最近網上看KMP算法,看了不少做者寫的文章,後來發現看不明白,貌似哪裏不正確,把代碼拷下來運行發現也有問題,致使誤導了不少人,我先舉幾個例子:html
def kmp(string, match): n = len(string) m = len(match) i = 0 j = 0 count_times_used = 0 while i < n: count_times_used += 1 if match[j] == string[i]: if j == m - 1: print "Found '%s' start at string '%s' %s index position, find use times: %s" % (match, string, i - m + 1, count_times_used,) return i += 1 j += 1 elif j > 0: j = j - 1 else: i += 1
代碼很簡潔,可是輸入:python
kmp("abcdeffg", "abcdefg")
明顯匹配不到,可是仍是會獲得結果。算法
拜讀《算法導論》後,纔將疑惑解除。code
下面是正確的代碼:htm
#include <stdio.h> #include <string.h> #include <stdlib.h> #define MAX_LEN 100 int * getnext(const char * pattern); int main() { char string[MAX_LEN]; char pattern[MAX_LEN]; int *next; scanf("%s", string); scanf("%s", pattern); int slen = strlen(string); int plen = strlen(pattern); if(plen > slen || slen > MAX_LEN || plen > MAX_LEN) { exit(0); } next = getnext(string); int q = -1; for(int i = 0; i < slen; i++) { while(q >= 0 && pattern[q+1] != string[i]) { q = next[q]; } if(pattern[q+1] == string[i]) { q++; } if(q == plen - 1) { printf("pattern occurs with shift %d\n", i + 1 - plen); q = next[q]; } } return 1; } int * getnext(const char * pattern) { int plen = strlen(pattern); int *next = (int *)malloc(MAX_LEN * sizeof(int)); next[0] = -1; int k = -1; for(int j = 1; j < plen; ++j) { while(k >= 0 && pattern[j] != pattern[k+1]) { k = next[k]; } if(pattern[k+1] == pattern[j]) { k++; } next[j] = k; } return next; }
運行結果:blog
abcdefghijkkkkkkkkkllllll k pattern occurs with shift 10 pattern occurs with shift 11 pattern occurs with shift 12 pattern occurs with shift 13 pattern occurs with shift 14 pattern occurs with shift 15 pattern occurs with shift 16 pattern occurs with shift 17 pattern occurs with shift 18