looks like this:

import numpy as np
from matplotlib import pyplot as plt

l = 4000000

def my_func(n):
    m = 3 * n
    z = n**2
    return(m,z)

am = []
az = []

for i in range(l): # question is referring to this loop
   am.append(my_func(i)[0])
   az.append(my_func(i)[1])

x = range(l)
plt.plot(x, am)
plt.plot(x, az)
plt.show()

In the i-loop my_func always runs twice and discards one of the two returned values, which sounds super inefficient to me. So how can I fill my am array with the first and my az array with the second values my_func returns, without running it twice for every i and only using half the returns every time?

How much of the time is spent on building the list and how much on plotting them? — Kelly Bundy
– Kelly Bundy, Commented May 2, 2022 at 16:57

Josh Clark · Accepted Answer · 2022-05-02 14:06:40Z

3

You can unpack the tuple returned from this function:

for i in range(l):
    m, z = my_func(i)
    am.append(m)
    az.append(z)

This blog post nicely explains this concept in Python.

edited May 2, 2022 at 14:06

answered May 2, 2022 at 14:00

Josh Clark

1,0222 gold badges10 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

timgeb · Accepted Answer · 2022-05-02 13:59:33Z

2

The obvious solution is to assign a variable to the return value of your function.

result = my_func(i)
am.append(result[0])
az.append(result[1])

Better yet, just unpack the result.

m, z = my_func(i)
am.append(m)
az.append(z)

answered May 2, 2022 at 13:59

timgeb

79.2k20 gold badges129 silver badges150 bronze badges

Comments

Schnitte · Accepted Answer · 2022-05-02 14:00:12Z

1

I would simply do without a separate function altogether and append the two arrays as

am.append(3*i)
az.append(i**2)

answered May 2, 2022 at 14:00

Schnitte

1,2276 silver badges16 bronze badges

Comments

Abhyuday Vaish · Accepted Answer · 2022-05-02 14:01:24Z

1

You can unpack the values:

m, z = my_func(i)
am.append(m)
az.append(z)

answered May 2, 2022 at 14:01

Abhyuday Vaish

2,3775 gold badges14 silver badges28 bronze badges

Comments

user2736738 · Accepted Answer · 2022-05-02 14:11:15Z

1

   am.append(my_func(i)[0])
   az.append(my_func(i)[1])

You can change to

   tup = my_func(i)
   am.append(tup[0])
   az.append(tup[1])

Then again, you don't need to call the function over here. You can easily do one thing.

x = np.arange(l)
plt.plot(x, 3*x)
plt.plot(x, x*x)
plt.show()

Timing

import numpy as np
from matplotlib import pyplot as plt
import time
start_time = time.time()

l = 4000000

def my_func(n):
    m = 3 * n
    z = n**2
    return(m,z)

am = []
az = []

for i in range(l): # question is referring to this loop
   am.append(my_func(i)[0])
   az.append(my_func(i)[1])

x = range(l)
plt.plot(x, am)
plt.plot(x, az)
plt.show()
print("--- %s seconds ---" % (time.time() - start_time))

Output:

Code-2

import numpy as np
from matplotlib import pyplot as plt
import time
start_time = time.time()

l = 4000000

# def my_func(n):
#     m = 3 * n
#     z = n**2
#     return(m,z)

# am = []
# az = []

# for i in range(l): # question is referring to this loop
#    am.append(my_func(i)[0])
#    az.append(my_func(i)[1])

x = np.arange(l)
plt.plot(x, 3*x)
plt.plot(x, x*x)
plt.show()
print("--- %s seconds ---" % (time.time() - start_time))

Output

Instead of using for loop, you can simply use vectorization to speed up the operation, as shown in the pictures( 5-6 times faster roughly). Also try to not call the same function multiple times to reuse the same value. Better to call it once and store the result.

edited May 2, 2022 at 14:11

answered May 2, 2022 at 14:01

user2736738

31k5 gold badges45 silver badges57 bronze badges

3 Comments

user2736738 Over a year ago

@MechanicPig Ah yes. I was adding something else. Thanks for mentioning.

Tim Over a year ago

Thanks for your answer! Although you're right timing-wise my_func is just a dummy function I put in order to ask the question, the actual function does a lot more, but in the end just returns two values, hence the dummy function, which I can't eliminate completely as you did in the last version. Unpacking worked, as many before have suggested tho!

user2736738 Over a year ago

@Tim.: Most likely not possible to answer in this scope, but usually you can almost always avoid for loop. That is the message. Even if your function is doing a lot of thing, you can (almost) always vectorize it.

Phoenixo · Accepted Answer · 2022-05-02 13:59:26Z

0

Just store the value in a variable :

results = my_func(i)
am.append(results[0])
az.append(results[1])

answered May 2, 2022 at 13:59

Phoenixo

2,1631 gold badge9 silver badges14 bronze badges

Comments

david · Accepted Answer · 2022-05-02 14:01:06Z

0

I would use np.vectorize

am, az = np.vectorize(my_func)(range(l))

answered May 2, 2022 at 14:01

david

1,5931 gold badge14 silver badges24 bronze badges

Comments

aquaplane · Accepted Answer · 2022-05-02 14:07:17Z

0

Do you mean something like the below? Save the returned tuple and append to the lists.

import numpy as np
from matplotlib import pyplot as plt

l = 4000000

def my_func(n):
    m = 3 * n
    z = n**2
    return(m,z)

am = []
az = []

for i in range(l): # question is referring to this loop
    m, z = my_func(i)
    am.append(m)
    az.append(z)

x = range(l)
plt.plot(x, am)
plt.plot(x, az)
plt.show()

answered May 2, 2022 at 14:07

aquaplane

1274 bronze badges

Collectives™ on Stack Overflow

Python: Using two function returns simultaneously

8 Answers 8

Comments

Comments

Comments

Comments

Timing

Output:

Code-2

Output

3 Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

8 Answers 8

Comments

Comments

Comments

Comments

Timing

Output:

Code-2

Output

3 Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related