Topics to Cover

Class 1: (Linear Array)

● What is an Array? Difference with other data structures.
● A sequence of data with the same type
● Why need an Array if a list can already do most of the tasks?
● NumPy Array (zeros)
● indexing (same as list and string)Why size can not be changed?
● benefits of array (random access)
● selection/bubble sort

Class 2: (Multidimensional Array)

● Cubic array example??
● Real life example (Matrix-2D Array), Higher Dimensional-3D (seat placement in a classroom)
● Matrix-matrix multiplication,
● vector-vector, matrix-vector, matrix row switch (For student’s self study)
1
 Linear Array

Array

● Sequence of data
● Single data type
● Fixed length
● Indexed
● Fixed-dimensional

2
 Linear Array

Difference between Array and List

List Array

Can store different types of data A single type of data

Dynamic and mutable Fixed length and mutable

Comparatively slower Fast because has contiguous memory

allocation

As data is stored in discontinuous As memory is contagious internal

memory space the internal memory memory allocation is simple. One data
allocation is complex. is stored after another.

It is easy to append elements to a It is easy perform arithmetic

list. operations on arrays.

3
 Linear Array

Introduction to NumPy

Array is not native to python, so we use a library called numpy

● Library import
import numpy as np

● Array creation
arr = np.zeros(size, dtype = datatype) arr = np.zeros(5, dtype = int)

(dtype is float by default)

● Array length
○ .size : attribute, returns the total number of
elements irrespective of dimensionality.
print(len(arr)) 5
○ len() : function, returns the total number of print(arr.size) 5
elements along the first dimension
4
 Linear Array

Array Operations

● Read and Write

var = arr[2] #read
arr[3] = 10 #write

● Array traversal
idx = 0 for idx in range(arr.size):
while idx < arr.size: print(arr[idx])
print(arr[idx])
idx += 1

5
 Linear Array

Array Operations

● Array creation from user input length = int(input("Enter array length: "))
arr = np.zeros(length, dtype = int)
○ Can be created from a single for i in range(length):
input string as well value = int(input("Enter a value: "))
arr[i] = value
print(arr)

● Array Manipulation
○ Updating values
○ Operation using/on indices
○ Array size increment => Creating a new array with increase size and copied values

6
 Linear Array

Benefits of Array

● Memory Efficient
● Random Access
● Faster Performance
● Better suited for Mathematical Operations (i.e. vectorized/matrix
operations)

7
 Linear Array
Bubble Sort

1st Pass 3rd Pass

-2 45 0 11 -9 -2 0 -9 11 45 arr = np.array([-2, 45, 0, 11, -9])

for i in range(len(arr)):
-2 0 45 11 -9 -2 -9 0 11 45
for j in range(0, len(arr) - i - 1):
-2 0 11 45 -9 if arr[j] > arr[j + 1]:
4th Pass
temp = arr[j]
-2 0 11 -9 45
-2 -9 0 11 45 arr[j] = arr[j+1]
arr[j+1] = temp
2nd Pass -9 -2 0 11 45
print(arr)
-2 0 11 -9 45

-2 0 -9 11 45

8
 Multi-Dimensional Array

Multi Dimensional Array

We can start by discussing from Vector to Matrix then Tensor.

9
 Multi-Dimensional Array

1D, 2D and 3D array

10
 Multi-Dimensional Array

RGB image

A RGB image is actually a

3D array containing three
2D Matrices for Red, Green
and Blue values.

11
 Multi-Dimensional Array

Matrix/2D Array Examples

● Seating Arrangement in a classroom.

● Linear Algebra Usage (Matrix representation).
● Displays are 2D array of pixels.

2D MATRIX

Dimensions of matrices
Rows x Columns

12
 Multi-Dimensional Array

2D Matrix in Python using Numpy

import numpy as np
arr2D = np.zeros(shape=(row,col), dtype = datatype)

arr2D = np.zeros(shape=(3,4), dtype = int) Output

print(arr2D) [[0 0 0 0]
[0 0 0 0]
[0 0 0 0]]

arr2D = np.zeros(shape=(3,4), dtype = float) Output

print(arr2D) [[0. 0. 0. 0.]
[0. 0. 0. 0.]
[0. 0. 0. 0.]]

arr2D = np.zeros(shape=(3,4), dtype = str) Output

print(arr2D) [['' '' '' '']
['' '' '' '']
['' '' '' '']]

13
 Multi-Dimensional Array

2D Matrix from Python List using Numpy

arr2D=np.array( [[1,2,4,6],[5,7,9,8]] ) Output

print(arr2D) [[1 2 4 6]
print(arr2D.dtype) [5 7 9 8]]
int64

arr2D=np.array( [[1,2,4.2,6],[5,7,9,8]] ) Output

print(arr2D) [[1. 2. 4.2 6. ]
print(arr2D.dtype) [5. 7. 9. 8. ]]
float64

14
 Multi-Dimensional Array

Indexing in 2D Matrix a

● a[row_num] will give us a

single linear array.

● a[row_num][col_num] can
access individual cell.

a = np.array( [[1,2,4,6],[5,6,7,8],[9,0,1,2]] ) Output

print(a) [[1 2 4 6]
print("First Row", a[0]) [5 6 7 8]
print("Second Row", a[1])
[9 0 1 2] ]
print("Third Cell", a[0][2])
First Row [1 2 4 6]
Second Row [5 6 7 8]
Third Cell 4 15
 Multi-Dimensional Array

Accessing the shape and size of a 2D Matrix

arr2D = np.zeros(shape=(3,4), dtype = int) Output

print(arr2D) [[0 0 0 0]
row, col = arr2D.shape[0], arr2D.shape[1] [0 0 0 0]
total_cells = arr2D.size [0 0 0 0]]
print(“Row Amount”, row) Row Amount 3
print(“Column Amount”, col) Column Amount 4
print(“Total cells”, total_cells)
Total cells 12

16
 Multi-Dimensional Array

Iteration of 2D Matrix

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

r_len , c_len = arr.shape[0], arr.shape[1] 1 2 4 6
#The Outer loop iterates rows 5 7 9 8
for r in range(r_len):
#The inner loop iterates column of each row
for c in range(c_len):
print(arr[r][c],end=' ')
print()

17
 Multi-Dimensional Array

Scalar & Matrix Multiplication

A=np.array( [[0,2,3],[1,1,0]] ) Output

print("Before",A) Before:
r_len , c_len = A.shape[0], A.shape[1] [[0 2 3]
#The outer loop iterates rows [1 1 0]]
for r in range(r_len): After:
#The inner loop iterates column of each row [[0 4 6]
[2 2 0]]
for c in range(c_len):
A[r][c] = 2*A[r][c]
print("After",A)

18
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4
n=3

= = 1x2 + 2x6 + 1x1

 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
 Multi-Dimensional Array

Matrix Multiplication Visual m=3

p=4

=
Matrix & Matrix Multiplication Code Multi-Dimensional Array

A=np.array( [[0,5],[1,2],[3,4]] ) OUTPUT

B=np.array( [[3,2,1],[1,2,4]] )
print("A(3,2):") A(3,2):
print(A) [[0 5]
print("B(2,3):") [1 2]
print(B) [3 4]]
Ar_len , Ac_len = A.shape[0], A.shape[1]
Br_len , Bc_len = B.shape[0], B.shape[1] B(2,3):
#Here, Bc_len==1 and Ac_len==Br_len [[3 2 1]
C=np.zeros( shape=(Ar_len,Bc_len), dtype=int) [1 2 4]]
Cr_len, Cc_len = C.shape[0], C.shape[1]
for i in range(Cr_len): A X B = C(3,3):
for j in range(Cc_len): [[ 5 10 20]
sum=0 [ 5 6 9]
for k in range(Br_len): [13 14 19]]
sum += A[i][k] * B[k][j]
C[i][j] = sum
print("A X B = C(3,3):")
print(C) 26
Rest of the parts won’t be covered in class.
Consider them as practice problems

27
 Multi-Dimensional Array

Matrix & Vector Multiplication Visual

28
Matrix & Vector Multiplication Code Multi-Dimensional Array

A=np.array( [[0,5],[1,2],[3,4]] ) :OUTPUT:

B=np.array( [[3],[2]] )
print("A(3,2):") A(3,2):
print(A) [[0 5]
print("B(2,1):") [1 2]
print(B) [3 4]]
Ar_len , Ac_len = A.shape[0], A.shape[1]
Br_len , Bc_len = B.shape[0], B.shape[1] B(2,1):
#Here, Bc_len==1 and Ac_len==Br_len [[3]
C=np.zeros( shape=(Ar_len,Bc_len), dtype=int) [2]]
Cr_len, Cc_len = C.shape[0], C.shape[1]
for i in range(Cr_len): C(3,1):
for j in range(Cc_len): [[10]
sum=0 [ 7]
for k in range(Br_len): [17]]
sum += A[i][k] * B[k][j]
C[i][j] = sum
#Here, j will be 0 since Cc_len==Bc_len==1
print("C(3,1):") 29
print(C)
 Multi-Dimensional Array

import numpy as np
A=np.array( [[3,4], A(3,2):
[4,5], [[3 4]
[10,11]] ) [4 5]
# The shape of A is (3,2) [10 11]]
B=np.array( [3,2] )
# The shape of A is (2,1) B(2,1):
# So, the shape of C will be (3,1) [3
C=np.zeros( shape=(3,1), dtype=int) 2]
for i in range(3):
sum=0 C(3,1):
for j in range(2): [[17]
sum += A[i][j] * B[j] [22]
[52]]
C[i][0]=sum
print("\nC(3,1):")
print(C)

30
 Multi-Dimensional Array

Vector & Vector Dot Product Code

v1 = np.array([2,3,1]) OUTPUT
v2 = np.array([5,-2,11])
Dot product: 15
dot_product = 0

for i in range(v1.size):
dot_product+= v1[i]*v2[i]

print(f'Dot product: {dot_product}')

31
 Multi-Dimensional Array

Matrix Row Switching v1

A = np.array([[4,3,2],[7,1,5],[-1,0,10]]) OUTPUT:
print('Before Switching')
Before Switching
print(A)
[[ 4 3 2]
#Switching row 2 and row 1 [ 7 1 5]
A[[1,2]] = A[[2,1]] [-1 0 10]]
print('After Switching')
After Switching
print(A)
[[ 4 3 2]
[-1 0 10]
[ 7 1 5]]

32
 Multi-Dimensional Array

Matrix Row Switching v2

A = np.array([[4,3,2],[7,1,5],[-1,0,10]]) OUTPUT:
print('Before Switching')
Before Switching
print(A)
[[ 4 3 2]
#Switching row 1 and row 2 [ 7 1 5]
#Backing up row 1 by creating a new array [-1 0 10]]
temp = np.array(A[1])
After Switching
#putting row 2 in row 1
[[ 4 3 2]
A[1] = A[2] [-1 0 10]
#putting row backup (row 2) in row 1 [ 7 1 5]]
A[2] = temp
print('After Switching')
print(A)

33
Matrix Row Switching v3 Multi-Dimensional Array

A = np.array([[4,3,2],[7,1,5],[-1,0,10]]) OUTPUT:
print('Before Switching')
Before Switching
print(A)
[[ 4 3 2]
row_len = A.shape[0] [ 7 1 5]
col_len = A.shape[1] [-1 0 10]]
#Switching Row index 1 and Row index 2
After Switching
for c in range(col_len):
[[ 4 3 2]
temp = A[1][c] [-1 0 10]
A[1][c] = A[2][c] [ 7 1 5]]
A[2][c] = temp
print("After Switching")
print(A)

34