KMP

一直对KMP有一种朦朦胧胧的感觉。今天看了几篇博文，得以理解。
http://www.cnblogs.com/yjiyjige/p/3263858.html
https://blog.csdn.net/yutianzuijin/article/details/11954939/
https://www.cnblogs.com/tangzhengyue/p/4315393.html

在KMP算法之前是朴素查找。

//假定主串和子串都不为空
//成功返回子串在主串中第一次出现的下标，失败返回-1
int searchSub(const char *str,const char *sub)
{
    int i = 0;
    int j = 0;

    while(str[i] != '\0' && sub[j] != '\0')
    {
        if(str[i] == sub[j])
        {
            ++i;
            ++j;
        }
        else
        {
            i = i-j+1;
            j = 0;
        }
    }

    if(sub[j] == '\0')
        return i-j;
    else
        return -1;
}

KMP算法

KMP算法是对朴素的改进，朴素算法中，当子串sub[i]和主串sub[j]失配时，主串需要回退到i-j+1的位置以重新匹配，KMP算法的改进在于，它不回退主串的i，只对子串的j进行修改，因为在失配之前的i-1个主串已经和子串经过比较了，朴素算法正式忽略了这个问题。

//KMP算法通过next数组获得子串j回退的位置
int* getNext(const char *sub)  
{
    int len = strlen(sub);
    int *next = new int[len];
    int i = 0;
    int j = -1;
    next[0] = -1;

    while (i < len) 
    {
        if (j == -1 || sub[i] == sub[j]) 
            next[++i] = ++j;
        else 
            j = next[j];
    }
    return next;
}  

一个例子：

“ABAB”的next数组为{-1,0,0,1}
当str[i]和sub[j]失配后j回退到1，sub[1]等于sub[j]等于B

所以回退到1并没有任何的意义，所以直接回退到0.

//优化后的next数组
int* getNextval(const char *sub)  
{
    int len = strlen(sub);
    int *next = new int[len];
    int i = 0;
    int j = -1;
    next[0] = -1;

    while (i < len) 
    {
        if (j == -1 || sub[i] == sub[j]) 
        {//aaaaaaab
            if (sub[i+1] == sub[j+1])
                next[++i] = next[++j];
            else
                next[++i] = ++j;
        }
        else
            j = next[j];
    }
    return next;
}

//KMP
int KMP(const char *str,const char *sub)
{
    int i = 0;
    int j = 0;

    int len = strlen(sub);
    int *next = new int[len+1];
    if(next == NULL)
        exit(1);
    memset(next,0,sizeof(int)*(len+1));
    Good_getNext(sub,next,len);

    while(str[i] != '\0' && sub[j] != '\0')
    {
        if(j == -1 || str[i] == sub[j])
        {
            ++i;
            ++j;
        }
        else
            j = next[j];
    }

    if(sub[j] == '\0')
    {
        delete []next;
        return i-j;
    }
    else
    {
        delete []next;
        return -1;
    }
}