The difference between axes and axis is, axes are a plural form, and an axis is a singular form. This means, in this function, we can mention axes on which we want to perform our operation.
Numpy apply_over_axes
The np.apply_over_axes() is a built-in Numpy library function used to perform any function over multiple axes in an nd-array repeatedly. The apply_over_axes() method applies the function frequently over multiple axes in an array.
Syntax
numpy.apply_along_axis(1d_func, array, axes, *args, **kwargs)
Parameters
The NumPy apply_over_axes() function have 5 parameters:
- 1d_func: This parameter is required to operate the 1D array. It can be applied in the 1D slices of an input array and along the particular axis.
- array: This is an array on which we want to work.
- axis: These are the required axes along which we want to operate.
- *args: This is the additional argument to 1D function (1d_func).
- **kwargs: Additional named argument to 1D function (1d_func).
Return Value
The NumPy apply_over_axes() function returns an output array. The shape of the output array can differ depending on whether the function (1d_func) changes the shape of its output as per its input.
Calculate the sum of 2D-array
See the following code.
#Importing numpy
import numpy as np
#We will create a 2D array
#Of shape 4x3
arr = np.array([(1, 10, 3), (14, 5, 6), (7, 8, 19), (50, 51, 52)])
#Printing the array
print("The array is: ")
print(arr)
#Declaring axes
axes = [1, -1]
print("Sum of array elements are:")
print(np.apply_over_axes(np.sum, arr, axes))
Output
The array is: [[ 1 10 3] [14 5 6] [ 7 8 19] [50 51 52]] Sum of array elements are: [[ 14] [ 25] [ 34] [153]]
Explanation
In this example, we have declared one 2D array of size 4×3, we have printed it and its shape.
We have declared axes=[1,-1] and called apply_over_axes() to calculate the sum of its elements. We can see that we have got a column of 4rows; each column has the sum of its row elements.
Calculate the sum of the 3D array
See the following code.
#Importing numpy
import numpy as np
#We will create a 2D array
#Of shape 4x3
arr = np.arange(12).reshape(2, 2, 3)
#Printing the array
print("The array is: ")
print(arr)
#Declaring axes
axes = [0, 2]
print("Sum of array elements are:")
print(np.apply_over_axes(np.sum, arr, axes))
Output
The array is: [[[ 0 1 2] [ 3 4 5]] [[ 6 7 8] [ 9 10 11]]] Sum of array elements are: [[[24] [42]]]
Explanation
In this example, we have first declared one 3D array of shapes (2x2x3), and we have printed that array.
We have declared axes with values [0,2], calculating the sum over axes 0 and 2.
We can see that we have the desired output when we have called the function.
