摘自网络：http://blog.minidx.com/2008/01/27/446.html，添加了些注释。
哈希算法将任意长度的二进制值映射为固定长度的较小二进制值，这个小的二进制值称为哈希值。哈希值是一段数据唯一且极其紧凑的数值表示形式。如果散列一段明文而且哪怕只更改该段落的一个字母，随后的哈希都将产生不同的值。要找到散列为同一个值的两个不同的输入，在概率上是极其微小的，所以数据的哈希值可以检验数据的完整性。
链表查找的时间效率为O(N)，二分法为log2N，B+Tree为log2N，但Hash链表查找的时间效率为O(1)。
设计高效算法往往需要使用Hash链表，常数级的查找速度是任何别的算法无法比拟的，Hash链表的构造和冲突的不同实现方法对效率当然也会有一定的影响，然而Hash函数是Hash链表最核心的部分，下面是几款经典软件中使用到的字符串Hash函数实现，通过阅读这些代码，我们可以在Hash算法的执行效率、离散性、空间利用率等方面有比较深刻的了解。
下面分别介绍几个经典软件中出现的字符串Hash函数。
// 推荐
// PHP 中出现的字符串Hash函数
static unsigned long hashpjw(char *arKey, unsigned int nKeyLength)
{
// check
if ((arKey == NULL) || (nKeyLength <= 0)) return 0;
unsigned long h = 0;
unsigned long g = 0;
char *arEnd = arKey+nKeyLength;
while (arKey < arEnd)
{
h = (h << 4) + *arKey++;
if ((g = (h & 0xF0000000))) // 处理溢出值
{
h = h ^ (g >> 24);
h = h ^ g;
}
}
return h;
}
// 推荐
// OpenSSL 中出现的字符串Hash函数
unsigned long lh_strhash(char *str)
{
// check
if (str == NULL) return 0;
unsigned long ret = 0;
int i, len;
unsigned short *sTemp;
len = (strlen(str)+1)/2;
sTemp = (unsigned short *)str;
for (i=0; i<len; i++)
ret ^= (sTemp[i]<<(i&0x0f)); // 1:异或^，没有溢出；最大左移16位，也没有溢出；
return ret;
}
/* The following hash seems to work very well on normal text strings
* no collisions on /usr/dict/words and it distributes on %2^n quite
* well, not as good as MD5, but still good.
*/
unsigned long lh_strhash(const char *s)
{
// check
if ((s == NULL) || (*s == ' ')) return 0;
unsigned long ret = 0;
unsigned long v;
int r;
long n = 0x100;
while (*s)
{
v = n|(*s);
r = (int)((v>>2)^v)&0x0f;
ret = (ret(32-r)); // vc6上编译不能通过
ret &= 0xFFFFFFFFL;
ret ^= v*v;
n += 0x100;
s ++;
}
return ((ret>>16)^ret);
}
// MySql中出现的字符串Hash函数
// Mysql中对字符串Hash函数还区分了大小写
#ifndef NEW_HASH_FUNCTION
/* Calc hashvalue for a key */
static unsigned int calc_hashnr(const byte *key, unsigned int length)
{
register unsigned int nr=1, nr2=4;
while (length--)
{
nr ^= (((nr&63)+nr2) * ((unsigned int)(unsigned char)*key++)) + (nr << 8); // 没看明白。当字符串很长时，nr2值较大，乘法可能溢出
nr2 += 3;
}
return ((unsigned int)nr);
}
/* Calc hashvalue for a key, case indepenently */
static unsigned int calc_hashnr_caseup(const byte *key,unsigned int length)
{
register unsigned int nr=1, nr2=4;
while (length--)
{
nr ^= (((nr&63)+nr2) * ((unsigned int)(unsigned char)toupper(*key++))) + (nr << 8);
nr2 += 3;
}
return((unsigned int) nr);
}
#else
/*
* Fowler/Noll/Vo hash
*
* The basis of the hash algorithm was taken from an idea sent by email to the
* IEEE Posix P1003.2 mailing list from Phong Vo (kpv@research.att.com) and
* Glenn Fowler (gsf@research.att.com). Landon Curt Noll (chongo@toad.com)
* later improved on their algorithm.
*
* The magic is in the interesting relationship between the special prime
* 16777619 (2^24 + 403) and 2^32 and 2^8.
*
* This hash produces the fewest collisions of any function that we've seen so
* far, and works well on both numbers and strings.
*/
unsigned int calc_hashnr(const byte *key, unsigned int len)
{
unsigned int hash = 0;
const byte *end = key+len;
for (hash = 0; key < end; key++)
{
hash *= 16777619; // magic number // 会有溢出
hash ^= (unsigned int)*(unsigned char*)key;
}
return (hash);
}
unsigned int calc_hashnr_caseup(const byte *key, unsigned int len)
{
unsigned int hash = 0;
const byte *end = key+len;
for (hash = 0; key < end; key++)
{
hash *= 16777619; // magic number // 会有溢出
hash ^= (unsigned int)(unsigned char)toupper(*key);
}
return (hash);
}
#endif
//
// 另一个经典字符串Hash函数
unsigned int hash(char *str)
{
register unsigned int h=0;
register unsigned char *p = (unsigned char *)str;
for(h=0; *p; p++)
h = 31 * h + *p; // 未考虑溢出的影响
return h;
}