Hadoop 中利用 mapreduce 读写 mysql 数据
有时候我们在项目中会遇到输入结果集很大，但是输出结果很小，比如一些 pv、uv 数据，然后为了实时查询的需求，或者一些 OLAP 的需求，我们需要 mapreduce 与 mysql 进行数据的交互，而这些特性正是 hbase 或者 hive 目前亟待改进的地方。
好了言归正传，简单的说说背景、原理以及需要注意的地方：
1、为了方便 MapReduce 直接访问关系型数据库（Mysql,Oracle），Hadoop提供了DBInputFormat和DBOutputFormat两个类。通过DBInputFormat类把数据库表数据读入到HDFS，根据DBOutputFormat类把MapReduce产生的结果集导入到数据库表中。
2、由于0.20版本对DBInputFormat和DBOutputFormat支持不是很好，该例用了0.19版本来说明这两个类的用法。
至少在我的 0.20.203 中的org.apache.hadoop.mapreduce.lib 下是没见到 db 包，所以本文也是以老版的 API 来为例说明的。
3、运行MapReduce时候报错：java.io.IOException: com.mysql.jdbc.Driver，一般是由于程序找不到mysql驱动包。解决方法是让每个tasktracker运行MapReduce程序时都可以找到该驱动包。
添加包有两种方式：
（1）在每个节点下的${HADOOP_HOME}/lib下添加该包。重启集群，一般是比较原始的方法。
（2）a)把包传到集群上： hadoop fs -put mysql-connector-java-5.1.0- bin.jar /hdfsPath/
b)在mr程序提交job前，添加语句：DistributedCache.addFileToClassPath(new Path(“/hdfsPath/mysql- connector-java- 5.1.0-bin.jar”), conf);
（3）虽然API用的是0.19的，但是使用0.20的API一样可用，只是会提示方法已过时而已。
4、测试数据：
- CREATE TABLE `t` (
`id` int DEFAULT NULL,
`name` varchar(10) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
CREATE TABLE `t2` (
`id` int DEFAULT NULL,
`name` varchar(10) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
insert into t values (1,"june"),(2,"decli"),(3,"hello"),
(4,"june"),(5,"decli"),(6,"hello"),(7,"june"),
(8,"decli"),(9,"hello"),(10,"june"),
(11,"june"),(12,"decli"),(13,"hello");
5、代码：
- package mysql2mr;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.File;
import java.io.IOException;
import java.sql.PreparedStatement;
import java.sql.ResultSet;
import java.sql.SQLException;
import mapr.EJob;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.filecache.DistributedCache;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.db.DBConfiguration;
import org.apache.hadoop.mapreduce.lib.db.DBInputFormat;
import org.apache.hadoop.mapreduce.lib.db.DBOutputFormat;
import org.apache.hadoop.mapreduce.lib.db.DBWritable;
/**
* Function: 测试 mr 与 mysql 的数据交互，此测试用例将一个表中的数据复制到另一张表中 实际当中，可能只需要从 mysql 读，或者写到
* mysql 中。
*
* @author administrator
*
*/
public class Mysql2Mr {
public static class StudentinfoRecord implements Writable, DBWritable {
int id;
String name;
public StudentinfoRecord() {
}
public String toString() {
return new String(this.id + " " + this.name);
}
@Override
public void readFields(ResultSet result) throws SQLException {
this.id = result.getInt(1);
this.name = result.getString(2);
}
@Override
public void write(PreparedStatement stmt) throws SQLException {
stmt.setInt(1, this.id);
stmt.setString(2, this.name);
}
@Override
public void readFields(DataInput in) throws IOException {
this.id = in.readInt();
this.name = Text.readString(in);
}
@Override
public void write(DataOutput out) throws IOException {
out.writeInt(this.id);
Text.writeString(out, this.name);
}
}
// 记住此处是静态内部类，要不然你自己实现无参构造器，或者等着抛异常：
// Caused by: java.lang.NoSuchMethodException: DBInputMapper.<init>()
// http://stackoverflow.com/questions/7154125/custom-mapreduce-input-format-cant-find-constructor
// 网上脑残式的转帖，没见到一个写对的。。。
public static class DBInputMapper extends
Mapper<LongWritable, StudentinfoRecord, LongWritable, Text> {
@Override
public void map(LongWritable key, StudentinfoRecord value,
Context context) throws IOException, InterruptedException {
context.write(new LongWritable(value.id), new Text(value.toString()));
}
}
public static class MyReducer extends Reducer<LongWritable, Text, StudentinfoRecord, Text> {
@Override
public void reduce(LongWritable key, Iterable<Text> values, Context context) throws IOException, InterruptedException {
String[] splits = values.iterator().next().toString().split(" ");
StudentinfoRecord r = new StudentinfoRecord();
r.id = Integer.parseInt(splits[0]);
r.name = splits[1];
context.write(r, new Text(r.name));
}
}
@SuppressWarnings("deprecation")
public static void main(String[] args) throws IOException, InterruptedException, ClassNotFoundException {
File jarfile = EJob.createTempJar("bin");
EJob.addClasspath("usr/hadoop/conf");
ClassLoader classLoader = EJob.getClassLoader();
Thread.currentThread().setContextClassLoader(classLoader);
Configuration conf = new Configuration();
// 这句话很关键
conf.set("mapred.job.tracker", "172.30.1.245:9001");
DistributedCache.addFileToClassPath(new Path(
"hdfs://172.30.1.245:9000/user/hadoop/jar/mysql-connector-java-5.1.6-bin.jar"), conf);
DBConfiguration.configureDB(conf, "com.mysql.jdbc.Driver", "jdbc:mysql://172.30.1.245:3306/sqooptest", "sqoop", "sqoop");
Job job = new Job(conf, "Mysql2Mr");
// job.setJarByClass(Mysql2Mr.class);
((JobConf)job.getConfiguration()).setJar(jarfile.toString());
job.setMapOutputKeyClass(LongWritable.class);
job.setMapOutputValueClass(Text.class);
job.setMapperClass(DBInputMapper.class);
job.setReducerClass(MyReducer.class);
job.setOutputKeyClass(LongWritable.class);
job.setOutputValueClass(Text.class);
job.setOutputFormatClass(DBOutputFormat.class);
job.setInputFormatClass(DBInputFormat.class);
String[] fields = {"id","name"};
// 从 t 表读数据
DBInputFormat.setInput(job, StudentinfoRecord.class, "t", null, "id", fields);
// mapreduce 将数据输出到 t2 表
DBOutputFormat.setOutput(job, "t2", "id", "name");
System.exit(job.waitForCompletion(true)? 0:1);
}
}
6、结果：
执行两次后，你可以看到mysql结果：
- mysql> select * from t2;
+------+-------+
| id | name |
+------+-------+
| 1 | june |
| 2 | decli |
| 3 | hello |
| 4 | june |
| 5 | decli |
| 6 | hello |
| 7 | june |
| 8 | decli |
| 9 | hello |
| 10 | june |
| 11 | june |
| 12 | decli |
| 13 | hello |
| 1 | june |
| 2 | decli |
| 3 | hello |
| 4 | june |
| 5 | decli |
| 6 | hello |
| 7 | june |
| 8 | decli |
| 9 | hello |
| 10 | june |
| 11 | june |
| 12 | decli |
| 13 | hello |
+------+-------+
26 rows in set (0.00 sec)
mysql>
