Java和Go性能的简单比较（mysql query）

通过简单地查询公司表(comp)的100条数据来直观地比较语言的性能。

环境：笔记本64位win10, jdk 1.8, go 1.11, 本机10.1.15-MariaDB。

 

1. java代码

package main;

import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.ResultSet;
import java.sql.Statement;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;

/**
 * 连续查询100条comp，测试时间：
 */
public class TestSqlStress {

    public static void main(String[] args) throws Exception {
        long start0 = System.currentTimeMillis();
        String url = "jdbc:mysql://127.0.0.1:3306/testdb?connectionCollation=utf8mb4_general_ci&serverTimezone=Asia/Shanghai&user=root&password=1";
        Connection cn = DriverManager.getConnection(url);

        long start = System.currentTimeMillis();

        String query = "SELECT id,sync_key,name,logo_url,capital,found_date,staff,tel,addr,intro,\n" +
                "            IFNULL(tag_codes,\"\") 'tag_codes', IFNULL(tag_names,\"\") 'tag_names',\n" +
                "            Unix_timestamp(create_time) 'create_time', is_disable, \n" +
                "            IFNULL(Unix_timestamp(disable_time),0) 'disable_time', IFNULL(disable_remark, \"\") 'disable_remark'\n" +
                "        FROM comp\n" +
                "        LIMIT 100";
        Statement stmt = cn.createStatement();
        ResultSet rs = stmt.executeQuery(query);

        long id;
        String sync_key, name, logo_url,capital,found_date,staff,tel,addr,intro;
        String tag_codes, tag_names;
        long create_time, is_disable, disable_time;
        String disable_remark;

        List arr = new ArrayList(200);
        while (rs.next()) {
            id = rs.getLong(1);
            sync_key = rs.getString(2);
            name = rs.getString(3);
            logo_url = rs.getString(4);
            capital = rs.getString(5);
            found_date = rs.getString(6);
            staff = rs.getString(7);
            tel = rs.getString(8);
            addr = rs.getString(9);
            intro = rs.getString(10);
            tag_codes = rs.getString(11);
            tag_names = rs.getString(12);
            create_time = rs.getLong(13);
            is_disable = rs.getLong(14);
            disable_time = rs.getLong(15);
            disable_remark = rs.getString(16);

            Map map = new HashMap();
            map.put("id", id);
            map.put("sync_key", sync_key);
            map.put("name", name);
            map.put("logo_url", logo_url);
            map.put("capital", capital);
            map.put("found_date", found_date);
            map.put("staff", staff);
            map.put("tel", tel);
            map.put("addr", addr);
            map.put("intro", intro);
            map.put("tag_codes", tag_codes);
            map.put("tag_names", tag_names);
            map.put("create_time", create_time);
            map.put("is_disable", is_disable);
            map.put("disable_time", disable_time);
            map.put("disable_remark", disable_remark);
            arr.add(map);
        }

        System.out.println("query took " + (System.currentTimeMillis() - start) + " ms");

        String str = arr.toString();
        System.out.println("str len=" + str.length());

        rs.close();
        stmt.close();
        cn.close();
        System.out.println("main took " + (System.currentTimeMillis() - start0) + " ms");
    }

}

 

2. Go代码

// 连续查询100条comp，测试时间
package main

import (
    "database/sql"
    "encoding/json"
    "fmt"
    "time"

    _ "github.com/ziutek/mymysql/godrv"
)

func main() {
    start0 := time.Now()
    defer func() {
        fmt.Println("main took", time.Since(start0).Nanoseconds() / 1000000, "ms")
    }()

    db, err := sql.Open("mymysql", "tcp:127.0.0.1:3306*testdb/root/1")
    if err != nil { panic(err) }
    defer db.Close()

    start := time.Now()

    query := `
        SELECT id,sync_key,name,logo_url,capital,found_date,staff,tel,addr,intro,
            IFNULL(tag_codes,"") 'tag_codes', IFNULL(tag_names,"") 'tag_names',
            Unix_timestamp(create_time) 'create_time', is_disable, 
            IFNULL(Unix_timestamp(disable_time),0) 'disable_time', IFNULL(disable_remark, "") 'disable_remark'
        FROM comp
        LIMIT 100
    `
    rows, err := db.Query(query)
    if err != nil { panic(err) }
    defer rows.Close()

    var id int64
    var sync_key, name, logo_url,capital,found_date,staff,tel,addr,intro string
    var tag_codes, tag_names string
    var create_time, is_disable, disable_time int64
    var disable_remark string

    arr := make([]map[string]interface{}, 0)
    for rows.Next() {
        err = rows.Scan(&id, &sync_key, &name, &logo_url, &capital, &found_date, &staff, &tel, &addr, &intro,
            &tag_codes, &tag_names, 
            &create_time, &is_disable, &disable_time,
            &disable_remark)
        if err != nil { panic(err) }

        ma := map[string]interface{} {
            "id": id,
            "sync_key": sync_key,
            "name": name,
            "logo_url": logo_url,
            "capital": capital,
            "found_date": found_date,
            "staff": staff,
            "tel": tel,
            "addr": addr,
            "intro": intro,
            "tag_codes": tag_codes,
            "tag_names": tag_names,
            "create_time": create_time,
            "is_disable": is_disable,
            "disable_time": disable_time,
            "disable_remark": disable_remark,
        }
        arr = append(arr, ma)
    }
    err = rows.Err()  // rows.Next 的错误
    if err != nil { panic(err) }

    fmt.Println("query took", time.Since(start).Nanoseconds() / 1000000, "ms")

    buf, err := json.Marshal(arr)
    if err != nil { panic("json marshal error: " + err.Error()) }
    fmt.Println("json len=", len(string(buf)))
}

 

3. 测试结果

sql_stress Go (ziutek/mymysql v1.5.4)

    query took 3 ms

    json len= 90520

    main took 6 ms

sql_stress Java

    -- mysql-connector-java-5.1.25

        query took 18 ms

        str len=50869

        main took 389 ms

    -- mysql-connector-java-8.0.15

        query took 38 ms

        str len=50869

        main took 473 ms

 

4. 结论分析

首先令人吃惊的是Go查询仅用了3毫秒，而Java快的也用了18毫秒。

其次mysql 官方的jdbc driver 从5.1升级到8.0后性能反而下降一倍左右：从18毫秒慢到了38毫秒。

 

初步看来，Go除了代码简洁外，运行性能也是令人放心的。