【LeetCode】187. Repeated DNA Sequences

时间:2022-03-12 18:08:00

题目:

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"].

提示:

这道题最初的想法就是使用HashMap来做,一开始的时候用了STL里的unordered_map:

class Solution {
public:
vector<string> findRepeatedDnaSequences(string s) {
unordered_set<string> string_set;
if (s.length() <= ) {
return {};
}
unordered_map<string ,int> m;
for (int i = ; i < s.length() - ; ++i) {
string tmp = s.substr(i, );
if (m.find(tmp) == m.end()) {
m[tmp] = ;
} else {
string_set.insert(tmp);
}
}
return vector<string>(string_set.begin(), string_set.end());
}
};

做下来以后虽然AC了,但是时间却很不理想,原因也很简单,对于这种特定的问题,使用自定义的map和hash方法会让算法的效率得到很大的提高,之于如何设计map和hash的问题,我们不妨先看一看这个问题输入所具有的特点:

  • 输入的字符串长度是固定的(10个字符)
  • 字符只有{'A', 'C', 'T', 'G'}四种
  • 结合上述两点,可以发现输入的可能性是可以计算出来的,即4^10个。

这样一来,我们就可以利用一个基本类型的数组作为map,而hash的算法则是可以想办法把4种字符分别映射到[1,4]这4个数字上,具体的做法可以看代码。

代码:

class Solution {
public:
vector<string> findRepeatedDnaSequences(string s) {
// 将该char数组作为一个map
char map[] = {};
int len = s.length(), num = ;
vector<string> res;
if (len <= ) {
return res;
}
for (int i = ; i < ; ++i) {
num <<= ;
// 可以试一下下面的计算方法,效果就是把4个字符映射到了1,2,3,4这四个数字上
num |= (s[i] - 'A' + ) % ;
}
for (int i = ; i < s.length(); ++i) {
num <<= ;
num |= (s[i] - 'A' + ) % ;
// 0xfffff代表了20个1组成的二进制序列,按位与之后,结果就是当前字符串代表的数值
num = num & 0xfffff;
if (map[num]++ == ) {
res.push_back(s.substr(i - , ));
}
}
return res;
}
};