1S5800 ASSIGNMENT 6

① h(x) = Xmod 10
,
Insert 7 3 ,
13 , 6 ,
3 ,
27 , 8]

modi
Inert 5 :

h(3) =
3

> 3
5
-

:
·
Insert 15 :

h(13) : 15 mod 10

= 3
"

e
Insert 6 :

h(6) : 6 mod 10
= 6

:
P

Insert 3 : I

h(3) =
3 mod 10
3 >
-
3

= 3 I

5 >
-
3 -> 15

↑
Insert 27 :

h(27) :
27 mod 10 I

=
7 3 >
-
3

"

i
P

Insert 8 : I

h (8) : I mod 10
3 >
-
3

I
= S
> 3 -> 15

F
-

O .
Universal
hashing :

Given :

1 = a
P
Size 20
:

mod p) modo 0=b

h(x) (ax + b) P
= <

p = 101

a)prove
I :

family of function is universal if

P(Ma b(X1 hap(x2)) +
the has =

,
20

consider two keys X1 and X2 . To

find the
probability that
they hash to the same

value , the collision condition is

(la . x, + b) modp) mod 20 =

(lax2 + b) modp) mod 20

b b
a (x1 -

x2) + -
= 0 (modp)
a (xi - *
2) =
0 / mod p)
we know that p 101 and since know it is prime we that
we can
say
:
,

a(X1 -X2) =
up

for some
integerm .
 we know that X , X2 and so X1-Xc 0.

is solution for mod p

... There
only one a .

a= 0 modp

There are in total (p-1) =

100 possible choices of a out of which there is
only one value

of a that satisfies the above condition

.

Given the initial

wod p that caused a coliion

: P(ha p(X1) ,
=
fra ,
p(x23) =
1
20

Hence ,
the
probability that ha b(X)
,
=

h(X) for any two distinct

integers X, and X is at

most 120
making the hash family universal.

b suppose we has so elements into the hash table ,

calculate expected number of collisions
in the Mash falls

Given hash elements into table of

that we are
trying to 50 a size so

of ways to choose elements from

The no
50-049-12 -0
2

The
probability that each individual pair collide under the

is
3 +
=
0 .
05-0
universal hash function
From 0 + O
: The expected no ·

of collision =
1225X0 05 .

=
61 25
.

...
Expected no
of collision while
hashing - 61 25.

50 elements to universal hash function

③. Hash table of size 1

Primary ash function

· : h , (x) = X mod I

he
secondary hash function (X) It (x mod 10)
· : =

9) Insert [27 ,
18 ,
29 ,
28 ,
39 ,
13 ,
16]

Insect 27 :

Mi (27) : 27 mod 11 16

= 5 2 13

position 39
so , adding 27 to 5 3

28
4

27

InsertIS (18)
3
-

:
18 mod 1 6 29

=
7 7 18

so ,
adding Is to
position 7 S

Ip
Insert
29 :

h , (29) :
29 mod 11

=
- (collision found)
he (29) : 1 + 129 mod 10)

= 1 + 9

= 10

new
position :
(7 + ixI0) mod 11

i 1 17 mod 11 6
say
: :
,

so ,
adding 29 in position 6
 Insert 28 :
-

hi (2) :
as mod 11

=
6
(collision found)
M2 (28) : 1 + 128 mod 10)
= 1 + 8

=
9

16 + ix 9) mod
new
position >
-
11

i 1, 13 mod 11 4
say
: :

--
so adding2s in position 4
.

39
set -
39 mod

=
6 (Collision found)
h 2 (39) :
1 + 139 mod 10)

=
1 - 9

= 10

new position :
(6 + ix10) mod I

say i : 1
,
I mod 11:3 / collision
found
i: 2 26 modi ↑ (collision founds
say ,
:

36 modi
sayi 3 3
: :
,

so , adding 39 in position 3

Insert
13 :

hi (13) : 13 mod 11

= 2

so ,
adding 13 in position 2
cert: 16 mod 11

=
5 (collision found)
he (16) : 1 + 116 mod its
= 1 + 6

= 7

new position :
(3 + ix7) mod
modli
say i 1 12 :
,

so ,
adding 16 in position
,

Final hash table :

16

2 13

3 39
28
4

3 27

6 29

7 18

Ip
 b) Explain how double
hashing reduces
clustering in
comparison to linear
proving :

Linear
probing :

linear probing when collision happens at a certain index the

In
algorithm checks
·

the next slot until an

empty slot is found.

·
This will lead to a
group of continuous slots being occupied and hence forming
primary clusters
.

Once cluster is formed a

future collision is more frequent in the same
region
·
a ,

Double Hashing :

·
Double hashing reduces
clustering by bringing a second hash .
function

When
·
collision occurs , instead of checking the next consecutive slot like linear

the doubl hashing has .

proving ,
uses a
secondary function

·
This makes the elements in the hash table more dispersed rather than

creating a
continuous cluster
.

ixhe(x)) modn
In double hashing position (h(x)
·
new : +
,
for
where i-no .

of times search occured

,
that key
h , (x) hash function
-

Primary
h2(x) -

Secondary has function

n-table size .

Eg : Linear
proving Double Hasking
suppose collision happens at position 6 .

Suppose collision happens at

position 6
we
try index 7, 8 , . -- and so -
on we use a
secondary has function to
determine a new
position which may

be 2 or a etc...
④ Implementation of a hash table using chaining :

import random
class UniversalHash: # Initialize universal hash
function with table size
def __init__(self, table_size, p=101):
self.table_size = table_size
self.p = p # largest prime no when compared to inpot
self.a = random.randint(1, p - 1)
self.b = random.randint(0, p - 1)
def hash(self, key): # Applying universal has function on
key
key_value = sum(ord(char) for char in key)
return ((self.a * key_value + self.b) % self.p) % self.table_size
class ChainedHashTable:
def __init__(self, size):#
Initialize hash table with
empty list
self.size = size
self.table = [[] for _ in range(size)]
self.hash_function = UniversalHash(size)
def set(self, key, value): #
inserting value
for key
index = self.hash_function.hash(key)
for i, (k, v) in enumerate(self.table[index]):
if k == key:
self.table[index][i] = (key, value) # existing key
return self.table[index].append((key, value))
#
inserting new
key details

def get(self, key):

# Get values for the
given key
index = self.hash_function.hash(key)
for k, v in self.table[index]:
if k == key:
return v
# when
return None key is not found
def display(self): # Printing Hash table
for i, chain in enumerate(self.table):
print(f"Index {i}: {chain}")
# Testing
data = [
("person1", "[email protected]"),
("person2", "[email protected]"),
("person3", "[email protected]"),
("person4", "[email protected]"),
("person5", "[email protected]"),
("person6", "[email protected]"),
↑
("person7", "[email protected]")
]

# creating chaired
hash table
hash_table = ChainedHashTable(size=8)
for name, mail in data: #
inserting data in hash table

hash_table.set(name, mail)
print("Printing the hash table:")
hash_table.display()

print("\nFinding the value for keys:")

print(f"Person 1: {hash_table.get('person1')}")
print(f"Person 2: {hash_table.get('person2')}")

print("\nFinding values for non-existing key:")

print(f"Person 8: {hash_table.get('person8')}")
ODPUT :

Printing the hash table ::

Index 0: []
Index 1: [('person1', '[email protected]')]
Index 2: [('person7', '[email protected]')]
Index 3: [('person4', '[email protected]')]
Index 4: [('person3', '[email protected]')]
Index 5: [('person2', '[email protected]')]
Index 6: [('person6', '[email protected]')]
Index 7: [('person5', '[email protected]')]

Finding the value for keys:

Person 1: [email protected]
Person 2: [email protected]

Finding values for non-existent keys:

Person 8: None

Total time complexity of unversal hash

function is O(C) where I is

length of key
.