Cuckoo Hashing – Worst case O(1) Lookup!

Background :
There are three basic operations that must be supported by a hash table (or a dictionary):

  • Lookup(key): return true if key is there in the table, else false
  • Insert(key): adds the item ‘key’ to the table if not already present
  • Delete(key): removes ‘key’ from the table

Collisions are very likely even if we have a big table to store keys. Using the results from the birthday paradox: with only 23 persons, the probability that two people share the same birth date is 50%! There are 3 general strategies towards resolving hash collisions:

  • Closed addressing or Chaining: store colliding elements in an auxiliary data structure like a linked list or a binary search tree.
  • Open addressing: allow elements to overflow out of their target bucket and into other spaces.

Although above solutions provide expected lookup cost as O(1), the expected worst-case cost of a lookup in Open Addressing (with linear probing) is Ω(log n) and Θ(log n / log log n) in simple chaining (Source : Standford Lecture Notes). To close the gap of expected time and worst case expected time, two ideas are used:



  • Multiple-choice hashing: Give each element multiple choices for positions where it can reside in the hash table
  • Relocation hashing: Allow elements in the hash table to move after being placed

 
Cuckoo Hashing :
Cuckoo hashing applies the idea of multiple-choice and relocation together and guarantees O(1) worst case lookup time!

  • Multiple-choice: We give a key two choices h1(key) and h2(key) for residing.
  • Relocation: It may happen that h1(key) and h2(key) are preoccupied. This is resolved by imitating the Cuckoo bird: it pushes the other eggs or young out of the nest when it hatches. Analogously, inserting a new key into a cuckoo hashing table may push an older key to a different location. This leaves us with the problem of re-placing the older key.
    • If alternate position of older key is vacant, there is no problem.
    • Otherwise, older key displaces another key. This continues until the procedure finds a vacant position, or enters a cycle. In case of cycle, new hash functions are chosen and the whole data structure is ‘rehashed’. Multiple rehashes might be necessary before Cuckoo succeeds.

Insertion is expected O(1) (amortized) with high probability, even considering the possibility rehashing, as long as the number of keys is kept below half of the capacity of the hash table, i.e., the load factor is below 50%.

Deletion is O(1) worst-case as it requires inspection of just two locations in the hash table.

 
Illustration :

Input:

{20, 50, 53, 75, 100, 67, 105, 3, 36, 39}

Hash Functions:

h1(key) = key%11
h2(key) = (key/11)%11

ch1
Let’s start with inserting 20 at its possible position in the first table determined by h1(20):

ch2Next: 50ch3

Next: 53. h1(53) = 9. But 20 is already there at 9. We place 53 in table 1 & 20 in table 2 at h2(20)

ch4


Next: 75. h1(75) = 9. But 53 is already there at 9. We place 75 in table 1 & 53 in table 2 at h2(53)

ch6

Next: 100. h1(100) = 1.

ch

Next: 67. h1(67) = 1. But 100 is already there at 1. We place 67 in table 1 & 100 in table 2

ch8

Next: 105. h1(105) = 6. But 50 is already there at 6. We place 105 in table 1 & 50 in table 2 at h2(50) = 4. Now 53 has been displaced. h1(53) = 9. 75 displaced: h2(75) = 6.

ch9

Next: 3. h1(3) = 3.

ch10


Next: 36. h1(36) = 3. h2(3) = 0.

ch11Next: 39. h1(39) = 6. h2(105) = 9. h1(100) = 1. h2(67) = 6. h1(75) = 9. h2(53) = 4. h1(50) = 6. h2(39) = 3.

Here, the new key 39 is displaced later in the recursive calls to place 105 which it displaced.

ch12

 
Implementation :
Below is the implementation of Cuckoo hashing

C++

filter_none

edit
close

play_arrow

link
brightness_4
code

// C++ program to demonstrate working of Cuckoo
// hashing.
#include<bits/stdc++.h>
  
// upper bound on number of elements in our set
#define MAXN 11
  
// choices for position
#define ver 2
  
// Auxiliary space bounded by a small multiple
// of MAXN, minimizing wastage
int hashtable[ver][MAXN];
  
// Array to store possible positions for a key
int pos[ver];
  
/* function to fill hash table with dummy value
 * dummy value: INT_MIN
 * number of hashtables: ver */
void initTable()
{
    for (int j=0; j<MAXN; j++)
        for (int i=0; i<ver; i++)
            hashtable[i][j] = INT_MIN;
}
  
/* return hashed value for a key
 * function: ID of hash function according to which
    key has to hashed
 * key: item to be hashed */
int hash(int function, int key)
{
    switch (function)
    {
        case 1: return key%MAXN;
        case 2: return (key/MAXN)%MAXN;
    }
}
  
/* function to place a key in one of its possible positions
 * tableID: table in which key has to be placed, also equal
   to function according to which key must be hashed
 * cnt: number of times function has already been called
   in order to place the first input key
 * n: maximum number of times function can be recursively
   called before stopping and declaring presence of cycle */
void place(int key, int tableID, int cnt, int n)
{
    /* if function has been recursively called max number
       of times, stop and declare cycle. Rehash. */
    if (cnt==n)
    {
        printf("%d unpositioned\n", key);
        printf("Cycle present. REHASH.\n");
        return;
    }
  
    /* calculate and store possible positions for the key.
     * check if key already present at any of the positions.
      If YES, return. */
    for (int i=0; i<ver; i++)
    {
        pos[i] = hash(i+1, key);
        if (hashtable[i][pos[i]] == key)
           return;
    }
  
    /* check if another key is already present at the
       position for the new key in the table
     * If YES: place the new key in its position
     * and place the older key in an alternate position
       for it in the next table */
    if (hashtable[tableID][pos[tableID]]!=INT_MIN)
    {
        int dis = hashtable[tableID][pos[tableID]];
        hashtable[tableID][pos[tableID]] = key;
        place(dis, (tableID+1)%ver, cnt+1, n);
    }
    else //else: place the new key in its position
       hashtable[tableID][pos[tableID]] = key;
}
  
/* function to print hash table contents */
void printTable()
{
    printf("Final hash tables:\n");
  
    for (int i=0; i<ver; i++, printf("\n"))
        for (int j=0; j<MAXN; j++)
            (hashtable[i][j]==INT_MIN)? printf("- "):
                     printf("%d ", hashtable[i][j]);
  
    printf("\n");
}
  
/* function for Cuckoo-hashing keys
 * keys[]: input array of keys
 * n: size of input array */
void cuckoo(int keys[], int n)
{
    // initialize hash tables to a dummy value (INT-MIN)
    // indicating empty position
    initTable();
  
    // start with placing every key at its position in
    // the first hash table according to first hash
    // function
    for (int i=0, cnt=0; i<n; i++, cnt=0)
        place(keys[i], 0, cnt, n);
  
    //print the final hash tables
    printTable();
}
  
/* driver function */
int main()
{
    /* following array doesn't have any cycles and
       hence  all keys will be inserted without any
       rehashing */
    int keys_1[] = {20, 50, 53, 75, 100, 67, 105,
                    3, 36, 39};
  
    int n = sizeof(keys_1)/sizeof(int);
  
    cuckoo(keys_1, n);
  
    /* following array has a cycle and hence we will
       have to rehash to position every key */
    int keys_2[] = {20, 50, 53, 75, 100, 67, 105,
                    3, 36, 39, 6};
  
    int m = sizeof(keys_2)/sizeof(int);
  
    cuckoo(keys_2, m);
  
    return 0;
}

chevron_right


Java

filter_none

edit
close

play_arrow

link
brightness_4
code

// Java program to demonstrate working of 
// Cuckoo hashing.
import java.util.*;
  
class GFG 
{
  
// upper bound on number of elements in our set
static int MAXN = 11;
  
// choices for position
static int ver = 2;
  
// Auxiliary space bounded by a small multiple
// of MAXN, minimizing wastage
static int [][]hashtable = new int[ver][MAXN];
  
// Array to store possible positions for a key
static int []pos = new int[ver];
  
/* function to fill hash table with dummy value
* dummy value: INT_MIN
* number of hashtables: ver */
static void initTable()
{
    for (int j = 0; j < MAXN; j++)
        for (int i = 0; i < ver; i++)
            hashtable[i][j] = Integer.MIN_VALUE;
}
  
/* return hashed value for a key
* function: ID of hash function according to which
    key has to hashed
* key: item to be hashed */
static int hash(int function, int key)
{
    switch (function)
    {
        case 1: return key % MAXN;
        case 2: return (key / MAXN) % MAXN;
    }
    return Integer.MIN_VALUE;
}
  
/* function to place a key in one of its possible positions
* tableID: table in which key has to be placed, also equal
  to function according to which key must be hashed
* cnt: number of times function has already been called
  in order to place the first input key
* n: maximum number of times function can be recursively
  called before stopping and declaring presence of cycle */
static void place(int key, int tableID, int cnt, int n)
{
    /* if function has been recursively called max number
    of times, stop and declare cycle. Rehash. */
    if (cnt == n)
    {
        System.out.printf("%d unpositioned\n", key);
        System.out.printf("Cycle present. REHASH.\n");
        return;
    }
  
    /* calculate and store possible positions for the key.
    * check if key already present at any of the positions.
    If YES, return. */
    for (int i = 0; i < ver; i++)
    {
        pos[i] = hash(i + 1, key);
        if (hashtable[i][pos[i]] == key)
        return;
    }
  
    /* check if another key is already present at the
       position for the new key in the table
    * If YES: place the new key in its position
    * and place the older key in an alternate position
    for it in the next table */
    if (hashtable[tableID][pos[tableID]] != Integer.MIN_VALUE)
    {
        int dis = hashtable[tableID][pos[tableID]];
        hashtable[tableID][pos[tableID]] = key;
        place(dis, (tableID + 1) % ver, cnt + 1, n);
    }
    else // else: place the new key in its position
    hashtable[tableID][pos[tableID]] = key;
}
  
/* function to print hash table contents */
static void printTable()
{
    System.out.printf("Final hash tables:\n");
  
    for (int i = 0; i < ver; i++, System.out.printf("\n"))
        for (int j = 0; j < MAXN; j++)
            if(hashtable[i][j] == Integer.MIN_VALUE) 
                System.out.printf("- ");
            else
                System.out.printf("%d ", hashtable[i][j]);
  
    System.out.printf("\n");
}
  
/* function for Cuckoo-hashing keys
* keys[]: input array of keys
* n: size of input array */
static void cuckoo(int keys[], int n)
{
    // initialize hash tables to a dummy value 
    // (INT-MIN) indicating empty position
    initTable();
  
    // start with placing every key at its position in
    // the first hash table according to first hash
    // function
    for (int i = 0, cnt = 0; i < n; i++, cnt = 0)
        place(keys[i], 0, cnt, n);
  
    // print the final hash tables
    printTable();
}
  
// Driver Code
public static void main(String[] args) 
{
    /* following array doesn't have any cycles and
    hence all keys will be inserted without any
    rehashing */
    int keys_1[] = {20, 50, 53, 75, 100
                    67, 105, 3, 36, 39};
  
    int n = keys_1.length;
  
    cuckoo(keys_1, n);
  
    /* following array has a cycle and hence we will
    have to rehash to position every key */
    int keys_2[] = {20, 50, 53, 75, 100
                    67, 105, 3, 36, 39, 6};
  
    int m = keys_2.length;
  
    cuckoo(keys_2, m);
}
  
// This code is contributed by Princi Singh

chevron_right


C#

filter_none

edit
close

play_arrow

link
brightness_4
code

// C# program to demonstrate working of 
// Cuckoo hashing.
using System;
      
class GFG 
{
  
// upper bound on number of
// elements in our set
static int MAXN = 11;
  
// choices for position
static int ver = 2;
  
// Auxiliary space bounded by a small 
// multiple of MAXN, minimizing wastage
static int [,]hashtable = new int[ver, MAXN];
  
// Array to store 
// possible positions for a key
static int []pos = new int[ver];
  
/* function to fill hash table 
with dummy value
* dummy value: INT_MIN
* number of hashtables: ver */
static void initTable()
{
    for (int j = 0; j < MAXN; j++)
        for (int i = 0; i < ver; i++)
            hashtable[i, j] = int.MinValue;
}
  
/* return hashed value for a key
* function: ID of hash function 
according to which key has to hashed
* key: item to be hashed */
static int hash(int function, int key)
{
    switch (function)
    {
        case 1: return key % MAXN;
        case 2: return (key / MAXN) % MAXN;
    }
    return int.MinValue;
}
  
/* function to place a key in one of 
its possible positions
* tableID: table in which key 
has to be placed, also equal to function
according to which key must be hashed
* cnt: number of times function has already 
been called in order to place the first input key
* n: maximum number of times function 
can be recursively called before stopping and
declaring presence of cycle */
static void place(int key, int tableID, 
                  int cnt, int n)
{
    /* if function has been recursively 
    called max number of times,
    stop and declare cycle. Rehash. */
    if (cnt == n)
    {
        Console.Write("{0} unpositioned\n", key);
        Console.Write("Cycle present. REHASH.\n");
        return;
    }
  
    /* calculate and store possible positions 
    * for the key. Check if key already present 
    at any of the positions. If YES, return. */
    for (int i = 0; i < ver; i++)
    {
        pos[i] = hash(i + 1, key);
        if (hashtable[i, pos[i]] == key)
        return;
    }
  
    /* check if another key is already present 
    at the position for the new key in the table
    * If YES: place the new key in its position
    * and place the older key in an alternate position
    for it in the next table */
    if (hashtable[tableID, 
              pos[tableID]] != int.MinValue)
    {
        int dis = hashtable[tableID, pos[tableID]];
        hashtable[tableID, pos[tableID]] = key;
        place(dis, (tableID + 1) % ver, cnt + 1, n);
    }
    else // else: place the new key in its position
    hashtable[tableID, pos[tableID]] = key;
}
  
/* function to print hash table contents */
static void printTable()
{
    Console.Write("Final hash tables:\n");
  
    for (int i = 0; i < ver;
             i++, Console.Write("\n"))
        for (int j = 0; j < MAXN; j++)
            if(hashtable[i, j] == int.MinValue) 
                Console.Write("- ");
            else
                Console.Write("{0} ",
                        hashtable[i, j]);
  
    Console.Write("\n");
}
  
/* function for Cuckoo-hashing keys
* keys[]: input array of keys
* n: size of input array */
static void cuckoo(int []keys, int n)
{
    // initialize hash tables to a 
    // dummy value (INT-MIN)
    // indicating empty position
    initTable();
  
    // start with placing every key 
    // at its position in the first 
    // hash table according to first 
    // hash function
    for (int i = 0, cnt = 0;
             i < n; i++, cnt = 0)
        place(keys[i], 0, cnt, n);
  
    // print the final hash tables
    printTable();
}
  
// Driver Code
public static void Main(String[] args) 
{
    /* following array doesn't have 
    any cycles and hence all keys 
    will be inserted without any rehashing */
    int []keys_1 = {20, 50, 53, 75, 100, 
                    67, 105, 3, 36, 39};
  
    int n = keys_1.Length;
  
    cuckoo(keys_1, n);
  
    /* following array has a cycle and 
    hence we will have to rehash to 
    position every key */
    int []keys_2 = {20, 50, 53, 75, 100, 
                    67, 105, 3, 36, 39, 6};
  
    int m = keys_2.Length;
  
    cuckoo(keys_2, m);
}
}
  
// This code is contributed by PrinciRaj1992

chevron_right


Output:

Final hash tables:
- 100 - 36 - - 50 - - 75 - 
3 20 - 39 53 - 67 - - 105 - 

105 unpositioned
Cycle present. REHASH.
Final hash tables:
- 67 - 3 - - 39 - - 53 - 
6 20 - 36 50 - 75 - - 100 -

Generalizations of cuckoo hashing that use more than 2 alternative hash functions can be expected to utilize a larger part of the capacity of the hash table efficiently while sacrificing some lookup and insertion speed. Example: if we use 3 hash functions, it’s safe to load 91% and still be operating within expected bounds (Source : Wiki)

This article is contributed by Yash Varyani. If you like GeeksforGeeks and would like to contribute, you can also write an article and mail your article to contribute@geeksforgeeks.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.

Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.



My Personal Notes arrow_drop_up



Article Tags :
Practice Tags :


3


Please write to us at contribute@geeksforgeeks.org to report any issue with the above content.