环境
|
1
2
|
go version go1.16.4 windows/amd64Intel(R) Core(TM) i7-7820HK CPU @ 2.90GHz 4核心8线程 |
项目需求
处理数个约5MB的小文件
从源目录读取文件并拷贝到目标目录
计算源文件MD5和目标文件MD5进行对比，如不相同则报错并终止程序执行
mapReduce使用说明
|
1
|
go get -u github.com/tal-tech/go-zero |
需求实现
判断上下文是否中止 → 读取数据 → 写入数据 → 校验MD5
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
|
func fnBuilder(name string) func() error { return func() error { // 判断上下文是否终止 select { case <-ctx.Done(): return ctx.Err() default: } // 读取源数据 data, _err := os.ReadFile(filepath.Join(sourcePath, fileName)) // 计算源数据MD5 sourceMD5 := hash.Md5(data) // 获取名称 fields := strings.Split(d.Name(), "-") // 目标文件路径 distFilePath := filepath.Join(distPath, fileName) // 拷贝数据 os.WriteFile(distFilePath, data, 0600) // 校验数据 distData, _err := os.ReadFile(distFilePath) distMD5 := hash.Md5(distData) if !bytes.EqualFold(sourceMD5, distMD5) { return errors.New("md5校验失败") } return nil }} |
业务逻辑
创建任务队列
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
|
type SourceMap = map[string]fs.DirEntryfunc CopyFileToDist(ctx context.Context, source SourceMap) (err error) { // 创建工作队列 work := make([]func() error, 0, len(source)) for _name := range source { // 创建任务 work = append(work, fnBuilder(_name)) } switch concurrency { default: // mapReduce case 1: // sync.waitGroup case 2: // 串行 }} |
执行方式1：MapReduce
|
1
2
3
4
5
|
func() { if err = mr.Finish(work...); err != nil { return err }} |
执行方式2：sync.WaitGroup
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
|
func() { var wg sync.WaitGroup wg.Add(len(work)) for k := range work { go func(index int) { defer wg.Done() if err = work[index](); err != nil { log.Errorln(err) return } }(k) } wg.Wait()} |
执行方式3：串行
|
1
2
3
4
5
6
7
|
func() { for _, fn := range work { if err = fn(); err != nil { return err } }} |
运行结果
MapReduce
耗时 109220900 ns
|
1
2
|
{"file":"D:/go/src/filenamesSorter/main.go:44","func":"main.init.0","level":"info","msg":"并发处理（0-mapReduce 1-Sync.WaitGroup 2-不并发） 0","time":"2021-06-02T13:32:05+08:00"}{"file":"D:/go/src/filenamesSorter/main.go:69","func":"main.main","level":"info","msg":"文件分类完毕","time":"2021-06-02T13:32:05+08:00","文件数":17,"耗时(ns)":109220900} |
sync.WaitGroup
耗时 109798000 ns
|
1
2
|
{"file":"D:/go/src/filenamesSorter/main.go:44","func":"main.init.0","level":"info","msg":"并发处理（0-mapReduce 1-Sync.WaitGroup 2-不并发） 1","time":"2021-06-02T13:31:28+08:00"}{"file":"D:/go/src/filenamesSorter/main.go:69","func":"main.main","level":"info","msg":"文件分类完毕","time":"2021-06-02T13:31:28+08:00","文件数":17,"耗时(ns)":109798000} |
串行
耗时 359307700 ns
|
1
2
|
{"file":"D:/go/src/filenamesSorter/main.go:44","func":"main.init.0","level":"info","msg":"并发处理（0-mapReduce 1-Sync.WaitGroup 2-不并发） 2","time":"2021-06-02T13:33:02+08:00"}{"file":"D:/go/src/filenamesSorter/main.go:69","func":"main.main","level":"info","msg":"文件分类完毕","time":"2021-06-02T13:33:02+08:00","文件数":17,"耗时(ns)":359307700} |
结论
- 在不严格的情况下，执行效率方面可以认为 mapReduce ≈ sync.WaitGroup
- 易用性（包括并发和错误处理），mapReduce 完胜 sync.WaitGroup
- mapReduce好用
引申阅读
以上就是Golang编程并发工具库MapReduce使用实践的详细内容，更多关于Golang并发工具库MapReduce的资料请关注服务器之家其它相关文章！
原文链接：https://blog.csdn.net/yes169yes123/article/details/117467920
