Find the size of numpy Array

Numpy is often used along with packages such as SciPy and matplotlib in python. Additionally, arrays in Numpy are faster compared to lists in Python due to which this module is extensively used for complex mathematical algorithms and problems. It has various functions, and methods, to make our task easy and simple for matrix operations.

In Numpy arrays there are two types of array −

• Single Dimensional Array

• Multi-Dimensional Array

Single Dimensional Array (1D array)

The datatypes of the elements stored in the array are - strings, integers, boolean or floating points. These arrays are called single or 1 Dimensional arrays. The shape of the so-defined array will be linear and it is more often represented as a 1D array.

Multi Dimensional Array

These arrays can store elements of any datatype - strings, integers, array, boolean etc. They have arrays inside arrays which are also called nested arrays and have a non-linear structure.

Usually, an array is measured in two types −

• Size of the array

• Memory size of the array

The array size refers to the length of an array or the total number of independent data elements present in that particular array.

Whereas, memory size of an array refers to the space (usually in bytes for smaller arrays) that the array has occupied in one’s system hardware to store data present in that particular array.

Method 1: Using Itemsize() Function

The itemsize() function returns the memory space occupied by each element in the array. The value returned is considered to be in bytes.

Syntax

<array_object_name>.itemsize

Returns integer value which is considered as no. of bytes.

Example 1

In the following example, we will compute memory occupied for each element stored in 1 Dimensional numpy array using the itemsize() function. Along with the size, length and total memory space occupied, the array is also displayed.

Algorithm

• Step 1 − Define the array

• Step 2 − Declare an arra

• Step 3 − Print size

• Step 4 − Print itemsize

import numpy

#creation of a numpy array
arr = numpy.array([1,2,3,4,5])

#Displays length of the array
print(f"Length of the array: {numpy.size(arr)}") 

#Displays memory size in bytes
print(f"Space occupied by one element: {arr.itemsize} bytes")

#Total memory space occupied
print(f"Total memory occupied by the element: {arr.itemsize*numpy.size(arr)} bytes")

Output

Length of the array: 5
Space occupied by one element: 8 bytes
Total memory occupied by the element: 40 bytes

Example 2

In the following example, we compute the same information for a multi-dimensional numpy array.

import numpy

#creation of a numpy array
arr = numpy.array([['a','b','c','d'], [1,2,3,4,5]])

arr_2 = numpy.array([ [[1,2,3], 'Tutorials Point'], ['a', 'e', 'i', 'o', 'u']])

#Displays length of the array
print(f'Length of the 2D array: {numpy.size(arr)}')

#Displays memory size in bytes
print(f"Space occupied by one element: {arr.itemsize} bytes")

#Total memory space occupied
print(f"Total memory occupied by the element: {arr.itemsize*numpy.size(arr)} bytes")

print()

print(f"Length of the 3D array: {numpy.size(arr_2)}")
print(f"Space occupied by one element: {arr_2.itemsize} bytes")
print(f"Total memory occupied by the element: {arr_2.itemsize*numpy.size(arr_2)} bytes")

Output

Length of 2D array: 2
Space occupied by one element: 8 bytes
Total memory occupied by the element: 16 bytes

Length of 3D array: 2
Space occupied by one element: 8 bytes
Total memory occupied by the element: 16 bytes

Method 2: Using Nbytes() Function

nbytes() is a method in the numpy module in python. This function returns the total space (in bytes) occupied by the numpy array, unlike the itemsize function which returns the memory space occupied by only one element.

Syntax

<array_object_name>.nbytes

Returns integer value which is considered as total byte space occupied by the array.

Example 1

In this example, a 1D numpy array is created to explain the memory size occupied, using the nbytes function. The length of the array is also displayed alongwith.

import numpy

arr = numpy.array([1,2,3,4,5])

#Displays length of the array
print(f"Length of the array: {numpy.size(arr)}") 

#Total memory space occupied
print(f"Total memory occupied by the element: {arr.nbytes} bytes")

Output

Length of the array: 5
Total memory occupied by the element: 40 bytes

Example 2

In the following example, we compute the length and memory size of multi-dimensional arrays with the same data values used in the above example.

import numpy

arr = numpy.array([['a','b','c','d'], [1,2,3,4,5]])
arr_2 = numpy.array([ [[1,2,3], 'Tutorials Point'], ['a', 'e', 'i', 'o', 'u']])

#Displays length of the array
print(f'Length of the 2D array: {numpy.size(arr)}')

#Total memory space occupied
print(f"Total memory occupied by the element: {arr.nbytes} bytes")

print()

print(f"Length of the 3D array: {numpy.size(arr_2)}")
print(f"Total memory occupied by the element: {arr_2.nbytes} bytes")

Output

Length of 2D array: 2
Total memory occupied by the element: 16 bytes

Length of 3D array: 2
Total memory occupied by the element: 16 bytes

Conclusion

Numpy arrays consume less memory compared to other data structures which is why they’re used in handling large datasets and perform complex mathematical calculations. The itemsize function would compute the size of each element whereas the nbytes function computes the memory space occupied by the entire array.

They’re used in various fields such as machine learning, data analyst and computational physics, transformation of data into various forms and many more. It is also used in various numerical and theoretical calculations of the data.