KMP之计算Next数组

KMP的Next数组：模式串的前缀与后缀的“相交”长度

KMP算法步骤：

1.先算next数组

2.若失配（此时模式串下标为j），利用Next数组求出失配后滑动的新位置

　　a.Next[j] \geq 0，则模式串右移 j-Next[j] 位

　　b.Next[j] = -1，则模式串右移 j+1 位

初始 i = 0, j = -1，Next[ 0 ] = -1（为了后面计算方便）;

Step1:

　　由于j == -1，i&j++，且Next[i] = j，即Next[1] = 0; 此时 i = 1, j = 0;

　　下一个循环：p[ i ] != p[ j ]，所以 j = Next[ j ] = Next[ 0 ] = -1，j回退

　　j == -1， i&j++，此时 i = 2, j = 0; Next[ i ] = j，即Next [ 2 ] = 0

Step2:

　　p [ 2 ] == p [ 0 ] -> i&j++ -> i = 3, j = 1, Next[ 3 ] = 1;

Step3:

　　p [ 3 ] == p [ 1 ] -> i&j++ -> i = 4, j = 2, Next[4] = 2;

Step4:

　　p [ 4 ] == p [ 2 ] -> i&j++ -> i = 5, j = 3, Next[5] = 3;

Step5:

　　p [ 5 ] == p [ 3 ] -> i&j++ -> i = 6, j = 4, Next[6] = 4;

Step6:

　　p [ 6 ] != p [ 4 ] -> 失配！ -> j = Next[ j ] = Next [ 4 ] = 2

Step7:

　　p [ 6 ] != p [ 2 ] -> 失配！ -> j = Next[ j ] = Next [2 ] = 0

Step8:

　　p [ 6 ] != p [ 0 ] -> 失配！ -> j = Next[ j ] = Next [0 ] = -1

Step9:

　　j == -1 -> i&j++ -> i = 7, j = 0 -> Next[7] = 0

Step10:

　　p [ 7 ] != p [ 0 ] -> 失配！ -> j = Next[ j ] = Next [0 ] = -1

Step11:

　　i !< strlen(p) -> 循环退出

int next[100] = { 0 };

void calNext(char *p, int *next)
{
    next[0] = -1;
    int i = 0, j = -1;
    while (i < strlen(p)) {
        if (j == -1 || p[i] == p[j])    //短路
        {
            i++;
            j++;
            next[i] = j;
        }
        else    //j指针回退过程
            j = next[j];
    }
}