Question

最近我遇到了这段代码片段：

// look for element which is the smallest max element from
// a given iterator
int diff = std::numeric_limits<int>::max();
auto it = nums.rbegin();
auto the_one = nums.rbegin();

for (; it != given; ++it) // this terminates
{
  int local_diff = *it - *given;
  // if the element is less than/equal to given we are not interested
  if (local_diff <= 0)
     continue;

  if (local_diff < diff)
  {
     // this update the global diff
     diff = local_diff;
     the_one = it;
  }
}

我想知道是否有人可以想到一个优雅的stl算法来取代上面的。基本上我们必须经历所有元素，并跟踪我们需要的元素。这与std::max_element不相似（至少我不能那样建模）。

Answer 1

auto the_one = std::min_element(nums.rbegin(), given,
  [given](int a, int b) {
    bool good_a = a > *given;
    bool good_b = b > *given;
    return (good_a && good_b) ? a < b : good_a;
  });

诀窍是编写一个比较函数，声明任何＆＃34; good＆＃34;元素（一个大于*given的元素）要小于任何＆＃34;不好的＆＃34;元件。两个＆＃34;好＆＃34;元素通常比较;两个＆＃34;坏＆＃34;元素总是声明等价。

STL算法用于最小的max元素而不是给定的值

1 个答案: