习惯与函数式编程语言的开发者,会认为for循环和if判断语句是冗长的代码,通过使用map和filter处理集合元素让代码更可读。本文介绍Go闭包实现集合转换和过滤功能。

定义集合功能函数

WorkWith
// WorkWith is the struct we'll
// be implementing collections for
type WorkWith struct {
Data string
Version int
}

针对该结构体定义filter和map函数:

// 基于判断函数过滤集合,返回符合条件的集合元素
func Filter(ws []WorkWith, f func(w WorkWith) bool) []WorkWith {
// depending on results, smaller size for result
// is len == 0
result := make([]WorkWith, 0)
for _, w := range ws {
if f(w) {
result = append(result, w)
}
}
return result
}
// 基于转换函数转换集合元素,返回集合的元素为转换后的元素
func Map(ws []WorkWith, f func(w WorkWith) WorkWith) []WorkWith {
// the result should always be the same
// length
result := make([]WorkWith, len(ws))
for pos, w := range ws {
newW := f(w)
result[pos] = newW
}
return result
}

实现具体功能函数

import "strings"
// LowerCaseData does a ToLower to the
// Data string of a WorkWith
func LowerCaseData(w WorkWith) WorkWith {
w.Data = strings.ToLower(w.Data)
return w
}
// IncrementVersion increments a WorkWiths
// Version
func IncrementVersion(w WorkWith) WorkWith {
w.Version++
return w
}
// OldVersion returns a closures
// that validates the version is greater than
// the specified amount
func OldVersion(v int) func(w WorkWith) bool {
return func(w WorkWith) bool {
return w.Version >= v
}
}

上面定义了三个函数,LowerCaseData修改WorkWith中Data值为小写形式,IncrementVersion让WorkWith中版本增加1,OldVersion基于参数过滤版本。

测试集合功能

定义测试用例文件:


import (
"fmt"
"testing"
)
func TestMap(t *testing.T) {
ws := []WorkWith{
{"Example", 1},
{"Example 2", 2},
}
fmt.Printf("Initial list: %#v\n", ws)
// first lower case the list
ws = Map(ws, LowerCaseData)
fmt.Printf("After LowerCaseData Map: %#v\n", ws)
// next increment all versions
ws = Map(ws, IncrementVersion)
fmt.Printf("After IncrementVersion Map: %#v\n", ws)
// lastly remove all versions older than 3
ws = Filter(ws, OldVersion(3))
fmt.Printf("After OldVersion Filter: %#v\n", ws)
}

运行 go test . -v

输出结果如下:

Initial list: []collections.WorkWith{collections.WorkWith{Data:"Example", Version:1}, collections.WorkWith{Data:"Example 2", Version:2}}
After LowerCaseData Map: []collections.WorkWith{collections.WorkWith{Data:"example", Version:1}, collections.WorkWith{Data:"example 2", Version:2}}
After IncrementVersion Map: []collections.WorkWith{collections.WorkWith{Data:"example", Version:2}, collections.WorkWith{Data:"example 2", Version:3}}
After OldVersion Filter: []collections.WorkWith{collections.WorkWith{Data:"example 2", Version:3}}

上面示例中,我们注意到函数都没有返回任何error对象,这遵循函数式编程思想,尽可能让函数纯粹:不修改原集合元素,即对原集合无副作用,而是生成新的集合。如果需要对集合应用多个功能,那么这种模式能够省去很多麻烦,并且测试也很简单。我们还可以将映射和过滤器链接在一起,让代码更简洁可读。

 ws := []WorkWith{
{"Example", 1},
{"Example 2", 2},
}
fmt.Printf("Initial list: %#v\n", ws)
result := Filter(Map(Map(ws, LowerCaseData), IncrementVersion), OldVersion(3))
fmt.Printf("After OldVersion Filter: %#v\n", result)

如果功能函数定义为集合类型的方法,并返回集合类型,则上述代码会更优雅。

泛型实现

上面代码仅能在特定类型上使用,我们自然想实现泛型函数,下面通过一个简单示例进行说明:

func map2[T, U any](data []T, f func(T) U) []U {
res := make([]U, 0, len(data))
for _, e := range data {
res = append(res, f(e))
}
return res
}

该函数接收类型T,转换后返回类型U,当然两者类型也可以一样。下面测试函数功能:

 // 字符串转大写
words := []string{"war", "cup", "water", "tree", "storm"}
result := map2(words, func(s string) string {
return strings.ToUpper(s)
})
fmt.Println(result)
// 生成原集合元素的平方集合
fmt.Println("-------------------")
numbers := []int{1, 2, 3, 4, 5, 6, 7, 8, 9, 10}
squares := map2(numbers, func(n int) int {
return n * n
})
fmt.Println(squares)
// 数值转为字符串
fmt.Println("-------------------")
as_strings := map2(numbers, func(n int) string {
return strconv.Itoa(n)
})
fmt.Printf("%q", as_strings)