Scribd
Upload
19 views

data structures

data structures chapter two

Uploaded by

Sam Amulya
Copyright
© © All Rights Reserved
Available Formats
Download as PDF or read online on Scribd
19 views

data structures

data structures chapter two

Uploaded by

Sam Amulya
Copyright
© © All Rights Reserved
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 24
Dictionaries and Hash Table Representation . SyYttasus Dictionaries: Linear List Representation, Skip List Representation, Operations - Insertion, Deletion and Searching, Hash Table Representation: Hash Functions, Collision Resolution - Separate Chaining, Open Addressing - Linear Probing, Quadratic Probing, Double Hashing, Rehashing, Extendible Hashing. (EEARNING OBJECTIVES | Introduction to Dictionaries Skip List and its Representation Various Operations Performed on Skip List Concept of Ideal Hashing Various Hash Functions Various Collision Resolution Techniques SOAS ES SK OS Concept of Extensible Hashing [A dictionary consists of « group of keys and their associated values. Its of the form (K, V) where, 'K! denotes the key and ‘V’ denotes the value of the key ‘K’. These keys are unique. Various operations performed on dictionaries are find, Insert and erase. A skip list is a probabl data structure for storing an ordered list of items. It was developed by William Pugh in 1990. It is based on a collection of linked lists. : Hashing is 0 technique that makes use of « hash function for mapping pairs to the corresponding (entries) positions in a hash table. Ideally a pair ‘P' with a key ‘K’ is stored at a position f(K) by the hash function “". Each hash table position can store one pair, Hash function is @ function that is-used to map the dictionary pairs to the corresponding entries. in the hash table. The various Hashing functions include Division method, Mid square method, Folding Method and Digit Analysis. 42 2.1 DicTiONARIES Q1. Explain briefly about dictionaries with an example. Answer : Dictionaries A dictionary consists of a group of keys and their associated values. It is of the form (KV) where, *K denotes the key and ‘17 denotes the value of the key “K”. These keys are unique. ‘Example 1 +The “company list” for the data structure “job” is a dictionary that consists ofa group of “employe Who are working in that company. When a new employee is appointed, a record for that employee is added to the dictionary. Similarly, when an employee leaves the company, his/her record gets deleted from the diction- ary. During the employment period, a manager may query the dictionary to get the record associated with a particular employee and can make changes in the record. ‘The employee-id (eid) is considered as the key’and the other details of an eniployee is the value of that key. A dictionary with duplicates is a dictionary that allows the same Key to be'used by two or more (key, value) pairs. The operations that can be performed on dictionary with duplicates include find, insert and erase. Example 2 A telephone directory. The ambiguities in ‘find” and ‘erase’ operations ina dictionary with duplicates can be eliminated by, (Finding a pair that contains the specified key or Gi)__By finding all the pai specified key. ° For the ‘erase? operation, a user must decide which pair to delete from among all the pairs and can either delete that pair or can delete all the pairs that are associated with that specified key. 4 The ‘find’ operation finds and retrieves the pairs randomly when the key is provided. Some applications makes use of sequential access to retrieve pages one-b: one in increasing order of their Q2. What are the operations that can be per= formed on a dictionary? How a dictionary is specified in a ADT? Answer : Operations @ Adding pair to a dictionary, that are associated with a (Deleting the pair associated with a specified key from the dietionary.- ii) Finding the pair that is associated with a given key. iv) Finding the number of pairs ie. size of adictionary. (v) Determining whether the dictionary is empty or ful DATA STRUCTURES [JNTU-HYDER, Aan) Dictionary in ADT Dictionary in ADT is specified as, Abstract Data Type Dictionary { Instances: a ‘group of pairs with unique keys for each pair, Operations . insent(p) : Insert the pair ‘p"to the dictionary; erase(k) : Delete the pair wth key ‘K’ from the dictionary, find(k) _: Finds and returns the pair with key k; ize()- : Finds the size of a dictionary; empty(): Returns true if the dictionary is empty ely returns false; } if The key and its associated value in'a pair ‘p* i, denoted as ‘p-first’ and‘ Q3.__ How to perform reassign operation ona dictionary? Answer : (Model Paper-tl, @4(a) | AprivMay-23(R18), a4) The reassign operation on a dictionary is per- formed when the existing value need to be modified which is associated with a key. Alternately, a new value can be mapped to a kéy which is already present in a dictionary. Syntax: Modify(c, ¢) Here, element ‘e’ is replaced/modified with its locator ‘c*. For remaining answer refer Unit-ll, Page No. 42, a it Representation Q4. Discuss the two classes that are used for the linear list representation method. Answer : : The two classes that are used for the linear list representation includes,” 1. Sorted Array List Class . The sorted array list class makes use of the ar ray representation of a linear list. This class follows the binary search technique. So, the find ‘operation requires the time of O(log n) for a. tionary consisting of n-paits. (@) Insertion: A pair is inserted into a dictionary only after ensuring that none of the already existing Pairs in a dictionary contains the same key 25 inserted pair. Hence, the total time taken to inseft a pair in a dictionary is the sum of, ‘Time for searching (find) + Time for inserting = Ollog n) + O(n) Deletion: A pair is deleted from the dictionary after searching it in the dictionary. Hence, the {otal time required for deletion is given as, Time for searching + Time for deleting => O(log n) + O(n) (by jonaries and Hash 3 UNIT: Table Representation Femento) 1 21 B11 13) (6 01 [eT Tso To [0 Bement) = 10 Eke} = 39 Ekement[2} = $0 Ekrmeni{3} = 99 Elemen(a] = 99 Blemeni[5] = empty Ekmeni( e em = ery 2 Sorted Chain Class . ‘The sorted chain class uses a linked I i l list representation method. In this class, the nodes are ta The cost of finding the pairs nereasing order of their keys is O(1) per pair. Example for Sorted Chain/Linked List Representation " Kor vate Key Vat Key Vale Key Vabe Key Vahe = 3 [0 [C4 [0 Ts Te Trl 2.1.2 Skip List Representation Q5. What is skip list? How it is different from a linear linked list? OR What is a skip list? 7 Answer : . ApriliMay-24R18), Q1(c) Skip List A skip list is a probabilistic data structure for ‘storing an ordered list of items. It was developed by William Pugh in 1990. Itis based on a collection of linked lists. These linked lists connect increasingly sparse subsequences of items as shown in the figure. Sm) 4 ‘ull 3 5 i) Figure: Skip List Example ‘The above list is called skip list because it skips certain items in the list. The last list in the skip list structure is a regular ordered linked list without any clement being skipped. The primary drawback associated with linked list structure is that, they need sequential scanning in order to locate a searched element. Ina linear linked list, the search process initiates from the start of the list and terminates only when the item being searched is found or when the complete lis is covered. The search process of linear linked list can be speed- up by ordering the items in the list. However, sequential scanning is still required. The skip list on the other hand is a special type of ordered linked list that allows non-sequential scanning of elements. In this list, the second element links to the succeeding node that is two positions ahead, the fourth element links to the succeeding node that is four positions ahead, and so on. Whereas, in the linear linked list, each element links only to its adjacent node. Q6. Explain about the representation of skip list structure and skipList class. Answer : ° . Skip List Structure * . _ Ina skip list structure, the header node requires specified number of pointer fields at each level while the tail node does not require any pointer field as each node consists, of an element field and the sufficient pointer fields (i.e., level number = 1). Element | Pointer Node} feid | “fields Figure: Skip List Structure SPECTRUM ALL-IN-ONE JOURNAL FOR ENGINEERING STUDENTS Where, [Number of pointer fields = Level number +1 Hence, astruct SkipNode is used that satisfies all the requirements of the nodes which is given as, template struct SkipNode ‘ typedefPair < const K, E > pairType; pairType SkipNode ** n; SkipNode(const pairType& P, int s) eP) {n= new SkipNode * (S];} 5 Where; The struct “SkipNode” ¢=Element, ; n=Next element p= Pair 3 S= Size. The pointer fields are denoted by the. airay n and the constructor allocates the memory space for the “n’ array of structures and keeps the pair into ‘e’. When the constructor is called, the size is given as, Size=Lev+1 for each ‘lev’ pair. Gy SkipList Class class SkipList t private: Data_members; public: constrictor; find(const K& key); insert(const pair & pair); erase(const K& key); B z Data Members The data members of the SkipList class are as follows. @ _~ SkipNode * headerNode It specifies the pointer to the header node. (i) SkipNode * tailNode It specifies the pointer to the tail node. (ili) SkipNode < K, E> ** Last Itrefers to the last node. Ifthe subscript ‘is mentioned with last as Last{/] then it denotes the last node of level ‘i*, Gv) int levels ‘It denotes the maximum number of non-émpty chains currently existing in the skip list, @) int dsize It represents the total number of < key, value > pairs in the dictionary, ; 3 book ) DATA STRUCTURES [JNTU-HYDERABAD, Int maxlevel It specifies the maximum number of levels thay are allowed in a particular chain. (vil) float cutort It is used in deciding the number of the levels, (vill) K tallKey Itrefers to a large key. class SkipList { private: ‘SkipNode < K, E> * headerNode; SkipNode < K, E> * tailNode; SkipNode ** Last; int levels, dsize, maxlevel; float cutOf; K tailkey; ho. fi Constructor The constructor of the SkipList class takes three parameters as, @ iy Larger Key : This is of type ‘K’. It is greater than any key in the dictionary. It is used in the tailNode. Maxpairs 5 It is of type ‘int’,and denotes the maximum number of pairs, a dictionary can have. Hence, the maxlevel is, maxlevel=[logyp maxpairs}-1 Prob Itis of type ‘float’ and denotes the probability that pair is present in both the level i and the level 4-1 chain. The constructor initializes the SkipList class data members as, tailkey maxlevel = (int) cell(logt ((float) maxpairs/ logf{I/prob)) ~ 1); headerNode = new skipNode (pair tailpair, maxlevel + 1); tailNode = new skipNode (pair tailpair, Q); Last = new skipNode *(maxlevel + 1); Where, tailpair.first = tailkey; This constructor sets aside the space for the “tail’ and the ‘header’ nodes. The array ‘last’ refers t0 the last node in each chain during the search operation. The skiplist contains ‘maxlevel + 1° Pointers from the “header’ to the ‘tail’ node. The constructors complexity is O(maxlevel), CRIMINAL act. Anyone found guilty is LIABLE to face LEGAL proceedings. UNIT-2: Dictionaries and Hash Taby lo Representati G7. Explain in detail about skip ate ha = in data on structure For answer refer Unit, Page No, 43 (Model Papers, 4b) | Maren-21(R18), 08) 8. Describe how the key Nn of a di awe a dictionary. ApritMay-23(R 10), 44a) ‘The number of comparisons can be dccrease tol 1 @ Maintaining a pointer to the middle pair. Gi) By comparing the pair to be sea Gi). For a smatter key p; ) for an n-pair dictionary by, irched with the middle pair. Search only the left half portion of the sorted chain. Search only the right half portion of the sorted ehain. parisons can be reduced, (iv) Fora larger key pai Hence, the number of com Example Consider the sorted chain consisting of 9-pai tail node and associated keys wit shown as Recent ith cach node, figure. This sorted chain has a header node, a shown as the numbers in the figure. As there are nine pairs, a search involves 9 key comparisons, Thi to S by keeping a pointer to the middle node. pepe: Tile ees of comparisons can be reduced . = PPS ay S-f1-PEP Sh Header Node ‘Middle Node Tail Node In order to search for a key, first it must be compared with the middle pair key. Ifits valuc is less than the riddle pair key value then the search operation is performed only on the left side portion of the middle pair. Else, ifvalue is greater than the middle pair key value then search is performed on the right hand side of the sorted chain ic. if a pair with a key 37 is to be searched, then 37 is compared with the middle key value 50 as 37 < 50, only the left side is examined. Similarly, for a pair with a key value 75 can be obtained by examining only the right side Portion of the sorted chain thereby reducing the number of comparisons. + ____ These comparisons can even be reduced by further maintaining two separate pointers at the two sides of the middle pair. For ‘n’ nodes, the level 0 includes all the node pairs, the level 1 includes every second pair and the level 2 includes every fourth pair and so on i.e, the level j includes every 2 #* pair. and Searchi 2:13" Skip List Of 1 and|Searching| Q9. Explain the insertion and searching operations of skip list with an example. Answer ; ion operation of a skip list is an extension to its search operation. It makes use of randomization tech- nique Sereda a eta! of references to the new item that mist be added to the skip list. First, the skip search(x) operation is performed for adding the new item (x, ¢) into skip list. This operation results in a position *pos' ofthe last-level item whose largest key is less than or equal to item x. The new item, e) is then inserted into the skip list immediately after the position pos. A function rand{ ) is invoked ora coin i tossed after inserting the new item which returns a random number between 0 and 1. If this number retumed is below 0.5 then, at to be “heads and backtrack to the higher level for inserting the item. If the number is above 0,5, consider it to tails and stop the processes. The insertion process is repeated until tail is obtained. “The tower for the new item (x, €) is constructed by linking all its references together. The’ algorithm for insertion operation of a skip list is given below. ‘SPECTRUM ALL-IN-ONE JOURNAL FOR ENGINEERING STUDENTS ipinsert(s. ©) pon item at lowest leve} ‘Algorithm ae Input: item(, €) jon that points 10 * Output: position pos — skipSearch(x) pos] « insertAferABOVE(POS: & position(pos!) while rand( )< 172 do . while above(pos)= null do yscanni pos before(pos) . pos < above(pos) pos] — insertAfter/ inpositi null, (x, €)) ing backward ‘Above(pos, posl. (6) Inserting item return pos? 3 Example: Insertion of new item into the skip list * a Ly = 7 I — Lb 1 15 45 I 1 I Lb |= 7 1s 21 45 I == I I ili L [2 7 15 |] 21 28 MM 45 tro Tesi ot Co 1 ; wheHe H7 Hel sH2He Hers H 45 i “To insen the key 32 into the list, a coin is tossed to get the insertion position. The tossed coin re consecutive heads followed by a tail. So, the skip list will be as follows, Le te 1 . fe] Ly |= 7 um oT I __ am bl= 7 15 22 45 te I I I »[= 7 1s 20 32 45 ! plea eal a) at I = 2S} T Cc r uo EHH He Helse Lee ‘The positions holding the newly inserted item aes item are shaded while th ns prec ith thick borders. The expected running time for the above insertion algorithms Sloane deserts en Q10. Explain the deletion o} oe peration of skip list with an example. Deletion Operation of a Skip List c wccessed by a a Sees cea ; by applying an above( ) method for moving up in th i time for deleting an item from the skip let x Example bk = ui 7 _L I to lL, |— 7 iis] _L, I I fa -— Hp : Ly |= 2 I I tk Hf a1 Hos a elfen I i I L Lp |=} 2 7 LEACH 10 Has Har fos LH He Hs LH Hs LH Some improvements associated with skp list (There is no need for storing pointers to items at ee Pointers to items at ii) There is no need s oS down oi to make use of above( ) and before) method because insertion and deletionsean be made in top-down and scan ina forward way thereby saving the memory tized By "uP and pre’ pont insertion, the deletion of an item is made asthe search moves down the skip list. The changed top-down structure employed by deletion operation elimi e ee structure are, ferent levels of skip list because pointers to keys will be ies te —_ - La [= 7 pe i I ; Lb 7 “S t i I I a id a5} T I i i ule247 34 45] [+= Lo L I I i Lo |= 2 a |_{ 34 LJ 40 | 45 tee To delete 15 from the skip list, firstly search for 15. The last pointer found on the chain is at level 2 with node 7. Thus, it is necessary to change the connection between the nodes from level 0 to level 3. At level 0, con- neci node 10 to 21. At levels 1 and 2 connect node 7 to 21 and at level 3, connect the node 7 with the node 32. By changing these connections, the node 15 will be removed. Q11. Explain the algorithms of skip list operations. OR Write an algorithm of skip list operations insertion and deletion. (epee OF TopIESY AIROF Yor IBERIORA ROP POR DeTetION) Answer : The algorithms for variou: @ Algorithm for Searching The algorithm for searching operation in s Algorithm SkipSeareh(L, key) Begin p=L{header] 2 for(i = level; i <=0;i-~) while(p = key —> forw[i] forwf P=p— forw[0) if (p + key = key) Retum p> value else return End 2, Algorithm for Insertion The algorithm for insertion operation in skip list is as follows, Algorithm SkipInsertion(L, key, item) Begin P= Lfheader); for(int i = level; i> -~) while(p = key -» forw[i] < key) P=p- forv{i] ili] =p P=p forv{o) if (p > key = key) ‘Update p[value] = item else IM = random _level() ifQ1> level) for (i= level +1 ip[i] = Lfheader] level = IvI; P=new snode (Ivl, key, value) forfint i P— forw[i] = ip{i] > forw{i] ipli] > forwfi End 3. Algorithm for Deletion The algorithm for deletion operation in skip list is as follows, Algorithm SkipDeletion(L, key) Begin p=Lfheader] forfint i= level; i> = 0; i--) while(p = key > forw{i] < key) P=p— forw(i] ipli]=p P=p- forw([0] if (Pp — key = key) for(int i = 0; i 1&&L| == NULL) level -- <=lvh iH) [header] + forw[level] End DATA STRUCTURES [INTU-HYDERAGay Q12. Explain the operations of the ski representation. (Model Papers, a) Marcha, ayy, OR P ling Describe the operations of skip list an example. Answer : Pee=t90R 18, 9, For answer refer Unit-II, Page Nos. 45, 46, 29, QI9, ABLE REPRESEy "ATION es Q13. Explain ideal hashing with an example, Answer : Hashing with A dictionary can even be represented by using , method called “hashing”. Hashing is a technique that makes use ofa hash function for mapping pairs to the corresponding (entries) positions in a hash table. Ideally a pair *P” with a key °K’ is stored at a position (K) by the hash funetion Each hash table position can store one pair. Inorder to search a pair with a key *K’, first) is calculated to determine the existence of a pair at a position (K) in the hash table. If a desired pair is found then it can be deleted, if required, else if, it does not ex- ists in a table then it can be inserted at a position f(A) Example: Consider a dictionary consisting of ‘employ- ce’ records. If employce id (eid) is used as-a key, then for atmost 100 employees, the eid’s ranges from 301000 and 302000. The hash function (K) = K — 301000 maps the eid’s to the table positions 0 to 1000. Table [1001] can be used to store dictionary pair pointers. The array table[#] is NULL for 0 int div_method(int a, int b) {retum a%%b+1; } void main) \_val=div_method(a,b)//Calling hash function Printf("The value of 26 is stored at index ‘%d" hash_val); getchQ; 3 Output i ‘count_dig(long int key) imte=1; while(key !=0) { key /= 10; » ed retum e; i Set_Nth_dig(long int n, int og long int dig=0,frac=0, len; = count digi), ot Gindpow(to, (len- Gig frac % 10; ety loc); DATA STRUCTURES [JNTU-HYDERA Bp, num= num num= num + dig: | jretum num; i int hash_midsq(long int key,int $7) t int hashval= 0; int keysq_len=0; * int mid_loc=0; long int keysq = key*hey; keysq_len = count_digikeysa); mid_loc= keysq_len /2; hashval= get_dig_rangelkeysq,(mid_locs Dezy, return hashval; j int main() clrser(); printf(“nThe middle digits of a key 5420 based digit address = “A” hash_midsq(5420,1)), printf(*\n\nThe middle digits of a key 5420 based g 2 digits address = ud" hash_midsq(3420.2) 9” Printf(“1n The middle digits of a key 5420 based ¢ 3 digits address = %d"hash_midsq(5420,3),, getch(); return 0; : ont + Dictionarios and Hash Ta xpiain tho algorithm for imptomenting quadratic probing on a hash table, (REPTOAY TOTS QUIIRTTE TIO, answer Aprninbay 24R1R), 08 cottston Resolatlon Collisions are said to occur when a new key is uahest tam autess hy using the hashing methods that tar provate one-t0-00 Mapping, There ie a nce for anating these collision. Por this purpose, there are many Waals that are independent of the hashing algorithm uNt used aypes of Collision Resolution The various types of collision resolution techniques ate as follows, 1, Open addressing (or) Closed hashing Separate chaining (or) Linked list (oe) Open hashing, Open Addressing (or) Closed Hashing Open addressing ean be defined as one of the methods used for resolving the collisions, This process ofeallision resolution takes place in an area called as the prime area that consists of the home addresses, Whenever collision takes place, the prime area addresses will be scanned until open (or) vacant element is found. When ‘an unoccupied element is found the new data will be placed in that position. Open addressing consists of four methods, @ Linear Probing In linear probing method of hashing, fis a linear function of i, typically fli) = i. Linear probing is a simple open addressing strategy for collision handling, In this method, the cells are sequentially searched to find an empty cell. Example: Assume that, keys: 16,23, 43, 18, 34, 59, 30,22 Size of Hash Table (HT) = 10 Then, modular arithmetic function will be, Hx) = x% (size of HT) = x% 10 Initially, the Hash Table is empty. 1 or 2 3 4 5 6 78D Inserting 16: For inserting she first key *16" into thi table, compute h(x). W(x) = x% 10 = 16% 10= So, the key 6 is inserted into 7" 16 ar, 4.5 67 8 9 So, the key 23 will be inserted into 3” pos 23 16 OT a8 4 5, en Seret OF 53 4196 10 Collision will occur if 43 is placed at 3" position pecause 21 is already present there in that position, So the next cell will be searched after J". which is 4" position. Thus the hey 43 is inserted in 4 position, nf 16 . 1.2) 4 Ss 6 7 8D Inserting 18: A(A) = 18%6 10= 8 J at 8* position. Thus, the key 18 will be plac as 16 18 o 1 2 ’ 4s 67 8 0% 104 Inserting 34: hk Ifthe key 44 is placed at 4* position, the collision will occur. $0, the next available cell is searched after 4 which is 5, Thus 34 will be inserted into S* position. nf of [6 18 .o 123s # 5 6 7 8 9 Inserting $9: h(k) = 59% 109 Thus, the key 59 will be inserted a | 9 [34 [16 OeUInaraierro nae eon Onn? erting 30: /(A)= 30% 10 = 0 ‘Thus, the key 30 will be placed at 0 position. 30 a | 3 [34] 6 18 59, 0 7 8 9 Insert 5 at 2" position. 30 2 fe [as] io 18 [59 (OM inate 2 rs Sura ers GIL 7b OE) ‘The implementation of linear probing is as follows, element* LinearProbing_Search(int key) { int curr_bkt, hm_bkt=hm(); for(curr_bkt =hm_bkt; ht{curr_bkt] && hifeurr_bkt] ey;) ( cuirr_bkt = (curr_bkt + 1) %p; itourr_bkt == hm_bkt) retum NULL; ) if{ht{curr_bkt] > k = = key) return hifeurr_bkt}; return NULI } ‘SPECTRUM ALL-IN-ONE JOURNAL FOR ENGINEERING STUDENTS. D = But quadratic probing ‘guarantees of inserting = “That is atleast half empty if the (i) Quadratic Probing, Quadratic probing overcomes the clusteritt problem of linear probing. This is a collision resolution ‘on function is quadratic. The method where the col quadric probing function is express method, when the collision occurs, in-a cell which is one cell away from 1 ed as fli) = ?. In this the key is inserted he collided cell Algorithm The algorithm that is used fo quadratic probing on a hash table is given below. ‘StepI: Start Step2: Initially read the key value (i.c., Key) or element to be inserted. Step3: Assume j = 0. ‘Step: Compute hash key, Le., Mkey) = key % HT_size. ‘StepS: Then compute index of the hash table where an element to be inserted i.e., Index = (hkey +j * D% HT_size. Step6: If clement is not present at that index then, insert the value at that index and go to step-9. Step7: Else, if element is already present at that index then, increment the value of §' (i.e.,j=5+ 0). Step8: Ifj = | a] _® oe a Hash Table with Double Hashing/Rehashing After Inserting {89, 18, 49, 58, 69} pe WARNING: Xerox/Photocopying of this book is a CRIMINAL act. Anyone found guilty is LIABLE to face LEGAL proceedings- Tre roeriing 4 te ecolved by evaluating the #49. This _stion TE oe iserted in el ee) ne inserting $8. I is resolved by hash es oS go $8 is inserted in position 3. Next whe asertd in position Oat distance of hash (60) cotG= Laway: There are some hod cases if insert y = js itoa series of collisions. Fr example if ion seating 60m psitonOthenitreits ina ee fie Hijisions at cells 3, 6, 9, since hash,(60) = Taek, ed until an empty spot is found, = ‘The hash table size in the above example i coe. Buttisdesinbletohave the able sneeeee ee pot then it is possible to run out of altemative vations prematurely. For example, if user try to inset Spove table it collides with $8. Since hash,(23) x 3 into a = § and table size is 10 (even) then user have Satyone alternative location which is already occupied. if double hashing is implemented correctly, then ge expected number of probes in double hashing is, Soest same as fora random collision resolution strategy. jrewever quadratic probing is simple and fast because it oss not require a second hash function. Random Probing ‘The random probing method determines the next _qrsilable slot using a random number generator. It uses hs formula, hash code = (hf(x) + ran#) % size of HT Here, ran# is a random number. All insertion and searches uses the same sequence of random numbers. Example: Consider that the keys 36, 25, $5 are to be imered in the hash table. Also assume that ran‘ 1 = 3, . Therefore, the probe sequence wo For 55: 55 +3 =58, 55+ ‘Advantages of Open Addressing (@ There is no need for a separate data structure for storing the items because they are directly stored in the hash table. (i Open addressing proves to be more efficient, interms of storage. Disadvantages of Open Addressing () It is mandatory to have distinct objects to be hashed. (i) Table size must be chosen care! does not work well. keys for the fully otherwise, it state flag (full, (ii) Each cell must contain a three~ ‘empty or deleted). 2. Separate Chaining (or) Linked List (or) Open Hashing In this type of hashing, the hash table stort a set of pointers (rather than data elements) -ach pointer Points to a separate linked list. The data elements TS “SPECTRUM ‘ALL-IN-ONE J 55 ich element ‘stored in the nodes of these linked I “hash funetion (hf (k)) in applied, which gives the hash code (index) of a particular bucket (cell) in the hash table. The element is stored in the linked list, whic assogiated with that bucket (cell). Algorithm Algorithm Separate Chaining Input - The key(K) of element to be searched. A flag called INSERT to perform insertion. Output Return the location of element which is to be searched or return NULL ‘Step1: Find the hash address of key using hash function LashFunetion(key) Step2: Assign pointer P to any node in List, P= H,{1] Step3: Assign flag with false for controlling the search Flag =F ‘Step4: While(P —- NULL) and (Flag = F) do Step5: If{ P+ DATA = key) then Step6: Flag =T Step7: Return(P) Step8: Exit Step9: Go to else ‘Step10: Move pointers to next node, Step: End If Step12: End while Stepl3: Ififlag = F) then ‘Step14: Printf “The key is not found” Step15: If{[INSERT_Value) then ‘Step16: Insert_End_SL[1] Step17: End If ‘Step18: End If Stepl Example Keys to be inserted (3417, 3132, 7122, 5199, 5344, 6796, 1898} Hash function hf{x) = x mod 10. | Computing the hash values for each of the given input using the hash function. nf{3417) 7 (7th bucket) 1nf(3132) = 3132 mod 10 = 2 2nd bucket) 7122 mod 102 (2nd bucket) (collision occurred) 5199 mod 10 =9 (9th bucket) f(5344) 5344 mod 10 =4 (4th bucket) 6796 mod 10 = 6 (6th bucket) 1898 mod 10 = 8 (8th bucket) = PLINK Stop. OURNAL FOR ENGINEERING STUDENTS % Advantages of Separate Chaining ©) ies sz=ple and efficient collision resolution. @) Gay (0 The keys of the elements to be hashed may or may toh be unique, Disadvantages of Separate Chaining G) A separate data structure for chains must be Answer: Medel Papers, O20) | Sep.2i¢ns8), 2) Given bash table size ~ 13 . List of elements are 65, 34, 79, 114, 26, 85, 55, 09,22; 9%, Insert 65° Jn Quadratic probing hash function, HO) = (+f) mod table size = (65 +) mod 13 = 65 mod 13 =0 i. 3-456 7 8.9 uit t WARMING: Kerox{Phatocopying of this book fz a CRIMINA ‘Anyone found guilty is LIABLE to fm DATA STRUCTURES [JNTU-HYp; Insert 34 HG)= (G4 + 0°) mod 13 = 34 mod 13 = g As & position is empty insert the element 3g Insert 79 HG) = (79 + 0°) mod 13 79 mod 13=1 As I" position is empty, insert the element 79, (elsT TT TTT Pty Oo 12345678 9G T Insert 114 H(x)= (114 + 0) mod 13 = 114 mod 13 = 10 As 10° position is empty, insert the element 14, ssli7 T I I 34] TsT J ] o1 23456789 0 np tr Aso value ie, 1 H(@)= (26+ 27 mod 13 =1 As 1* position is al 7 value ie, 2. A(z) = (26 + 22) mod 13 = (26 +4) mod 13 =30 mod 13= 4 As 4* position is empty, 65] 79 26 I o 12s Position is not empty, increment the‘ so not empty, increment the insert the element 26. Bal Tay 1) $567 8 on Insert 85 HG) = (85 + 02) mod 13 Position is empty, insert the element 85. 26] T [stsa] [ay Td 10 2 Ome 2 ia Insert $9 HG) = (89 + 0°) mod 13 = 89 mog As 11® position is empty a sin 65 | 79 . a ar 34 ia ps9 Insert 22 sae 8 9 0 WW H(x)= (22 + 0°) mod ‘ 13 = 22 mod 13-9 AS 9° position is empty, insert the clement 23 6S 79 ~ : ~ x 85 | 34 | 22 [114] 89 2 5s 6 7 8 9 1 Wh 12 Insert 98 ‘ HQ) = 8 + 0°) mod 13° =98 mod 13=7 As 7* position is not empty, increment the ‘/' value i.e, | H(&)= 98 + 1) mod 13 a =99 mod 13=8 As 8* position is not empty, increment the ‘/’ value i.c,, 2. H()= (98 +22) mod 13 = 102 mod 13 = 11 As 11® position is also not empty, increment the */° value i. (x)= (98 + 37) mod 13 = 107.mod 13 =3 As 3" position is not empty, increment the ‘j’ valuz i.e, (x) = (98 + 4) mod 13 = 114 mod 13=10 As 10* position is also not empty, increment the ‘/ value i.e., 5. (x)= (98 + 5?) mod 13 = 123 mod 13=6 * , ' ‘As 6" position is empty, insert the element 98. ‘The resultant hash table is, . 65 | 79 | 55 | 26 os [85 | 34 | 22 [14] 89 5 2 ya SGC o 1 2 § ; 22. Write C programs to implement different collision resolution techniques. * Answer = r plemented by using the following programs, ‘ ate: i “The collision resolution techniques can Pe 1. Open Addressing Program : include #inelude ‘int HT. 7 < Susie; __cye journal SPECTRUM Lt DATA STRUCTURES [JNTU-HYDERAp, $$ Aly So haa : ss HT (int key) a ar int x; ' : . rintf(*inThe keys of th ae BIVEN belay, Sa aga “ rie for (x-0:x2HIT size;xt++){ ) : print(\nThe key at index % is %dx typ, int rehash_inear(int key) _//Liniear Probing Jbreak ; r cas intx; ) printf(“Enter size of hash table: "); x= (key+I)%HT_size; . scant (“¥%6d",&HT_size)s return x for (x=0:x

: #include #define HT_size 8 struct node { -int info; struct node*next; hb struct node *head[HT_siz. void insert() { struct node*nnode; int key; (NULL},*chain; Printf(“Enter value of a ash table: "); scanf(“%d",&key), Jrkey%HT_size; nnode=(struct node *)malloc(sizeof(struet node)); y=key; hnode->next NULL; While(chain->next!= NULL) , {chain=chain->next; } Void search) { int ind.key; Printf(“\nEnter value to search: Seanfl"%d" &key); ind=key%HT_size; Miheadlind] = NULL) Printf(“Element not found!\n"); else{ for(chain=headtind};chai ¢ ‘NUL Lichain=chain->next) if{chain>info= key){ Printf("Element found! break; , } if{chain-SNULL) ); - Printf(“Element not found!\n"); H void display(){ int for=05",chain-info); a» printf"); : } void main() t intkey,jchs clrser(); 7" while(1) : f : i ni \ sintf(*\n~-Separate Chaining, Method~\n1 Insert m2 Display genet ExitinEnter your choice: "), scanf("%d" deh); next) = Awe IANDNG! EOD ENGmEcRINS ene ch(ch) rchQ); ak; © Azexit(0): ‘ult: ntf(“Enter Correct Choice!”); tput See Ret Met eoess Bye per) oor Sacha ascites Pome Ls ere Seo eet ReR TN fii fees Siete 7 ere eee eerie ese Ceaser ares eng aires emai ores Brome ya me ts iL) elas 12a eae nae ete at ere eu ce Tae a Scere AE DOSRarOTL Chu speed manlt0% cyles Famestp 0) Programe iC We ase fs ea Bae ae ieee separate Chaining Method — Witeaas ANE dE) ioral (irs eateries Ties Enter value ta search: 3132 Recaray cbs .2 Extendible Hashing Q23. Explain in detail about extendible hashing. Model Paper, O42) OR List and explain the advantages of extendible hashing. (Refer Only Topic: Advantages) Answer: March-22(R18), 4) Extendible Hashing Extendible hashing is a dynamic hashing technique that is capable of’ dividing the buckets 25 database grows and rejoining them as it shrinks. This requires reorganization of buckets in hash table. Aa important feature of extendible hashing is that it reduces the performance overhead of hashing because it performs organization only on one bucket at a time. In extendible hashing, the hash. function “ chosen such that it has sufficient qualities of uniformity and randomness. And it is capable of generating valves over a lafge range, often known as b-bit binary inteset “The value of 6 is usually 32. Therefore, it can create 2 i.e., more than 4 billion buckets. However, all buckets 2 not created at once, instead they are created on dem: as records are inserted into hash table. tae tye Pes big Neg ee te tee ea ov Bucket | 1»... [oS Ne bene see Bucket 2 pa : Bocket Address Bucket S Table CAL Bucket ‘n° top of bucket address table. This value indicates that ‘d’ bits of hash value ix) are needed to find the appropriate bucket for item ‘x Note that, in figure, each bucket has an integer value associated with it, labelled asd, dd, value specifies the length of common hash prefix that is present in several consecutive table the same bucket. The number of entries that point to a bucket ‘nis given by 3-4), Example entries which point to Consider an example to insert some search key values into memory using extendible hashing. In this, a hash function h(x) = x mod 8 is considered for insertion. Given hash function, Hx) = x mod 8 Search-key values = 2, 3, 5, 7, 11, 17, 19, 23, 29, 31 Size of bucket = 3 Stept ‘Compute the hash values for each of the given search key value, ‘h(2) = 2 mod 8 = 2 43) =3 mod 8 = 3 A(S) = 5 mod 8 = 5 ‘A(7) = 7 mod 8=7 (11) = 11 mod 8= 3 WIT) = 17 mod 8 = 1 (19) = 19 mod 8 =3 4(23) = 23 mod 8 =7 (29) = 29 mod 8 = 5 G1) = 31 mod 8 =7 SPECTRUM ALL-IN-ONE JOURNAL FOR ENGINEERING STUDENTS 62 Step2 Compute the resultant hash valu HQ)=2=010 inary values for each of the WQ)= 3-01 WS) =5=101 W)=7= 11 : W)=3=011 W7)=1= 001 19) 11 nQ3)=7= 11 (29) = 5 = 101 W431) u Step3 Compute the global depth value and place it on the top of the bucket address table consisting of 8 slots numbered (000, 001, 010, 011, 100, 101, 110, 111) Global depth = 3 (+° The size of the directory = 3] Step4 Insert the key values in accordance to their binary values. Here, the number of buckets used are S because five different hash values are generated, 1,2,3,5,7. Now, consider the insertion ofthe search key value 2. Since, the computed binary value of 2 is 010, the appropriate bucket (ie., bucket 1) is searched. Since, the bucket is empty, the data entry 2 can be inserted. After inserting, a pointer is added from 010 address in the bucket address table to bucket 1. In the similar fashion, the remaining data entry values are inserted. Steps Verify whether any of the address slots are empty. If they are empty, then pointer needs to be inserted from the address slots to the respective buckets. This can be done by comparing the first two MSB's of the’ empty address slot with the first two MSB's of the succeeding address slot. If both the MSB’s matches then only a pointer is added to the bucket. Steps Compute the local depth for each of the bucket. This is done by taking into account the number of pointers attached to every bucket. Ifa bucket is assigned. two pointers, then the MSB’s of this bucket is compared with the MSB’s of the next bucket. If the MSBs of both buckets binary address matches, then the local depth is assigned the value depending on the number of bits that, ‘are matching. On the other hand, if'a bucket is assigned only one pointer, then the local depth value of that bucket is equal to the global depth value. ir IT scary ma = oO 7 4 EJ Petes ovo a 3 ou TT} Becker 1m 4 -—__5 101 }29 | Buckets 0 : Mm j=_==x_"4 | Bek Bucket addressable Z| toca Figure: Extendible Hash Structure Advantages ° The advantages of extendible hashing are as follows, 1 Variable Number of Buckets: In extendible hashing, buckets are divided when database increases. However, when data size decreases, buckets are rejoined or shrinked. 2. Minimum Number of Buckets: As bucket grows and shrinks, memory is utilized properly without getting wasted. 3. Increase in file size does not degrade the perfor- mance, 4. In terms of computation, data retrieval is less expensive 5. _ Itean access desired records immediately. Disadvantages é The disadvantages of extendible hashing are as. follows, 1. The records are not accessed sequentially. 2. The data may be erased unexpectedly, thereby requiring special precautions, RY SHOR ve T QUESTIONS WITH SOLUTIONS (VSQS) / dictionary consists ofa grouy ‘a group of keys of Keys and their associated values. His ofthe Form (K. V) where, “K" denotes ey and 17 denotes the value of the key "A" These keys af These keys are unique. we Gz. Define skip list, answer = skip list is a probabilisti _Askip list is a probabilistic data structure for storing an ordered yy William List of items. It was developed by sh in, 1990. itis based on a collection of linked lists 3. Define hash function. Answe! Model Papers, 21(¢) the hash Hash function isa fi is a finetion that is used to map the dictionary pairs tothe corresponding entries in tl sable. Q4. _ Define hash table. Answer = Hash table isa data structure that is used to store dictionary pairs. It is of fixes to apart of the dictionary pair called the key. The size of the hash table is re ize which ranges from 0 to table size — 1 ise se Tipp avnblealed Q5, List the operations in dictionaries. Answer = Model Papers, 1(4) “The operations in dictionaries are, (Adding pair to a dictionary. (i) Deleting the pair associated wi ated with a given key. ith a specified key from the dictionary. Gi) Finding the pair that is assoc (iv) _Determining whether the dictionary is empty of full. Q6. Write about midsquare method. Answer: solving the middle digits In midsquare meth from the number. The num! jtself and address is obtained by multiplied by ison the table size. 1od, the key is ected usually depend: ber of digits se? Q7. Givean ‘example for digital analysis method. Model Papers, 1) sition 3 to 6 and reversing 2164 by selecting digits in po: n should be used Answer: ‘nd the same arrangement patte Akey 7546123 is trans! their order. For a given key set consistently. Q8. Define collision. Answer : fo the address formed t position in the key some cn two different keys want 10 0¢cUPY the same home bucket. As the urs wh ‘an easily be handled. ‘situation that occ Acollision is ne pait, collision © bucket can have more than oF DATA STRUCTURES UNTUHYDERARAD, PORTANT QUESTIONS Eachus Srey about dictionaries with an eTaeTS Imvportant Qerton wie Uninth Page So Qt - ethod. — : Yor the near list representation m™ SS] Sisauss the Gwe classes that are used Important Question aswe anower meter Unt, Page Nn £2. C4 ; from a linear linked list? Lo Whats stip st? How itis different Important Question amporant Question | ApriUMay-23(R18), 4a) 5 (Samad CS Expizin the coerations of the skip list representation. easel) (March 22{R18), O3(2) | Dec.-19(R18), 25) ines) Important Question (important Question | March-22(R18), Q4{s)) Witte the dicorithm for calculating Hash function using division method. Aaswer Important Question For mswer refx Unie Il Page No. S1,Q18. ) ‘CG. Whet ere Gifierent methods of collision resolution in hashing? Explain in brief. paswes = (Aertiay-24RAE}, OS | AprUMay-23(R18), Q1(d) | March-22(R18), Q3(b) er, G2 De-6(8, Dees, ign |_O For anewer teie Univll, Paze No. $2, Q20. 0. ieaert the folowing ist of elements into the hash table Hash table is 12) 65, 24, 79, 114, 26, using Qu 35, 55, 89, 22 oon! USING Quadratic probing (size of (important Question | Sep.-21(R18), Q2) a

You might also like