大多数惯用的`[a - > a] - > （a - > a）` [英] Most idiomatic implementation of `[a -> a] -> (a -> a)`

问题描述

如果我有一个函数列表，每个类型 a - >一个类型的，结合它们的最短，最优雅和最习惯的方式是什么？最好不添加额外的依赖项？

一些变体包括：

$ $ $ $ $ $ $ $ $ x：xs）= x。 （foo xs）
foo [] = id

和

  foo = foldr（。）id 
  

和

  foo = appEndo。 mconcat。 map Endo 
  

但由于某种原因，我期望找到更好的东西。

解决方案

我会说你不会击败

  comp = foldr（。）id 
  

为什么？那么我们有一个列表，我们试图以正确的关联方式来减少它。

如果您查看和的实现， sum ，最大值以及类似的，你会发现这是它们在标准库中的实现方式，我认为你不会比这更习惯于使用它：）

切线：我不愿意添加注释中提到的 foldr1 变体，因为我会说这是意外行为不是说最大值，这显然是必须的。

If I have a list of functions, each of the type a -> a for some type, what is the most shortest, elegant and idiomatic way to combine them; preferably without adding extra dependencies?

Some variants include

foo (x:xs) = x . (foo xs)
foo [] = id

and

foo = foldr (.) id

and

foo = appEndo . mconcat . map Endo

but for some reason I’m expecting to find something nicer.

解决方案

I'd say you're not going to beat

comp = foldr (.) id

Why? Well we have a list of things and we're trying to reduce it in a right associative way.

If you look at the implementations of and, sum, maximum and similar, you'll see that this is how they're implemented in the standard library, I don't think you get more idiomatic than that :)

Tangent: I hesitate to add the foldr1 variant mentioned in comments because I'd say that it's unexpected behavior for this to be partial, unlike say maximum where it clearly must be.

这篇关于大多数惯用的`[a - > a] - > （a - > a）`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！