spark之JDBC开发（连接数据库测试）

spark之JDBC开发（连接数据库测试）

---

以下操作属于本地模式操作：

1、在Eclipse4.5中建立工程RDDToJDBC，并创建一个文件夹lib用于放置第三方驱动包

[hadoop@CloudDeskTop software]$ cd /project/RDDToJDBC/
[hadoop@CloudDeskTop RDDToJDBC]$ mkdir -p lib
[hadoop@CloudDeskTop RDDToJDBC]$ ls
bin lib src

2、添加必要的环境

2.1、将MySql的jar包拷贝到工程目录RDDToJDBC下的lib目录下
[hadoop@CloudDeskTop software]$ cp -a /software/hive-1.2.2/lib/mysql-connector-java-3.0.17-ga-bin.jar /project/RDDToJDBC/lib/
2.1、将Spark的开发库Spark2.1.1-All追加到RDDToJDBC工程的classpath路径中去(可以通过添加用户库的方式来解决)；Spark2.1.1-All中包含哪些包，请点击此处

3、准备spark的源数据：

[hadoop@CloudDeskTop spark]$ cd /home/hadoop/test/jdbc/
[hadoop@CloudDeskTop jdbc]$ ls
myuser  testJDBC.txt
[hadoop@CloudDeskTop jdbc]$ cat myuser
lisi 123456 165 1998-9-9
lisan 123ss 187 2009-10-19
wangwu 123qqwe 177 1990-8-3

4、开发源码：

package com.mmzs.bigdata.spark.core.local;

import java.io.File;
import java.sql.Connection;
import java.sql.Date;
import java.sql.DriverManager;
import java.sql.PreparedStatement;
import java.sql.SQLException;

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.VoidFunction;

import scala.Tuple4;

public class TestMain {
    /**
     * 全局计数器
     */
    private static int count;

    /**
     * 数据库连接
     */
    private static Connection conn;

    /**
     * 预编译语句
     */
    private static PreparedStatement pstat;

    private static final File OUT_PATH=new File("/home/hadoop/test/jdbc/output");

    static{
        delDir(OUT_PATH);
        try {
            String sql="insert into myuser(userName,passWord,height,birthday) values(?,?,?,?)";
            String url="jdbc:mysql://127.0.0.1:3306/test?useUnicode=true&characterEncoding=utf8";
            Class.forName("com.mysql.jdbc.Driver");
            conn=DriverManager.getConnection(url, "root", "123456");
            pstat=conn.prepareStatement(sql);
        } catch (ClassNotFoundException e) {
            e.printStackTrace();
        } catch (SQLException e) {
            e.printStackTrace();
        }
    }
    /**
     * 删除任何目录或文件
     * @param f
     */
    private static void delDir(File f){
        if(!f.exists())return;
        if(f.isFile()||(f.isDirectory()&&f.listFiles().length==0)){
            f.delete();
            return;
        }
        File[] files=f.listFiles();
        for(File fp:files)delDir(fp);
        f.delete();
    }

    private static void batchSave(Tuple4<String,String,Double,Date> line,boolean isOver){
        try{
            pstat.setString(1, line._1());
            pstat.setString(2, line._2());
            pstat.setDouble(3, line._3());
            pstat.setDate(4, line._4());

            if(isOver){//如果结束了循环则直接写磁盘
                pstat.addBatch();
                pstat.executeBatch();
                pstat.clearBatch();
                pstat.clearParameters();
            }else{ //如果没有结束则将sql语句添加到批处理中去
                pstat.addBatch();
                count++;
                if(count%100==0){ //如果满一个批次就提交一次批处理操作
                    pstat.executeBatch();
                    pstat.clearBatch();
                    pstat.clearParameters();
                }
            }
        }catch(SQLException e){
            e.printStackTrace();
        }
    }

    /**
     * 将RDD集合中的数据存储到关系数据库MYSql中去
     * @param statResRDD
     */
    private static void saveToDB(JavaRDD<String> statResRDD){
        final long rddNum=statResRDD.count();
        statResRDD.foreach(new VoidFunction<String>(){
            private long count=0;
            @Override
            public void call(String line) throws Exception {
                String[] fields=line.split(" ");
                String userName=fields[0];
                String passWord=fields[1];
                Double height=Double.parseDouble(fields[2]);
                Date birthday=Date.valueOf(fields[3]);
                Tuple4<String,String,Double,Date> fieldTuple=new Tuple4<String,String,Double,Date>(userName,passWord,height,birthday);
                if(++count<rddNum){
                    batchSave(fieldTuple,false);
                }else{
                    batchSave(fieldTuple,true);
                }
            }
        });

        try{
            if(null!=pstat)pstat.close();
            if(null!=conn)conn.close();
        }catch(SQLException e){
            e.printStackTrace();
        }
    }

    public static void main(String[] args) {
        SparkConf conf=new SparkConf();
        conf.setAppName("Java Spark local");
        conf.setMaster("local");

        //根据Spark配置生成Spark上下文
        JavaSparkContext jsc=new JavaSparkContext(conf);

        //读取本地的文本文件成内存中的RDD集合对象
        JavaRDD<String> lineRdd=jsc.textFile("/home/hadoop/test/jdbc/myuser");

        //...........其它转换或统计操作................

        //存储统计之后的结果到磁盘文件中去
        //lineRdd.saveAsTextFile("/home/hadoop/test/jdbc/output");
        saveToDB(lineRdd);

        //关闭Spark上下文
        jsc.close();
    }
}

5、初始化MySql数据库服务(节点在192.168.154.134上)

A、启动MySql数据库服务

[root@DB03 ~]# cd /software/mysql-5.5.32/multi-data/3306/
[root@DB03 3306]# ls
data my.cnf my.cnf.bak mysqld
[root@DB03 3306]# ./mysqld start
Starting MySQL...

B、建立test库

[root@CloudDeskTop 3306]# cd /software/mysql-5.5.32/bin/
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "show databases;"
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
+--------------------+
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "create database test character set utf8;"
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "show databases;"
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
| test               |
+--------------------+

C、建立myuser表：

[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "create table if not exists test.myuser(uid int(11) auto_increment primary key,username varchar(30),password varchar(30),height double(10,1),birthday date)engine=myisam charset=utf8;"
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "use test;show tables;"
+-------------------+
| Tables_in_test    |
+-------------------+
| myuser            |
+-------------------+
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "use test;desc test.myuser;"
+----------+--------------+------+-----+---------+----------------+
| Field    | Type         | Null | Key | Default | Extra          |
+----------+--------------+------+-----+---------+----------------+
| uid      | int(11)      | NO   | PRI | NULL    | auto_increment |
| username | varchar(30)  | YES  |     | NULL    |                |
| password | varchar(30)  | YES  |     | NULL    |                |
| height   | double(10,1) | YES  |     | NULL    |                |
| birthday | date         | YES  |     | NULL    |                |
+----------+--------------+------+-----+---------+----------------+

#目前数据库表中还没有数据
[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "select * from test.myuser;"

6、运行并查看数据库中结果

6.1、在Eclipse4.5中直接运行Spark代码，观察Eclipse控制台输出

6.2、检查在关系数据库MySql中是否已经存在数据

[root@CloudDeskTop bin]# ./mysql -h192.168.154.134 -P3306 -uroot -p123456 -e "select * from test.myuser;"
+-----+----------+----------+--------+------------+
| uid | username | password | height | birthday   |
+-----+----------+----------+--------+------------+
|   1 | lisi     | 123456   |  165.0 | 1998-09-09 |
|   2 | lisan    | 123ss    |  187.0 | 2009-10-19 |
|   3 | wangwu   | 123qqwe  |  177.0 | 1990-08-03 |
+-----+----------+----------+--------+------------+