LeetCode Repeated DNA Sequences

class Solution {
private:
    char tbl[256];
public:
    vector<string> findRepeatedDnaSequences(string s) {
        vector<string> res;
        
        int len = s.size();
        if (len < 10) {
            return res;
        }
        vector<bool> exist(1<<20, false);
        vector<bool> add(1<<20, false);
        
        tbl['A'] = 0x00;
        tbl['C'] = 0X01;
        tbl['G'] = 0x02;
        tbl['T'] = 0x03;
        
        int mask= (1<<20) - 1;
        int pattern = 0;
        
        for (int i=0; i<10; i++) {
            pattern = mask & ((pattern << 2) | tbl[s[i]]);
        }
        exist[pattern] = true;
        
        for (int i=10; i<len; i++) {
            int start = i - 10 + 1;
            pattern = mask & ((pattern << 2) | tbl[s[i]]);
            if (exist[pattern] && !add[pattern]) {
                res.push_back(s.substr(start, 10));
                add[pattern] = true;
            } else {
                exist[pattern] = true;
            }
        }
        return res;
    }
};

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"].

将10个连续的DNA碱基序列看着是一个10位4进制的数，这样的数共有4^10=2^20个。用两个vector<bool>来分别表示，是否存在，是否已经添加到结果中即可。

第二轮：

简化一下用一个hashmap去存，不过发现速度下降很多，因为vector<bool>是一个特化模板实际只占一个bit的空间，相比用整数来存状态空间少很多，而且2^20个数据也就1MB个bit几百KB的空间占用，综合起来这个版本反而倒退了，卧槽：

class Solution {
public:
    vector<string> findRepeatedDnaSequences(string s) {
        // 2bit * 10 = 20bit
        unordered_map<int, int> cache;
        vector<string> res;
        
        int mappings[256];
        mappings['A'] = 0x0;
        mappings['C'] = 0x1;
        mappings['G'] = 0x2;
        mappings['T'] = 0x3;
        
        int hash = 0;
        int mask = 0x000fffff;
        
        int pos = 0;
        int len = s.size();
        while (pos < 10) {
            hash = (hash<<2) | mappings[s[pos++]];     
        }
        cache[hash]++;
        while (pos < len) {
            hash = mask & ((hash << 2) | mappings[s[pos++]]);
            if (cache[hash] > 0) {
                res.push_back(s.substr(pos - 10, 10));
                cache[hash] = -1;
            } else if (cache[hash] == 0){
                cache[hash]++;
            }
        }
        return res;
    }
};