字符串匹配算法

-：Sunday算法

最后一个是Sunday 算法，实际上比Boyer-Moore 还快，呵呵。长江后浪推前浪。

原始论文：Daniel M. Sunday, A very fast substring search algorithm, Communications of the ACM, v.33 n.8, p.132-142, Aug. 1990

看原始论文的题目，D.M. Sunday 貌似是故意想气气Boyer-Moore 两位大牛似的。呵呵。不过实际上的确Sunday 算法的确比BM 算法要快，而且更简单。

Sunday 的算法思想和Horspool 有些相似，但是。当出现不匹配的时候，却不是去找匹配串中不匹配的字符在模式串的位置，而是直接找最右边对齐的右一位的那个字符在模式串的位置。

比如：

匹配串：abcbc zdxzc

模式串：zbcac

恩，这里我们看到b-a 没有对上，我们就看匹配串中的z 在模式串的位置，然后，嘿嘿。

匹配串：abcbczdxzc

模式串：     zbcac

如果模式串中的没有那个字符怎么办呢？很简单，跳过去呗。

匹配串：abcbc edxzcs

模式串：zbcac

e 不在模式串中出现

那么我们就

匹配串：abcbcedxzcs

模式串：      zbcac

int findStr(const char* text,const char* pattern)
{
    if(text==NULL||pattern==NULL)
        return -1;
    int i,j,temp;
    int textLen=strlen(text);
    int patternLen=strlen(pattern);
    int size=1<<(sizeof(char)*8);
    int* moveTable=new int[size];

    //precompute the moving table, pattern will move  'patternLen+1' steps by default
    for(i=0;i<size;i++)
        moveTable[i]=patternLen+1;
    for(i=0;i<patternLen;i++)
        moveTable[pattern[i]]=patternLen-i;

    //compare the text and pattern backwards
    for(i=patternLen-1;i<textLen;)
    {
        j=patternLen-1;
        temp=i;
        while(pattern[j]==text[temp]&&j>=0)
        {
            j--;
            temp--;
        }
        if(j<0)
            return temp+1;
        i+=moveTable[text[i+1]];
    }
    return -1;
}

int main()
{
    cout << "Hello World!" << endl;
    const char* text="we call on the western nations to commit more money to the poorest country";
    const char* pattern="more";
    cout<<findStr(text,pattern);
    return 0;
}

 1 import redis
  2 import random
  3
  4 def fillDB(conn):
  5         conn.execute_command('CONFIG SET maxmemory-policy noeviction')
  6         count=0
  7         try:
  8                 while 1:
  9                         conn.execute_command('SET user:%d super_hero'%count )
 10                         count+=1
 11         except Exception as e:
 12                 print e
 13                 conn.execute_command('CONFIG SET maxmemory-policy allkeys-lru')
 14                 return count
 15
 16 def simulate(conn,dbSize,dataPro,probability,times):
 17         '''
 18         conn--------------the redis client instance
 19         dbSize------------the size of the DB,actually it's the number of string stored in it
 20         dataPro-----------the proportion of data that are frequently visited
 21         probability-------the probability with which the frequent data will be accessed
 22         times-------------the number of simulation times
 23         '''
 24         dataAmount=dbSize*2
 25         item=-1
 26         for i in range(0,times):
 27                 visitPro = random.random()
 28                 if visitPro<=probability:
 29                         item=random.randint(0,dataAmount*dataPro)
 30                 else:   
 31                         item=random.randint(dataAmount*dataPro+1,dataAmount)
 32                 
 33                 if not conn.execute_command('GET user:%d'%item):
 34                         conn.execute_command('SET user:%d super_hero'%item)
 35                         
 36 if __name__=='__main__':
 37         print 'before running this,please make sure that you have limited the maxmemory'
 38         conn = redis.StrictRedis()
 39         '''
 40         cacheSize = fillDB(conn)
 41         print 'the cacheSize is %d'%cacheSize
 42         simulate(conn,cacheSize,0.2,0.8,10000)
 43 '''     
 44         info = conn.execute_command('info')
 45         info = info.split('\r\n')
 46         
 47         hits=0
 48         misses=0
 49         for s in info:
 50                 if 'keyspace_hits' in s:
 51                         hits = int(s[len('keyspace_hits')+1:len(s)])    
 52                 elif 'keyspace_misses' in s:
 53                         misses = int(s[len('keyspace_misses')+1:len(s)])
 54                         break
 55
 56         print 'keyspace_hits = %d\n'%hits,'keyspace_misses = %d\n'%misses
 57         print 'hit_rate = %f'%(hits/float(hits+misses))