1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

NUMPY Complete notes by Amar Sharma

Import numpy library

In [1]: import numpy as np #import library

array creation

In [2]: my_array1 = np.array([1,5,3], np.int8) #The 8-bit integer has a range from -128 to 127 [-2^(7) to 2^

In [3]: my_array1

Out[3]: array([1, 5, 3], dtype=int8)

In [4]: my_array1[0] #indexing operation

Out[4]: 1

In [5]: my_array2 = np.array([[1,5,3]])

In [6]: my_array2[0,2]

Out[6]: 3

In [7]: my_array2[0]

Out[7]: array([1, 5, 3])

In [8]: my_array2[0,1]

Out[8]: 5

Array shape

In [9]: my_array2.shape #This indicates that the array is 2-dimensional with a shape of 1 row and 3 columns

Out[9]: (1, 3)

In [10]: my_array1.shape #This indicates that the array is 1-dimensional with a length of 3.

Out[10]: (3,)

In [11]: my_array3 = np.array([[[1,5,3]]])

In [12]: my_array3.shape #This indicates that the array is 3-dimensional with a shape of 1 depth, 1 row, and

Out[12]: (1, 1, 3)

In [13]: my_array4 = np.array([[[[1,5,3]]]])

my_array4.shape #4-dimensional with a shape of 1 depth, 1 row, 1 column, and 3 elements along the

Out[13]: (1, 1, 1, 3)

Array data type

In [14]: my_array1.dtype #Data rype

Out[14]: dtype('int8')

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 1/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

In [15]: my_array2

Out[15]: array([[1, 5, 3]])

Change on array

In [16]: my_array2[0,1] = 45 #changing value

In [17]: my_array2

Out[17]: array([[ 1, 45, 3]])

Array creation : conversion from other python structures

In [18]: listarray = np.array([[1,2,3],[5,8,5],[0,3,5]])

In [19]: listarray

Out[19]: array([[1, 2, 3],

[5, 8, 5],
[0, 3, 5]])

In [20]: listarray.shape

Out[20]: (3, 3)

Check array size

In [21]: listarray.size

Out[21]: 9

In [22]: listarray.dtype

Out[22]: dtype('int32')

Make array by zeros

In [23]: zeros = np.zeros((2,5))

In [24]: zeros

Out[24]: array([[0., 0., 0., 0., 0.],

[0., 0., 0., 0., 0.]])

Make array by arange

In [25]: rng = np.arange(15)

In [26]: rng

Out[26]: array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14])

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 2/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

divide equally spaced

In [27]: lspace = np.linspace(1,10,4) #4 elements from 1 to 10 with eaually spaced

lspace

Out[27]: array([ 1., 4., 7., 10.])

Make identity matrix

In [28]: ide = np.identity(45) #formed identity array of row column 45

In [29]: ide

Out[29]: array([[1., 0., 0., ..., 0., 0., 0.],

[0., 1., 0., ..., 0., 0., 0.],
[0., 0., 1., ..., 0., 0., 0.],
...,
[0., 0., 0., ..., 1., 0., 0.],
[0., 0., 0., ..., 0., 1., 0.],
[0., 0., 0., ..., 0., 0., 1.]])

In [30]: ide.shape

Out[30]: (45, 45)

In [31]: arr = np.arange(99)

arr

Out[31]: array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,

17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,
51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67,
68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,
85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98])

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 3/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

Re shape array
In [32]: arr.reshape(33,3) #Reshaping

Out[32]: array([[ 0, 1, 2],

[ 3, 4, 5],
[ 6, 7, 8],
[ 9, 10, 11],
[12, 13, 14],
[15, 16, 17],
[18, 19, 20],
[21, 22, 23],
[24, 25, 26],
[27, 28, 29],
[30, 31, 32],
[33, 34, 35],
[36, 37, 38],
[39, 40, 41],
[42, 43, 44],
[45, 46, 47],
[48, 49, 50],
[51, 52, 53],
[54, 55, 56],
[57, 58, 59],
[60, 61, 62],
[63, 64, 65],
[66, 67, 68],
[69, 70, 71],
[72, 73, 74],
[75, 76, 77],
[78, 79, 80],
[81, 82, 83],
[84, 85, 86],
[87, 88, 89],
[90, 91, 92],
[93, 94, 95],
[96, 97, 98]])

In [33]: arr.ravel() #back to 1D array

Out[33]: array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,

17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,
51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67,
68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,
85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98])

In [34]: arr.shape

Out[34]: (99,)

Numpy Array Axis

In [35]: ar = np.array([[1,2,3],[4,5,6],[7,1,0]])

In [36]: ar

Out[36]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

axis=0: Operations along the rows (sum along each column in a 2D array).
axis=1: Operations along the columns (sum along each row in a 2D array).

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 4/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

In [37]: ar.sum(axis=0) #1+4+7,2+5+1,3+6+0

Out[37]: array([12, 8, 9])

In [38]: ar.sum(axis=1) #1+2+3,4+5+6,7+1+0

Out[38]: array([ 6, 15, 8])

In [39]: ar.T #Transpose

Out[39]: array([[1, 4, 7],

[2, 5, 1],
[3, 6, 0]])

In [40]: ar

Out[40]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

Check dimention
In [41]: ar.ndim #dimension of array

Out[41]: 2

Make it Iterable
In [42]: ar.flat #made it Iterable

Out[42]: <numpy.flatiter at 0x1adf961bc80>

In [43]: for i in ar.flat:

print(i)

1
2
3
4
5
6
7
1
0

Check memory captured

In [44]: ar.nbytes #bytes consume by that array

Out[44]: 36

In [45]: arr = np.array([8,55,2,75])

index of smallest and largest

In [46]: arr.argmin() #returns the index (or indices) corresponding to the smallest element

Out[46]: 2

In [47]: arr.argmax() #returns the index (or indices) corresponding to the largest element

Out[47]: 3

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 5/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

In [48]: ar

Out[48]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

In [49]: ar.argmin() #1,2,3,4,5,6,7,1,0

Out[49]: 8

In [50]: ar.argmax()

Out[50]: 6

In [51]: ar.argmin(axis=1)

Out[51]: array([0, 0, 2], dtype=int64)

In [52]: ar

Out[52]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

Use where
In [53]: np.where(ar>5) # use of where

Out[53]: (array([1, 2], dtype=int64), array([2, 0], dtype=int64))

In [54]: ar[1,2]

Out[54]: 6

In [55]: ar[2,0]

Out[55]: 7

In [56]: ar

Out[56]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

Count non zero

In [57]: np.count_nonzero(ar)

Out[57]: 8

In [58]: ar

Out[58]: array([[1, 2, 3],

[4, 5, 6],
[7, 1, 0]])

In [59]: np.nonzero(ar) #0,0 & 0,1 & 0,2 & 1,0 & 1,1 & 1,2 & 2,0 & 2,1 are non zero, except ar[2,2]

Out[59]: (array([0, 0, 0, 1, 1, 1, 2, 2], dtype=int64),

array([0, 1, 2, 0, 1, 2, 0, 1], dtype=int64))

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 6/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

check how much memomy normal python array is consuming

In [60]: import sys

In [61]: py_array = [5,2,3,4]

In [62]: np_array = np.array(py_array)

In [63]: sys.getsizeof(1)* len(py_array) #check how much memomy normal python array is consuming

Out[63]: 112

In [64]: np_array.itemsize * np_array.size #how much numpy array is consuming

Out[64]: 16

Operations in matrix

In [65]: aa1 = np.array([[1,2,3],[4,5,6],[7,8,9]])

aa2 = np.array([[5,4,6],[2,5,8],[2,4,6]])

In [66]: aa1

Out[66]: array([[1, 2, 3],

[4, 5, 6],
[7, 8, 9]])

In [67]: aa2

Out[67]: array([[5, 4, 6],

[2, 5, 8],
[2, 4, 6]])

In [68]: aa1+aa2

Out[68]: array([[ 6, 6, 9],

[ 6, 10, 14],
[ 9, 12, 15]])

In [69]: aa1*aa2 # not possible in python list

Out[69]: array([[ 5, 8, 18],

[ 8, 25, 48],
[14, 32, 54]])

In [70]: np.sqrt(aa1)

Out[70]: array([[1. , 1.41421356, 1.73205081],

[2. , 2.23606798, 2.44948974],
[2.64575131, 2.82842712, 3. ]])

In [71]: np.sqrt(aa2)

Out[71]: array([[2.23606798, 2. , 2.44948974],

[1.41421356, 2.23606798, 2.82842712],
[1.41421356, 2. , 2.44948974]])

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 7/8

1/7/24, 8:42 PM NUMPY Notes by Amar Sharma - Jupyter Notebook

Sort & argsort

In [72]: aa3 = np.array([[8,2,3],[4,6,6],[7,4,9]])

aa3

Out[72]: array([[8, 2, 3],

[4, 6, 6],
[7, 4, 9]])

In [73]: np.sort(aa3, axis = 1) #the sorting operation is performed along the column for each rows.

Out[73]: array([[2, 3, 8],

[4, 6, 6],
[4, 7, 9]])

In [74]: np.sort(aa3, axis = 0) #the sorting operation is performed along the rows for each column.

Out[74]: array([[4, 2, 3],

[7, 4, 6],
[8, 6, 9]])

In [75]: aa3

Out[75]: array([[8, 2, 3],

[4, 6, 6],
[7, 4, 9]])

In [76]: np.argsort(aa3, axis = 1) #sort of indices

Out[76]: array([[1, 2, 0],

[0, 1, 2],
[1, 0, 2]], dtype=int64)

In [77]: np.argsort(aa3, axis = 0)

Out[77]: array([[1, 0, 0],

[2, 2, 1],
[0, 1, 2]], dtype=int64)

All methods and attributes check it from here

https://docs.scipy.org/doc/numpy-1.6.0/reference/generated/numpy.ndarray.html (https://docs.scipy.org/doc/numpy-
1.6.0/reference/generated/numpy.ndarray.html)

127.0.0.1:8889/notebooks/NUMPY Notes by Amar Sharma.ipynb#array-creation 8/8