如何连接两个现有范围::视图? [英] How to concat two existing ranges::view?

问题描述

我想使用现有视图进行串联.在代码中:

I want to use an existing view for concatenation. In code:

auto rng = view::empty<vector<int>>();

for(int i{0}; i < 5; ++i)
{
    vector<int> const & v{foo()}; // returns a reference
    rng |= view::concat(v); // doesn't compile - error: no viable overloaded '|='
};

换句话说-我如何创建多个矢量的视图，这些矢量的编号直到运行时才知道?

In other words - how can I create a view to multiple vectors whose number is not known until runtime?

推荐答案

您不能以这种方式编写视图.串联视图将产生具有不同类型的对象.您无法将其分配回原始视图，因为其类型不同.

You can't compose views this way. Concatenating a view yields an object with a different type. You can't assign it back to the original view because its type is different.

您可以结合使用view::cycle(采用一个范围并无限重复)和view::take(采用范围的第一个 N 元素)来获得想要的效果.

You can get the effect you're after with a combination of view::cycle (takes one range and repeats it infinitely), and view::take (takes the first N elements of a range).

vector<int> const & v{foo()}; // returns a reference
auto rng = v | view::cycle | view::take(5 * v.size());

编辑

如果foo()每次都可以返回对不同向量的引用，那么除了view::take之外，您还可以使用view::generate和view::join:

If foo() can return a reference to a different vector each time, then you can use view::generate and view::join, in addition to view::take:

auto rng = view::generate(foo) | view::take(5) | view::join;

这篇关于如何连接两个现有范围::视图?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！