[LeetCode] Repeated DNA Sequences

Repeated DNA Sequences

 

All DNA is composed of a series of nucleotides abbreviated as A, C, G, and T, for example: "ACGAATTCCG". When studying DNA, it is sometimes useful to identify repeated sequences within the DNA.

Write a function to find all the 10-letter-long sequences (substrings) that occur more than once in a DNA molecule.

For example,

Given s = "AAAAACCCCCAAAAACCCCCCAAAAAGGGTTT",Return:["AAAAACCCCC", "CCCCCAAAAA"].

解题思路：

1、用map存储已经扫描过的子串，并对之计数。时间复杂度为O(n)。代码如下：

class Solution {public:    vector<string> findRepeatedDnaSequences(string s) {        map<string, int> count;                vector<string> result;                int len = s.length();        for(int i=0; i<len-10; i++){            string str = s.substr(i, 10);            map<string, int>::iterator it=count.find(str);            if(it!=count.end()){                count[str]=1;            }else{                if(it->second==1){                    result.push_back(str);                }                count[str]++;            }        }                return result;    }};

但是会报内存溢出错误。

2、为AGCT分别编码，共有4种，故只需两位便能编码。共十个字符，只需20位就能表示任何一种组合。int类型32位，因此可以用一个int类型来存储10个字符串。每检查一个字符之后，需要将最高位置为0。代码如下：

class Solution {public:    vector<string> findRepeatedDnaSequences(string s) {        const int subStrLen = 10;        const int mask = 0x3ffff;                map<int, int> count;                map<char, int> cCode;        cCode['A']=0;        cCode['C']=1;        cCode['G']=2;        cCode['T']=3;                vector<string> result;                int len = s.length();                int code=0;                if(len>subStrLen){            string str = s.substr(0, subStrLen);            for(int i=0; i<subStrLen; i++){                code <<= 2;                code |= cCode[str[i]];            }            count[code] = 1;        }                for(int i=subStrLen; i<len; i++){            code &= mask;    //清空最高位            code <<= 2;            code |= cCode[s[i]];            count[code]++;            if(count[code]==2){                result.push_back(s.substr(i-subStrLen + 1, subStrLen));            }        }                return result;    }};

在leetCode中，我跑上面的代码需要269ms，看了很多其他人的时间，都比我快很多。不知各位有更好的方法没有。