近期有一个需求：获取多个文件 md5 校验和判断是否存在重复文件，因为文件数量较多，有的文件还比较大，需要处理的文件还没有到位，我就考虑了一下效率的问题。目前我已知的 Golang 中获取 md5 校验和的方法有两个，这里直接给出实现源码。

package main

import (
    "crypto/md5"
    "flag"
    "fmt"
    "io"
    "io/IoUtil"
    "os"
)

var which = flag.Bool("which",true,"")
var path = flag.String("path","","")
var cnt = flag.Int("cnt", 100,"")

func aaa() {
    f,err := os.Open(*path)
    if err != nil {
        fmt.Println("Open",err)
        return
    }

    defer f.Close()

    body,err := IoUtil.ReadAll(f)
    if err != nil {
        fmt.Println("ReadAll",err)
        return
    }

    md5.Sum(body)
    //fmt.Printf("%x\n",md5.Sum(body))
}

func bbb() {
    f,err)
        return
    }

    defer f.Close()

    md5hash := md5.New()
    if _,err := io.Copy(md5hash,f); err != nil {
        fmt.Println("Copy",err)
        return
    }

    md5hash.Sum(nil)
    //fmt.Printf("%x\n",md5hash.Sum(nil))
}

func main() {
    flag.Parse()

    for i := 0; i < *cnt; i++ {
        if *which {
            aaa()
        } else {
            bbb()
        }
    }
}

还有可供参考的获取 md5 校验和的 Shell 命令

md5 -- calculate a message-digest fingerprint (checksum) for a file
md5 [-pqrtx] [-s string] [file ...]

测试文件是公司项目的日志文件

banjakukutekiiMac:shell panshiqu$ ls -an | grep by
-rw-r--r--   1 501  20   7285957 11 17 16:14 by.out

banjakukutekiiMac:shell panshiqu$ cp by.out by2.out
banjakukutekiiMac:shell panshiqu$ cat by.out >> by2.out

banjakukutekiiMac:shell panshiqu$ ls -an | grep by
-rw-r--r--   1 501  20   7285957 11 17 16:14 by.out
-rw-r--r--   1 501  20  14571914 11 17 17:03 by2.out

下面效率展示

banjakukutekiiMac:shell panshiqu$ time ./gomd5 -cnt=1 -which=true -path="by.out"

real    0m0.027s
user    0m0.017s
sys 0m0.012s
banjakukutekiiMac:shell panshiqu$ time ./gomd5 -cnt=1 -which=true -path="by2.out"

real    0m0.048s
user    0m0.033s
sys 0m0.018s
banjakukutekiiMac:shell panshiqu$ time ./gomd5 -cnt=1 -which=false -path="by.out"

real    0m0.018s
user    0m0.012s
sys 0m0.004s
banjakukutekiiMac:shell panshiqu$ time ./gomd5 -cnt=1 -which=false -path="by2.out"

real    0m0.031s
user    0m0.024s
sys 0m0.005s
banjakukutekiiMac:shell panshiqu$ time md5 by.out
MD5 (by.out) = 9d79e19a00cef1ae1bb6518ca4adf9de

real    0m0.023s
user    0m0.019s
sys 0m0.006s
banjakukutekiiMac:shell panshiqu$ time md5 by2.out
MD5 (by2.out) = 0a029a460a20e8dcb00d032d6fab74c6

real    0m0.042s
user    0m0.037s
sys 0m0.009s

总结：
* 不管什么方法都会随着文件变大时间会变长，上面的例子大约都是2倍
* io.Copy 方法效率最高，建议大家这样使用