### Introduction

Numpy is python library that provides computation on large array elements and matrices. Numpy provides fast and efficient processing on n-dimensional arrays. Array elements in numpy are stored in continuous memory location so that the processing of element is fast and efficeint, unlike in list where the elements are stored in random memory location.

### Installation of numpy

``pip install numpy``

### Creation of numpy array

First of all, we've to import numpy module

``import numpy as np``

Here, numpy module is imported as np means numpy is shortened to np.

#### One dimensional array

Let's create a numpy array

``````arr = np.array([1,3,5,6,8,9])
print("Array Created: ", arr)``````

Output

``Array Created:  [1 3 5 6 8 9]``

Now, let's see the type of arr we just created

``print("Type of arr is: ", type(arr))``

Output

``Type of arr is:  <class 'numpy.ndarray'>``

We can create one dimensional array using arange() function

``````arr = np.arange(10)
print("Array created: ", arr)``````

Output

``Array created:  [0 1 2 3 4 5 6 7 8 9]``

arange function in numpy is similar to range() function in python

Using list we can create one dimensional array too

``````a = [1,2,3,4,5]
arr = np.array(a)
print("Array created from list: ", arr)``````

Output

``Array created from list:  [1 2 3 4 5]``

#### Two dimensional array

``````a  = [1,2,3,4,5]
b = [6,7,8,9,10]
arr = np.array([a, b])
print( arr)``````

Output

``````[[ 1  2  3  4  5]
[ 6  7  8  9 10]]``````

This is how we can create two dimensional array in python using list

``````arr = np.array([[1,2,3,4,5],[6,7,8,9,10]])
print( arr)``````

Output

``````[[ 1  2  3  4  5]
[ 6  7  8  9 10]]``````

This is another method where we can give elements directly to create two dimensional array

### Shape of an array

We can determine the shape of array using shape() function

``````arr = np.array([[1,2,3,4,5],[6,7,8,9,10]])
print("Shape of array is: ", arr.shape)``````

Output

``Shape of array is:  (2, 5)``

Shape function returns the no of rows and columns present in array. In this case there are 2 rows and 5 columns

#### Dimension of an array

``````arr = np.array([[1,2,3,4,5],[6,7,8,9,10]])
print("Dimension of array is: ", arr.ndim)``````

Output

``Dimension of array is:  2``

#### Let's see another examples

``````arr = np.array([1,2,3,4,5])
print("Dimension of array is: ", arr.ndim)``````

Output

``Dimension of array is:  1``

``````arr = np.array([[[1,2,3,4,5], [1,2,3,4,5]], [[1,2,3,4,5], [1,2,3,4,5]]])
print(arr)
print("Dimension of array is: ", arr.ndim)``````

Output

``````[[[1 2 3 4 5]
[1 2 3 4 5]]

[[1 2 3 4 5]
[1 2 3 4 5]]]
Dimension of array is:  3``````

### Size of an array

``````a = [1,3,5,6,8,9,6,8]
arr = np.array(a)
print("size of arr is: ", arr.size)``````

Output

``size of arr is:  8``

Size function returns the number of elements in an array

``````arr = np.array([[[1,2,3,4,5], [1,2,3,4,5]], [[1,2,3,4,5], [1,2,3,4,5]]])
print("size of array is: ", arr.size)``````

Output

``size of array is:  20``

### Accessing elements of an array

Array elements can be accessed using index same as list and tuple

``````a = [1,2,3,4,5]
arr = np.array(a)
print(arr)
print(arr)
print(arr)``````

Output

``````1
3
5``````

Elements can also be accessed using loop

``````for ele in arr:
print(ele)``````

Output

``````1
2
3
4
5``````

Let's see example of 2-D array

``````arr = np.array([[1,2,3,4,5], [23, 45, 67 ,98, 100]])
print(arr)
print(arr)
print(arr)
print(arr)``````

Output

``````2
1
98
23``````
##### Explanation

In above example, we passed two lists inside single list where first list index as 0 and another one index as 1 to make 2-dimensional array with 2 rows and 5 columns. Thus, if we have to access element from first row two indices must be passed i.e first index for selection of row and second index for selecting column. The first 'print statement' prints the value of first row and second column which is 2. Like as, third 'print statement' prints element from second row and fourth cloumn which is 98.

### Slicing array

###### Syntax
``array_name[start : end : step]``

Array slicing is similar to list slicing in python.

Let's see slicing of one dimensional array

``````a = [1,2,3,4,5]
arr = np.array(a)
print(arr[0:4])``````

Output

``array([1, 2, 3, 4])``

Index of array starts with 0 and end with one less than length of an array. Here in this example index starts from 0 and end with 4. We know the end index or upper bound is exclusive the above example retrieve the elements indexing 0 to 3.

``print(arr[-5:-1])``

Output

``array([1, 2, 3, 4])``

Indexing is assigned in negative as well. The negative indexing starts from -1 which is assigned to last element of array and ends with negative of length of an array.

##### Let's see slicing of two dimensional array
###### Syntax
``array_name[start_row : end_row: step_row, start_column : end_column: step_column]``

``````arr = np.array([[1,2,3],[4,5,6],[7,8,9]])
print(arr)``````

Output

``````[[1 2 3]
[4 5 6]
[7 8 9]]``````

Now, let's take elements from first two rows and last two columns

``print(arr[0:2, 1:])``

Output

``````[[2 3]
[5 6]]``````

Lets take out last element of array

``print(arr[2:, 2:])``

Output

``[]``

Lets take out last two element from second row

``print(arr[1:2, 1:])``

Output

``[[5 6]]``

### Reshaping an array

Using reshape() function, we can define new array from previously defined array

``````arr1 = np.array([1,2,3,4,5,6])
arr2 = arr1.reshape(3,2)
print(arr2)``````

Output

``````[[1 2]
[3 4]
[5 6]]``````

``````arr1 = np.array([1,2,3,4,5,6])
print("shape of arr1: ", arr1.shape, "\n")
arr2 = arr1.reshape(3,2)
print(arr2, "\n")
print("shape of arr2: ", arr2.shape)``````

Output

``````shape of arr1:  (6,)

[[1 2]
[3 4]
[5 6]]

shape of arr2:  (3, 2)``````

Here we've changed shape of arr1. All we have to care about during reshaping is that the no of elements must be same in both new and previous array

``````arr2 = arr1.reshape(2,2)
print(arr2)``````

Output

``````---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-70-a75024147a88> in <module>
1 arr1 = np.array([1,2,3,4,5,6])
----> 2 arr2 = arr1.reshape(2,2)
3 print(arr2)

ValueError: cannot reshape array of size 6 into shape (2,2)
``````

Here, we got ValueError because we are trying to reshaping array having 6 element to array with 4 element in it.

``````arr1 = np.array([[1,2,3,4,5,6], [7,8,9,10,11,12]])
arr2 = arr1.reshape(12,1)
print(arr2)``````

Output

``````[[ 1]
[ 2]
[ 3]
[ 4]
[ 5]
[ 6]
[ 7]
[ 8]
[ 9]


]``````

In both arrays, the number of element are same just shape is changed.

``````arr2 = arr1.reshape(4, 3)
print(arr2)``````

Output

``````[[ 1  2  3]
[ 4  5  6]
[ 7  8  9]
[10 11 12]]``````

### Appending & Inserting row and column in array

#### Using append() function

##### Row wise appending
###### Syntax
``np.append(previous_array, [array_to_be_add], axis =0)``

``````a = np.array([20,21,22])
np.append(arr2,[a],axis=0)``````

Output

``````array([[ 1,  2,  3],
[ 4,  5,  6],
[ 7,  8,  9],
[10, 11, 12],
[20, 21, 22]])``````

##### Column wise appending
###### Syntax
`` np.append(previous_array, array_to_be_add, axis =1)``

``````a = np.array([20,21,22,23])
b= a.reshape(4,-1)
np.append(arr2,b,axis=1)``````

Output

``````array([[ 1,  2,  3, 20],
[ 4,  5,  6, 21],
[ 7,  8,  9, 22],
[10, 11, 12, 23]])``````

#### Using insert()function

##### Row wise inserting
###### Syntax
``np.insert(previous_array, inserting_index, array_tobe_inserted, axis=0)``

``````arr = np.array([[1,2,3],[4,5,6],[7,8,9]])
inserting_arr = np.array([11,12,13])
print("Before insertion: ")
print(arr)
print("after insertion at index 2: ")
print(np.insert(arr, 2, inserting_arr, axis=0))``````

Output

``````Before insertion:
[[1 2 3]
[4 5 6]
[7 8 9]]
after insertion at index 2:
[[ 1  2  3]
[ 4  5  6]
[11 12 13]
[ 7  8  9]]``````

##### Column wise inserting
###### Syntax
``np.insert(previous_array, inserting_index, array_tobe_inserted, axis=1)``

``````arr = np.array([[1,2,3],[4,5,6],[7,8,9]])
inserting_arr = np.array([11,12,13])
print("Before insertion: ")
print(arr)
print("after insertion at index 1: ")
print(np.insert(arr, 1, inserting_arr, axis=1))``````

Output

``````Before insertion:
[[1 2 3]
[4 5 6]
[7 8 9]]
after insertion at index 1:
[[ 1 11  2  3]
[ 4 12  5  6]
[ 7 13  8  9]]``````

### Matrix generation using numpy

We can generate martices having elements all one and zero using ones() and zeros() function

``````zero_matrix = np.zeros([3,3], dtype=int)
print(zero_matrix)``````

Output

``````[[0 0 0]
[0 0 0]
[0 0 0]]``````

Let's take another example

``````ones_matrix = np.ones([4,5], dtype=float)
print(ones_matrix)       ``````

Output

``````[[1. 1. 1. 1. 1.]
[1. 1. 1. 1. 1.]
[1. 1. 1. 1. 1.]
[1. 1. 1. 1. 1.]]``````

#### Using random, we can generate metrices too

###### Syntax
``np.random.rand(size of array)``

``````arr = np.random.rand(3,5)
print(arr)
print(arr.shape)``````

Output

``````[[0.74206847 0.44733595 0.10237527 0.34372174 0.87838503]
[0.48042584 0.46966427 0.318181   0.88341896 0.46838867]
[0.02591508 0.58777176 0.07273747 0.80669176 0.69172011]]
(3, 5)``````

#### Element wise operation of array

##### One dimensional array
``````arr1 = np.array([1,2,3,4])
arr2 = np.array([4,3,2,1])
print("sum: ", arr1+arr2)
print("Difference: ", arr1-arr2)
print("Multiplication: ", arr1*arr2)
print("Division: ", arr1/arr2)``````

Output

``````sum:  [5 5 5 5]
Difference:  [-3 -1  1  3]
Multiplication:  [4 6 6 4]
Division:  [0.25       0.66666667 1.5        4.        ]``````

##### Two dimensional array
``````arr1 = np.array([[1,2,3,4], [4,3,2,1]])
arr2 = np.array([[4,3,2,1], [4,3,2,1]])
print("sum:\n ", arr1+arr2)
print("Difference:\n ", arr1-arr2)
print("Multiplication:\n ", arr1*arr2)
print("Division:\n ", arr1/arr2)``````

Output

``````sum:
[[5 5 5 5]
[8 6 4 2]]
Difference:
[[-3 -1  1  3]
[ 0  0  0  0]]
Multiplication:
[[ 4  6  6  4]
[16  9  4  1]]
Division:
[[0.25       0.66666667 1.5        4.        ]
[1.         1.         1.         1.        ]]``````

### Conclusion

Numpy is a powerful library that provides fast computation on large array of element and matrices. Elements are stored in continuous memory location so the processing of array element is faster than in list. Numpy has usecase in image processing too as opencv sees images as array of 1's and 0's. Numpy can be used with matplotlib library to plot various bar charts, histogram etc. So numpy is the powerful and most useable library in python.

Happy Learning :-)

### Reference

https://numpy.org/