字符串查找算法：bm算法

今天有空，认真的对比了一下经典的字符串查找算法BM算法和C库查找函数 strstr 的区别，两者各有优缺点，总结一下：

bm算法的应用场合：适合海量数据搜索，比如数据库，磁盘文件等，总之是数据量越大，性能越高；

strstr，数据量较少时，比较适合，尤其是在一个几千字节的字符串中查找不同的字符串，这时候bm被strstr甩出几条街，原因就是每次查找，bm都要建立搜索模型，而strstr立马上阵。其次，搜索字符在总字符里面出现概率较少，相同率很低的情况下，strstr优势再次展露，比如在http协议头部关键字，像Host、Get、Post、User-Agent、Accept等，这些关键字一次出现几次的概率都非常少，此时用strstr较快。

废话这么多，上代码证明一下吧，同时附加一个改进过的 bm 算法代码：

#include <stdio.h>
#include <string.h>
#include <sys/time.h>
#include <stdlib.h>

#define MAX_CHAR 128    // 键盘字符，从32-126，总共95个
#define SIZE     128
#define MAX(x, y) (x) > (y) ? (x) : (y)

void PreBmBc(char *pattern, int m, int bmBc[])
{
int i;

for (i=0; i<MAX_CHAR; i++)
{
bmBc[i] = m;
}
for (i = 0; i < m - 1; i++)
{
bmBc[pattern[i]] = m - 1 - i;
}
}

void suffix_old(char *pattern, int m, int suff[])
{
int i, j;

suff[m - 1] = m;

for(i = m - 2; i >= 0; i--)
{
j = i;
while(j >= 0 && pattern[j] == pattern[m - 1 - i + j]) j--;

suff[i] = i - j;
}
}

void suffix(char *pattern, int m, int suff[])
{
int f, g, i;

suff[m - 1] = m;
g = m - 1;
for (i = m - 2; i >= 0; --i)
{
if (i > g && suff[i + m - 1 - f] < i - g)
suff[i] = suff[i + m - 1 - f];
else
{
if (i < g)
g = i;
f = i;
while (g >= 0 && pattern[g] == pattern[g + m - 1 - f])
--g;
suff[i] = f - g;
}
}
}

void PreBmGs(char *pattern, int m, int bmGs[])
{
int i, j;
int suff[SIZE];

// 计算后缀数组
suffix(pattern, m, suff);

// 先全部赋值为m，包含Case3
for(i = 0; i < m; i++)
{
bmGs[i] = m;
}

// Case2
j = 0;
for(i = m - 1; i >= 0; i--)
{
if(suff[i] == i + 1)
{
for(; j < m - 1 - i; j++)
{
if(bmGs[j] == m)
bmGs[j] = m - 1 - i;
}
}
}

// Case1
for(i = 0; i <= m - 2; i++)
{
bmGs[m - 1 - suff[i]] = m - 1 - i;
}

//  print(bmGs, m, "bmGs[]");
}

const char* BM_strstr(char *pattern, int patternLen, const char *text, int textLen)
{
int i, j, bmBc[MAX_CHAR], bmGs[SIZE];

// Preprocessing
PreBmBc(pattern, patternLen, bmBc);
PreBmGs(pattern, patternLen, bmGs);

// Searching
j = 0;
while(j <= textLen - patternLen)
{
for(i = patternLen - 1; i >= 0 && pattern[i] == text[i + j]; i--);
if(i < 0)
{
//printf("Find it, the position is %d\n", j);
return (text + j);
}
else
{
j += MAX(bmBc[text[i + j]] - patternLen + 1 + i, bmGs[i]);
}
}

//printf("No find.\n");
return NULL;
}

const char* line_strstr(const char* src, int srcLen, const char* dst, int dstLen)
{
const char* pline = src;
const char* pEnd = src + srcLen;

while(pline < pEnd)
{
if (*pline == '\n' || *pline == '\r')
{
++pline;
while (*pline == '\n' || *pline == '\r' || *pline == ' ' || *pline == '\t') ++pline;
if (memcmp(pline, dst, dstLen) == 0)
{
return pline;
}
}
else
{
++pline;
}
}
return NULL;
}

const char* sun_strstr(const char *text, int textLen, const char *patt, int pattLen)
{
unsigned int temp[256];
unsigned int *shift = temp;

int i;

for( i=0; i < 256; i++ )
{
*(shift+i) = pattLen + 1;
}
for( i=0; i < pattLen; i++ )
{
*(shift + (unsigned char)(*(patt+i))) = pattLen-i;
}

//shift['s']=6 步,shitf['e']=5 以此类推
size_t limit = textLen - pattLen+1;
for(i=0; i < limit; i += shift[ text[i+pattLen] ])
{
if( text[i] == *patt )
{
const char *match_text = text + i + 1;
size_t match_size = 1;
do
{
// 输出所有匹配的位置
if( match_size == pattLen )
{
return (text+i);
}
}while((*match_text++) == patt[match_size++]);
}
}
return NULL;
}

const char *text = "GET /qq HTTP/1.1\n\r\
Connection: keep-alive\n\r\
Accept: image/webp,*/*;q=0.8\n\r\
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) \
Chrome/41.0.2272.89 Safari/537.36\n\r\
Referer: \ http://www.jd.com/?cu=true&utm_source=p.yiqifa.com&utm_medium=tuiguang&utm_campaign=t_1_792977&utm_term=2d74c473e39447508534846d9f847e2f\n\r\ Accept-Encoding: gzip, deflate, sdch \n\r\
Host: jcm.jd.com \n\r\
Accept-Language: zh-CN,zh;q=0.8\n\r";

int main()
{
#define TEST_CNT    1000000
char pattern[256] = {"Host"};
int patLen = strlen(pattern);
int textLen = strlen(text);

printf("patLen=%d, textLen=%d\n", patLen, textLen);
int i = 0;

struct timeval t;
gettimeofday(&t, 0);
unsigned long long a = t.tv_sec * 1000000ULL + t.tv_usec;
for (i=0; i<TEST_CNT; ++i)
{
const char* pstr = BM_strstr(pattern, patLen, text, textLen);
//if (pstr)
//printf("%c%c%c%c\n", pstr[0], pstr[1], pstr[2], pstr[3]);
if (!pstr)
{
printf("error\n");
exit(1);
}
}
gettimeofday(&t, 0);
unsigned long long b = t.tv_sec * 1000000ULL + t.tv_usec;
printf("BM_strstr: time=%lld\n", b-a);

gettimeofday(&t, 0);
a = t.tv_sec * 1000000ULL + t.tv_usec;
for (i=0; i<TEST_CNT; ++i)
{
const char* pstr = strstr(text, pattern);
//if (pstr)
//printf("%c%c%c%c\n", pstr[0], pstr[1], pstr[2], pstr[3]);
if (!pstr)
{
printf("error, i=%d\n", i);
exit(1);
}
}
gettimeofday(&t, 0);
b = t.tv_sec * 1000000ULL + t.tv_usec;
printf("strstr: time=%lld\n", b-a);

gettimeofday(&t, 0);
a = t.tv_sec * 1000000ULL + t.tv_usec;
for (i=0; i<TEST_CNT; ++i)
{
const char* pstr = line_strstr(text, textLen, pattern, patLen);
//if (pstr)
//printf("%c%c%c%c\n", pstr[0], pstr[1], pstr[2], pstr[3]);
if (!pstr)
{
printf("error, i=%d\n", i);
exit(1);
}
}
gettimeofday(&t, 0);
b = t.tv_sec * 1000000ULL + t.tv_usec;
printf("linestrstr: time=%lld\n", b-a);

gettimeofday(&t, 0);
a = t.tv_sec * 1000000ULL + t.tv_usec;
for (i=0; i<TEST_CNT; ++i)
{
const char* pstr = sun_strstr(text, textLen, pattern, patLen);
//if (pstr)
//printf("%c%c%c%c\n", pstr[0], pstr[1], pstr[2], pstr[3]);
if (!pstr)
{
printf("error, i=%d\n", i);
exit(1);
}
}
gettimeofday(&t, 0);
b = t.tv_sec * 1000000ULL + t.tv_usec;
printf("SUNDAY: time=%lld\n", b-a);

return 0;
}