Rust中sort_by多条件排序的冗余模式匹配如何简化？

直接用Ordering的链式方法then/then_with就能彻底消除嵌套match的冗余，这是Rust标准库专门为多条件排序设计的语法糖。

先看你可能写的冗余实现：

use std::cmp::Ordering;

#[derive(Debug, PartialEq)]
struct Player {
    name: String,
    championships: u32,
    wins: u32,
}

fn main() {
    let mut players = vec![
        Player { name: "Alice".into(), championships: 2, wins: 100 },
        Player { name: "Bob".into(), championships: 3, wins: 90 },
        Player { name: "Charlie".into(), championships: 3, wins: 100 },
    ];

    // 冗余的嵌套match写法
    players.sort_by(|a, b| {
        let cmp_champs = b.championships.cmp(&a.championships);
        match cmp_champs {
            Ordering::Equal => {
                let cmp_wins = b.wins.cmp(&a.wins);
                match cmp_wins {
                    Ordering::Equal => a.name.cmp(&b.name),
                    other => other,
                }
            }
            other => other,
        }
    });
}

简化后的实现 #

用then方法链式调用，完全替代嵌套match：

use std::cmp::Ordering;

#[derive(Debug, PartialEq)]
struct Player {
    name: String,
    championships: u32,
    wins: u32,
}

fn main() {
    let mut players = vec![
        Player { name: "Alice".into(), championships: 2, wins: 100 },
        Player { name: "Bob".into(), championships: 3, wins: 90 },
        Player { name: "Charlie".into(), championships: 3, wins: 100 },
    ];

    // 简洁的链式排序
    players.sort_by(|a, b| {
        b.championships.cmp(&a.championships)
            .then(b.wins.cmp(&a.wins))
            .then(a.name.cmp(&b.name))
    });
}

为什么这能行？ #

• Ordering::then方法会先判断当前排序结果：如果不是Equal，直接返回当前结果；如果是Equal，才会执行后面的比较逻辑，完美替代嵌套match的分支判断。
• 降序需求直接通过b.field.cmp(&a.field)实现（默认a.cmp(&b)是升序，反过来就是降序），升序则用a.field.cmp(&b.field)。
• 如果你的排序逻辑需要延迟计算（比如要先对字段做转换再比较），可以用then_with替代then，它接受一个返回Ordering的闭包：

// 比如name需要转成小写再比较升序
players.sort_by(|a, b| {
    b.championships.cmp(&a.championships)
        .then_with(|| b.wins.cmp(&a.wins))
        .then_with(|| a.name.to_lowercase().cmp(&b.name.to_lowercase()))
});

这种写法不仅消除了冗余的match代码，还让排序逻辑的层级关系一目了然，可读性拉满。

内容的提问来源于stack exchange，提问作者Patrick Bucher