NumPy’s sum() function is extremely useful for summing all elements of a given array in Python. In this article, we’ll be going over how to utilize this function and how to quickly use this to advance your code’s functionality.
Let’s go over how to use these functions and the benefits of using this function rather than iteration summation. Let’s first write the underlying algorithm to do simple summation ourselves. This will take the form of a function as follows:
Python3
# Let's define our function# Parameters: Input Arraydef sum(array): # Set variable for our final answer sum = 0 # Parse through our array for i in array: # Continuously add current element # to final sum sum += i # Return our sum return sum# Create a test arraytestArray = [1, 3, 34, 92, 29, 48, 20.3]# Test our functionprint('The sum of your numbers is ' + str(sum(testArray))) |
Output:
The sum of your numbers is 227.3
Now that we’ve seen how many lines it takes to write just this simple summation function, let’s test out NumPy’s sum() function to see how it compares.
Python3
# Import NumPy Libraryimport numpy as np# Let's begin with an example array# Initialize our arrayarray = [1, 4, 2.5, 3, 7.4, 8]# Utilize the sum() functionprint('The sum of these numbers is ' + str(np.sum(array))) |
Output:
The sum of these numbers is 25.9
Let’s see some more examples for understanding the usage of this function. One thing to note before going any further is that if the sum() function is called with a two-dimensional array, the sum() function will return the sum of all elements in that array.
Example 1:
In this example, we’ll just do another simple summation of a one-dimensional array, just as we have seen before. It’ll get more exciting later on though!
Python3
# Import NumPyimport numpy as np# Initialize our test arrayarray = [0.5, 1.5]# Call our sum() functionprint(np.sum(array)) |
Output:
2.0
Example 2:
In this example, we’ll go over summing a two-dimensional array. Still pretty basic.
Python3
# Import NumPy Libraryimport numpy as np# Initialize our arrayarray = [[1, 3.4, 4.5], [3.45, 5.6, 9.8], [4.5, 5, 6.3]]# Call our sum() functionprint(np.sum(array)) |
Output:
43.55
Example 3:
Let’s try to use the optional parameters to try and manipulate our output. Let’s try out the axis parameter.
Python3
# Import NumPyimport numpy as np# Initialize our arrayarray = [[0, 1], [0, 5]]# Let's say we want to sum each sub array# Sums will be returned separately in array# format Call sum() functionprint(np.sum(array, axis = 1)) |
Output:
[1 5]
Example 4:
Let’s try using the initial parameter to initialize our sum value. Essentially, the initial value gets added to the sum of the elements of the array. This is useful in certain problems utilizing counters.
Python3
# Import NumPy Libraryimport numpy as np# Initialize our arrayarray = [1.5, 3, 5.6]# Call our sum() function# initial = 3print(np.sum(array, initial = 3)) |
Output:
13.1
Example 5:
If the accumulator for the sum is too small, then we get issues as far as overflow. Though this won’t directly throw an error, we will get issues with the results being unreliable. Let’s take a look at an example. We’ll use NumPy’s ones() function to automatically create an array of a given length that is filled with ones.
Python3
# Import NumPy Libraryimport numpy as np# Initialize our arrayarray = np.ones(250, dtype=np.int8)# Call our sum() function with a specified# accumulator typeprint(np.sum(array, dtype=np.int8))# Expected Output: 250 |
Output:
-6
