What are universal functions for n-dimensional arrays in Python?

In this article, we will explain universal functions in python and how they are applied to n-dimensional arrays.

A universal function (or ufunc) is a function that operates on ndarrays element by element, providing array broadcasting, type casting, and a variety of other standard features. A ufunc, in other words, is a "vectorized" wrapper around a function that accepts a fixed number of scalar inputs and returns a fixed number of scalar outputs.

They are simple mathematical functions in the NumPy library. Numpy includes a number of universal functions that support a wide range of operations.

These functions contain standard trigonometric functions, arithmetic operations, complex number handling, statistical functions, and so forth. The following are some of the characteristics of universal functions −

• These functions work with ndarray (N-dimensional array), which is Numpy's array class.

• It provides quick array operations on elements.

• It provides a variety of functions such as array broadcasting, type casting, and so on.

• Numpy universal functions are objects in the numpy.ufunc class.

• Python functions can also be made universal by utilizing the frompyfunc library function.

• When the corresponding array arithmetic operator is applied, some ufuncs are called automatically. When two arrays are added element by element using the '+' operator, np.add() is called internally.

Trigonometric Functions

Since these trigonometric functions operate in radians, angles must be converted to radians by multiplying by pi/180. Only then can we refer to trigonometric functions. They accept an array of arguments as input.

Example

# importing numpy module with an alias name
import numpy as np

# creating an array of angles 
inputAngles = np.array([0, 30, 45, 60, 90, 180])

# converting input angles from degrees into radians
# using deg2rad() function
radians_angles = np.deg2rad(inputAngles)

# printing the sine values of input angles
print('Sine values of input array angles:')
sineValues = np.sin(radians_angles)
print(sineValues)
print()

# printing the inverse sine of sine values
print('Inverse Sine values of sine values:')
print(np.rad2deg(np.arcsin(sineValues)))
print()

# printing the hyperbolic sine of input angles
print('Hyperbolic Sine values of input angles:')
sineh_value = np.sinh(radians_angles)
print(sineh_value)
print() 

# printing the inverse sine hyperbolic 
print('Inverse Sine hyperbolic values of input angles:')
print(np.sin(sineh_value)) 
print()

# printing the hypotenuse of right angled triangle the 
triangle_base = 4
triangle_height = 3
print('The hypotenuse of right angled triangle = ', np.hypot(triangle_base, triangle_height))

Output

On executing, the above program will generate the following output −

Sine values of input array angles:
[0.00000000e+00 5.00000000e-01 7.07106781e-01 8.66025404e-01
 1.00000000e+00 1.22464680e-16]

Inverse Sine values of sine values:
[0.0000000e+00 3.0000000e+01 4.5000000e+01 6.0000000e+01 9.0000000e+01
 7.0167093e-15]

Hyperbolic Sine values of input angles:
[ 0.          0.54785347  0.86867096  1.24936705  2.3012989  11.54873936]

Inverse Sine hyperbolic values of input angles:
[ 0.          0.52085606  0.76347126  0.94878485  0.74483916 -0.85086591]

The hypotenuse of right angled triangle =  5.0

Here, numpy.array() function- returns an ndarray. The ndarray is an array object that satisfies the given requirements

Statistical Functions

The statistical functions calculates the mean, median, variance, minimum of array elements. Following are the functions it contains −

Example

# importing numpy module with an alias name
import numpy as np

# creating an array
inputArray = np.array([20, 45, 50.5, 60, 55.5, 10])

# printing the minimum value in an array
print('Minimum value in an input Array:')
print(np.amin(inputArray))

# printing the maximum value in an array
print('Maximum value in an input Array:')
print(np.amax(inputArray))

# printing the peak to peak value i.e 
# range of inputArray(maximum-minimum) values 
print('Range of inputArray(maximum-minimum) values:')
print(np.ptp(inputArray))

# printing the 50 percentile of input array 
print('50 percentile of input array: ')
print(np.percentile(inputArray, 50))

# printing the mean of input array 
print('Mean of an input array:')
print(np.mean(inputArray))

# printing the median of input array 
print('Median of an input array:')
print(np.median(inputArray))

# printing the standard deviation of input array 
print('Standard deviation of an input array:')
print(np.std(inputArray))

# printing the variance of input array 
print('Variance of an input array: ')
print(np.var(inputArray))

# printing the average of input array 
print('Average of an input array: ')
print(np.average(inputArray))

Output

On executing, the above program will generate the following output −

Minimum value in an input Array:
10.0
Maximum value in an input Array:
60.0
Range of inputArray(maximum-minimum) values:
50.0
50 percentile of input array: 
47.75
Mean of an input array:
40.166666666666664
Median of an input array:
47.75
Standard deviation of an input array:
18.598088312751095
Variance of an input array: 
345.88888888888886
Average of an input array: 
40.166666666666664

Bit-twiddling Functions

Bit-twiddling functions take integer numbers as input arguments and perform bitwise operations on their binary representations. Following are the functions it contains −

Example

# importing numpy module with an alias name
import numpy as np

# creating two arrays 
inputArray_1 = np.array([3, 2, 7, 5])
inputArray_2 = np.array([5, 1, 6, 4])

# printing the bitwise "and" operation of two arrays
print('Bitwise "and" operation of two arrays:')
print(np.bitwise_and(inputArray_1, inputArray_2))

# printing the bitwise "or" operation of two arrays
print('Bitwise "or" operation of two arrays:')
print(np.bitwise_or(inputArray_1, inputArray_2))

# printing the bitwise "xor" operation of two arrays
print('Bitwise "xor" operation of two arrays:')
print(np.bitwise_xor(inputArray_1, inputArray_2))

# printing the inversion/not of inputArray_1
print('The inversion/not of inputArray_1:')
print(np.invert(inputArray_1))

# left shifting inputArray_1 elements by 1 bit
print('left shifting inputArray_1 elements by 1 bit')
print(np.left_shift(inputArray_1, 1))

# right shifting inputArray_1 elements by 2 bits
print('right shifting inputArray_1 elements by 2 bits:')
print(np.right_shift(inputArray_1, 2))

Output

On executing, the above program will generate the following output −

Bitwise "and" operation of two arrays:
[1 0 6 4]
Bitwise "or" operation of two arrays:[7 3 7 5]
Bitwise "xor" operation of two arrays:
[6 3 1 1]
The inversion/not of inputArray_1:
[-4 -3 -8 -6]
left shifting inputArray_1 elements by 1 bit
[ 6  4 14 10]
right shifting inputArray_1 elements by 2 bits:
[0 0 1 1]

Conclusion

In this article, we learned about all of the universal functions in Python's n-dimensional arrays, along with examples.

Vikram Chiluka

Updated on: 20-Oct-2022

682 Views

Kickstart Your Career

Get certified by completing the course

Get Started