STL 算法:<algorithm> 的威力
掌握最重要的 STL 算法──从搜索和排序到变换和累加。学习算法如何与迭代器组合以编写富有表达力且高效的代码。
<algorithm> 头文件
头文件包含超过 100 个在迭代器范围上操作的泛型算法。这些算法凝聚了数十年的计算机科学成果──经过实战检验、高度优化且正确可靠。使用它们代替手写循环可以使你的代码更简短、更有表达力,通常也更快。
算法按类别组织:
- 非修改类:find、count、search、all_of/any_of/none_of、equal、mismatch
- 修改类:copy、transform、fill、generate、replace、remove
- 排序与有序类:sort、stable_sort、partial_sort、nth_element
- 二分查找类:lower_bound、upper_bound、binary_search、equal_range
- 数值类(在 中):accumulate、inner_product、partial_sum、iota
每个算法都以迭代器范围 [first, last) 作为主要输入。
非修改算法
非修改算法在不改变元素的情况下检查它们。它们是查询和验证的主力工具。
#include <algorithm>
#include <vector>
#include <string>
#include <iostream>
int main() {
std::vector<int> nums = {3, 1, 4, 1, 5, 9, 2, 6, 5};
// find — returns iterator to first match (or end)
auto it = std::find(nums.begin(), nums.end(), 5);
if (it != nums.end()) {
std::cout << "Found 5 at index " << std::distance(nums.begin(), it) << '\n';
}
// find_if — with a predicate
auto even = std::find_if(nums.begin(), nums.end(),
[](int x) { return x % 2 == 0; });
std::cout << "First even: " << *even << '\n'; // 4
// count / count_if
int fives = std::count(nums.begin(), nums.end(), 5);
int odds = std::count_if(nums.begin(), nums.end(),
[](int x) { return x % 2 != 0; });
std::cout << "Fives: " << fives << ", Odds: " << odds << '\n';
// all_of / any_of / none_of
bool all_positive = std::all_of(nums.begin(), nums.end(),
[](int x) { return x > 0; });
bool has_negative = std::any_of(nums.begin(), nums.end(),
[](int x) { return x < 0; });
std::cout << std::boolalpha;
std::cout << "All positive: " << all_positive << '\n'; // true
std::cout << "Has negative: " << has_negative << '\n'; // false
// equal — compare two ranges
std::vector<int> nums2 = {3, 1, 4, 1, 5, 9, 2, 6, 5};
bool same = std::equal(nums.begin(), nums.end(), nums2.begin());
std::cout << "Equal: " << same << '\n'; // true
}修改算法
修改算法改变范围内的元素值或重新排列元素。
#include <algorithm>
#include <vector>
#include <iostream>
#include <iterator>
int main() {
std::vector<int> src = {1, 2, 3, 4, 5};
// transform — apply a function to each element
std::vector<int> squared;
std::transform(src.begin(), src.end(), std::back_inserter(squared),
[](int x) { return x * x; });
// squared: {1, 4, 9, 16, 25}
// transform with two input ranges (binary transform)
std::vector<int> b = {10, 20, 30, 40, 50};
std::vector<int> sums;
std::transform(src.begin(), src.end(), b.begin(),
std::back_inserter(sums),
[](int a, int b) { return a + b; });
// sums: {11, 22, 33, 44, 55}
// fill and generate
std::vector<int> zeros(10);
std::fill(zeros.begin(), zeros.end(), 42); // all 42s
int counter = 0;
std::generate(zeros.begin(), zeros.end(),
[&counter]() { return counter++; });
// zeros: {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
// remove-erase idiom (pre-C++20)
std::vector<int> data = {1, 2, 3, 2, 4, 2, 5};
auto new_end = std::remove(data.begin(), data.end(), 2);
data.erase(new_end, data.end()); // actually shrinks the vector
// data: {1, 3, 4, 5}
// C++20: simpler
std::erase(data, 3);
// data: {1, 4, 5}
// replace
std::replace(data.begin(), data.end(), 4, 99);
// data: {1, 99, 5}
for (int x : data) std::cout << x << ' ';
std::cout << '\n';
}排序与有序算法
STL 提供了多种具有不同权衡的排序算法。std::sort 是通用的主力工具(平均 O(n log n),内省排序),而 partial_sort 和 nth_element 在你只需要部分排序时更高效。
#include <algorithm>
#include <vector>
#include <string>
#include <iostream>
struct Student {
std::string name;
double gpa;
};
int main() {
std::vector<int> nums = {5, 2, 8, 1, 9, 3, 7, 4, 6};
// sort — O(n log n), not stable
std::sort(nums.begin(), nums.end());
// nums: {1, 2, 3, 4, 5, 6, 7, 8, 9}
// sort descending with comparator
std::sort(nums.begin(), nums.end(), std::greater<int>{});
// nums: {9, 8, 7, 6, 5, 4, 3, 2, 1}
// stable_sort — preserves relative order of equal elements
std::vector<Student> students = {
{"Alice", 3.8}, {"Bob", 3.5}, {"Charlie", 3.8}, {"Diana", 3.5}
};
std::stable_sort(students.begin(), students.end(),
[](const Student& a, const Student& b) {
return a.gpa > b.gpa; // descending GPA
});
// Alice before Charlie (both 3.8) — relative order preserved
// partial_sort — sort only the first k elements
std::vector<int> data = {5, 2, 8, 1, 9, 3};
std::partial_sort(data.begin(), data.begin() + 3, data.end());
// First 3 are sorted: {1, 2, 3, ...rest unspecified}
// nth_element — partition so the nth element is in its sorted position
std::vector<int> vals = {5, 2, 8, 1, 9, 3, 7};
std::nth_element(vals.begin(), vals.begin() + 3, vals.end());
// vals[3] == 5 (the median), left side <= 5, right side >= 5
// lower_bound / upper_bound — binary search on sorted data
std::vector<int> sorted = {1, 2, 4, 4, 4, 7, 9};
auto lo = std::lower_bound(sorted.begin(), sorted.end(), 4); // first 4
auto hi = std::upper_bound(sorted.begin(), sorted.end(), 4); // past last 4
std::cout << "Count of 4s: " << std::distance(lo, hi) << '\n'; // 3
}数值算法(<numeric>)
头文件提供了用于对范围进行数学计算的算法。
#include <numeric>
#include <vector>
#include <iostream>
#include <functional>
int main() {
std::vector<int> nums = {1, 2, 3, 4, 5};
// accumulate — fold/reduce with initial value
int sum = std::accumulate(nums.begin(), nums.end(), 0);
std::cout << "Sum: " << sum << '\n'; // 15
// accumulate with custom operation (product)
int product = std::accumulate(nums.begin(), nums.end(), 1,
std::multiplies<int>{});
std::cout << "Product: " << product << '\n'; // 120
// inner_product — dot product
std::vector<int> weights = {2, 3, 1, 4, 2};
int dot = std::inner_product(nums.begin(), nums.end(),
weights.begin(), 0);
std::cout << "Dot product: " << dot << '\n'; // 1*2 + 2*3 + 3*1 + 4*4 + 5*2 = 37
// partial_sum — running total
std::vector<int> running;
std::partial_sum(nums.begin(), nums.end(), std::back_inserter(running));
// running: {1, 3, 6, 10, 15}
// iota — fill with incrementing values
std::vector<int> seq(10);
std::iota(seq.begin(), seq.end(), 1); // {1, 2, 3, ..., 10}
// C++17: reduce (parallel-friendly version of accumulate)
// int sum2 = std::reduce(nums.begin(), nums.end()); // same as accumulate for +
for (int x : running) std::cout << x << ' ';
std::cout << '\n';
}std::remove 并不真正从容器中移除元素。它将「保留」的元素移到前面,并返回新逻辑末尾的迭代器。容器的实际大小不变──你必须调用 erase() 来物理移除尾部元素:
std::vector<int> v = {1, 2, 3, 2, 4};
auto new_end = std::remove(v.begin(), v.end(), 2);
// v is now: {1, 3, 4, ?, ?} with size still 5
v.erase(new_end, v.end()); // NOW size is 3在 C++20 中,std::erase(container, value) 和 std::erase_if(container, predicate) 一次调用即可处理这个惯用法。在可用时优先使用它们。
算法在组合使用时大放异彩。不要写复杂的循环,而是将算法链接在一起:
// Get sorted unique elements
std::sort(v.begin(), v.end());
v.erase(std::unique(v.begin(), v.end()), v.end());
// Copy only even numbers, squared, to output
std::vector<int> result;
std::copy_if(v.begin(), v.end(), std::back_inserter(result),
[](int x) { return x % 2 == 0; });
std::transform(result.begin(), result.end(), result.begin(),
[](int x) { return x * x; });这种风格清晰、正确,且通常优于手写循环,因为库实现使用了 SIMD、无分支代码和其他微优化。
头文件提供了 100 多个在任何迭代器范围上工作的泛型算法std::sort是 O(n log n) 的内省排序;std::stable_sort保持相等元素的顺序std::transform是函数式的「映射」──对每个元素应用一个函数std::remove不会删除元素──使用 remove-erase 惯用法,或 C++20 的std::erase_ifstd::accumulate(来自)是通用的 fold/reduce 操作- 优先使用算法而非手写循环──它们更清晰、更安全,通常也更快
测验 — 检验你的知识
(20 经验值)1. `std::remove` 返回什么?
2. 你会使用哪个算法来检查范围中的所有元素是否都满足某个谓词?
3. `std::nth_element` 的时间复杂度是多少?