#include <map>
map<string,int> dict;
map是基于红黑树实现的,可以快速查找一个元素是否存在,是关系型容器,能够表达两个数据之间的映射关系。
dict.insert(make_pair("abc",1));
dict.count("mn"); 看看dict中含有 mn的个数,因为元素是唯一的,所以这个返回值只有两种情况,0或者1. (multi_map就不是这样啦)
dict.find("mn");如果不存在就返回 dict.end(),如果存在就返回指向该元素的 iterator
dict["pq"]++ 相当于取出 value 来进行 ++ 再存回去,下标访问不存在的元素,将导致在map中添加一个新的元素,新的键就是该下标, value的值为默认值。
所以建立一个map的时候可以:
vector<string> L;
unordered_map<string,int> dictL; for(int i = 0; i< L.size(); i++) dictL[L[i]] += 1;
删除元素: dict.eraser("abc"); 返回值为删除元素的个数
dict.eraser(itr); 删除 itr 指向的元素,并且 itr 所指向元素必须是存在的,不能是 dict.end(). 这种形式的返回值为 void
遍历: for(map<string, int>::iterator itr = dict.begin(); itr != dict.end(); itr++)
cout<< itr->first <<" " << itr->second <<endl;
map中的元素是按照健,有序的.
#include<unordered_map>
unordered_map是基于hash实现的。
unordered_map的插入、删除、查找 性能都优于 hash_map 优于 map,所以首选为 unordered_map.
它的缺陷是元素是无序的,当使用时候需要元素是有序的时候,不可以用它。
性能比较参考:http://keary.cn/?p=779
下面是它比较的代码
#include <iostream> #include <stdlib.h> #include <Windows.h> #include <map> #include <hash_map> #include <unordered_map> #include <algorithm> bool MapTest() { const unsigned int NUM_COUNT = 0xffffff; // 数组长度 const int DEFAULT_VALUE = 0; // 键值 const unsigned int NUM_RANGE = 0xffffff; // 随机数范围的最大值 int* szNumA = new int[NUM_COUNT]; int* szNumB = new int[NUM_COUNT]; srand(::GetTickCount()); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { szNumA[uiNum] = (rand() * rand()) % NUM_RANGE; szNumB[uiNum] = szNumA[uiNum]; } for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { std::swap(szNumB[uiNum], szNumB[(rand() * rand()) % NUM_COUNT]); } std::map<int, int> mapNum; std::hash_map<int, int> hMapNum; std::unordered_map<int, int> unMapNum; DWORD dwMap, dwHMap, dwUnMap; // 插入测试 dwMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { mapNum.insert(std::map<int, int>::value_type(szNumA[uiNum], 0)); } dwMap = ::GetTickCount() - dwMap; dwHMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { hMapNum.insert(std::hash_map<int, int>::value_type(szNumA[uiNum], 0)); } dwHMap = ::GetTickCount() - dwHMap; dwUnMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { unMapNum.insert(std::unordered_map<int, int>::value_type(szNumA[uiNum], 0)); } dwUnMap = ::GetTickCount() - dwUnMap; std::cout << "insert time of map is :" << dwMap << "ms" <<std::endl; std::cout << "insert time of hash_map is :" << dwHMap << "ms" <<std::endl; std::cout << "insert time of unordered_map is :" << dwUnMap << "ms" <<std::endl << std::endl; // 查找测试 dwMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { mapNum.find(szNumB[uiNum]); } dwMap = ::GetTickCount() - dwMap; dwHMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { hMapNum.find(szNumB[uiNum]); } dwHMap = ::GetTickCount() - dwHMap; dwUnMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { unMapNum.find(szNumB[uiNum]); } dwUnMap = ::GetTickCount() - dwUnMap; std::cout << "search time of map is :" << dwMap << "ms" <<std::endl; std::cout << "search time of hash_map is :" << dwHMap << "ms" <<std::endl; std::cout << "search time of unordered_map is :" << dwUnMap << "ms" <<std::endl << std::endl; // 删除测试 dwMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { mapNum.erase(szNumB[uiNum]); } dwMap = ::GetTickCount() - dwMap; dwHMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { hMapNum.erase(szNumB[uiNum]); } dwHMap = ::GetTickCount() - dwHMap; dwUnMap = ::GetTickCount(); for (unsigned int uiNum = 0; uiNum < NUM_COUNT; ++uiNum) { unMapNum.erase(szNumB[uiNum]); } dwUnMap = ::GetTickCount() - dwUnMap; std::cout << "delete time of map is :" << dwMap << "ms" <<std::endl; std::cout << "delete time of hash_map is :" << dwHMap << "ms" <<std::endl; std::cout << "delete time of unordered_map is :" << dwUnMap << "ms" <<std::endl << std::endl; delete [] szNumA; delete [] szNumB; return true; } int main(int argc, char* argv[]) { MapTest(); system("pause"); return 0; }