哈希表的简单实现,使用分离连接法处理散列“冲突”(即多个key映射到相同表地址)
#include<stdio.h>
#include<stdlib.h>
#define HASH_LEN 10 //表长
struct hash_node{
hash_node * next;
int key;
int value;
};
//取留数法
static int hash(int key)
{
return key % HASH_LEN;
}
//链表法处理冲突
static void collision(hash_node * vec[], int hash_addr, hash_node * h)
{
if (vec[hash_addr] == NULL)
vec[hash_addr] = h;
else
{
h->next = vec[hash_addr];
vec[hash_addr] = h;
}
}
static void ord_num_print(int i)
{
if (i == 1)
printf("the 1st element: ");
else if (i == 2)
printf("the 2nd element: ");
else if (i == 3)
printf("the 3rd element: ");
else
printf("the %dth element: ", i);
}
static void print_table(hash_node * vec[])
{
hash_node * tmp;
for (int i = 0; i < HASH_LEN; i++)
{
if (vec[i] == NULL)
{
ord_num_print(i + 1);
printf("NULL\n");
}
else
{
ord_num_print(i + 1);
tmp = vec[i];
while (tmp != NULL)
{
printf("%d ", tmp->key);
tmp = tmp->next;
}
printf("\n");
}
}
}
static void create_table(hash_node * vec[], int keys[], int num)
{
hash_node * tmp;
int hash_addr;
for (int i = 0; i < num; i++)
{
tmp = (hash_node *)malloc(sizeof(hash_node));
tmp->key = keys[i];
tmp->next = NULL;
hash_addr = hash(keys[i]);
collision(vec, hash_addr, tmp);
}
}
void insert_element(hash_node * vec[], int key)
{
hash_node * new_node;
new_node = (hash_node*)malloc(sizeof(hash_node));
new_node->next = NULL;
new_node->key = key;
int hash_addr;
hash_addr = hash(key);
collision(vec, hash_addr, new_node);
}
bool delete_element(hash_node * vec[], int key)
{
int hash_addr = hash(key);
if (vec[hash_addr] == NULL)
{
printf("no such key!\n");
return false;
}
else
{
hash_node * tmp = vec[hash_addr];
hash_node * last = NULL;
bool first = true;
while (tmp->key != key)
{
if (tmp->next == NULL)
{
printf("no such key!\n");
return false;
}
last = tmp;
tmp = tmp->next;
first = false;
}
if (first)
vec[hash_addr] = tmp->next;
else
last->next = tmp->next;
free(tmp);
}
}
int main()
{
hash_node * vec[HASH_LEN];
for (int i = 0; i < HASH_LEN; i++)
{
vec[i] = NULL;
}
int keys[HASH_LEN] = { 19, 14, 23, 68, 20, 84, 27, 55, 11, 10 };
printf("first time\n");
create_table(vec, keys, HASH_LEN);
print_table(vec);
printf("second time\n");
insert_element(vec, 5);
print_table(vec);
printf("third time\n");
delete_element(vec, 5);
print_table(vec);
getchar();
return 0;
}
如果不使用链表,另一个解决冲突的方法是当冲突发生时,就散列到表的另一单元。根据如下公式计算出下一个散列地址。
i=0,1,2...依次寻找
根据f()不同可分为:
(1)线性探测
f(i)=i
(2)平方探测
f(i)=i^2
/*
功能:哈希表,平方探测处理冲突
描述:表有一半以上空时,才能保证元素插入;
若表中元素大于一半时,要再散列(rehash),
将表大小扩大一倍,遍历原表元素,使之
散列到新表中。
*/
#pragma once
#include<vector>
template<typename HashObj>
class HashTable
{
public:
HashTable(int size = 10);
~HashTable();
bool Contains(const HashObj & x);
bool Insert(const HashObj & x);
bool Delete(const HashObj & x);
int TableSize() { return _Lists.size(); }
private:
enum EntryType{EMPTY, ACTIVE};
struct HashEntry {
HashObj elem;
EntryType state;//表中元素状态
HashEntry(const HashObj & e = HashObj(), EntryType s = EMPTY) : elem(e), state(s){}
};
std::vector<HashEntry> _Lists;
int currSize;
int myHash(const HashObj & x);
bool isActive(int curPos) { return _Lists[curPos].state == ACTIVE; }
int findPos(const HashObj & x);
void rehash();
};
template<typename HashObj>
HashTable<HashObj>::HashTable(int size)
{
_Lists = std::vector<HashEntry>(size);
currSize = 0;
for (int i = 0; i < size; i++)
{
_Lists[i].state = EMPTY;
}
}
template<typename HashObj>
HashTable<HashObj>::~HashTable()
{
}
template<typename HashObj>
int HashTable<HashObj>::myHash(const HashObj & x)
{
int TableSize = _Lists.size();
int pos;
pos = x % TableSize;
if (pos < 0)
pos += TableSize;
return pos;
}
template<typename HashObj>
int HashTable<HashObj>::findPos(const HashObj & x)
{
int TableSize = _Lists.size();
int offset = 1;
int curPos = myHash(x);
while (_Lists[curPos].state != EMPTY && _Lists[curPos].elem != x)
{
curPos += offset;
offset += 2;//f(i)=i^2也可写成:f(i)=f(i-1)+2i-1
if (curPos > TableSize)
curPos -= TableSize;
}
return curPos;
}
template<typename HashObj>
bool HashTable<HashObj>::Contains(const HashObj & x)
{
return isActive(findPos(x));
}
template<typename HashObj>
bool HashTable<HashObj>::Insert(const HashObj & x)
{
int TableSize = _Lists.size();
int curPos = findPos(x);
if (isActive(curPos))
return false;
_Lists[curPos] = HashEntry(x, ACTIVE);
if (++currSize > TableSize / 2)
rehash();
return true;
}
template<typename HashObj>
bool HashTable<HashObj>::Delete(const HashObj & x)
{
int curPos = findPos(x);
if (!isActive(curPos))
return false;
_Lists[curPos].state = EMPTY;
currSize--;
return true;
}
template<typename HashObj>
void HashTable<HashObj>::rehash()
{
std::vector<HashEntry> old_Lists = _Lists;
_Lists.resize(old_Lists.size() * 2);
for (int i = 0; i < old_Lists.size(); i++)
{
EntryType state = old_Lists[i].state;
HashObj elem = old_Lists[i].elem;
if (state == ACTIVE)
{
Insert(elem);
}
}
}