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：

import java.io.DataInput;import java.io.DataOutput;import java.io.IOException;import java.sql.PreparedStatement;import java.sql.ResultSet;import java.sql.SQLException;import java.util.Iterator;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.JobClient;import org.apache.hadoop.mapred.JobConf;import org.apache.hadoop.mapred.MapReduceBase;import org.apache.hadoop.mapred.Mapper;import org.apache.hadoop.mapred.OutputCollector;import org.apache.hadoop.mapred.Reducer;import org.apache.hadoop.mapred.Reporter;import org.apache.hadoop.mapred.lib.IdentityReducer;import org.apache.hadoop.mapred.lib.db.DBConfiguration;import org.apache.hadoop.mapred.lib.db.DBInputFormat;import org.apache.hadoop.mapred.lib.db.DBOutputFormat;import org.apache.hadoop.mapred.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。 * date: 2013-7-29 上午2:34:04 <br/> * @author june */public class Mysql2Mr {// DROP TABLE IF EXISTS `hadoop`.`studentinfo`;// CREATE TABLE studentinfo (// id INTEGER NOT NULL PRIMARY KEY,// name VARCHAR(32) NOT NULL);public static class StudentinfoRecord implements Writable, DBWritable {int id;String name;public StudentinfoRecord() {}public void readFields(DataInput in) throws IOException {this.id = in.readInt();this.name = Text.readString(in);}public String toString() {return new String(this.id + " " + this.name);}@Overridepublic void write(PreparedStatement stmt) throws SQLException {stmt.setInt(1, this.id);stmt.setString(2, this.name);}@Overridepublic void readFields(ResultSet result) throws SQLException {this.id = result.getInt(1);this.name = result.getString(2);}@Overridepublic 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 MapReduceBase implementsMapper<LongWritable, StudentinfoRecord, LongWritable, Text> {public void map(LongWritable key, StudentinfoRecord value,OutputCollector<LongWritable, Text> collector, Reporter reporter) throws IOException {collector.collect(new LongWritable(value.id), new Text(value.toString()));}}public static class MyReducer extends MapReduceBase implementsReducer<LongWritable, Text, StudentinfoRecord, Text> {@Overridepublic void reduce(LongWritable key, Iterator<Text> values,OutputCollector<StudentinfoRecord, Text> output, Reporter reporter) throws IOException {String[] splits = values.next().toString().split(" ");StudentinfoRecord r = new StudentinfoRecord();r.id = Integer.parseInt(splits[0]);r.name = splits[1];output.collect(r, new Text(r.name));}}public static void main(String[] args) throws IOException {JobConf conf = new JobConf(Mysql2Mr.class);DistributedCache.addFileToClassPath(new Path("/tmp/mysql-connector-java-5.0.8-bin.jar"), conf);conf.setMapOutputKeyClass(LongWritable.class);conf.setMapOutputValueClass(Text.class);conf.setOutputKeyClass(LongWritable.class);conf.setOutputValueClass(Text.class);conf.setOutputFormat(DBOutputFormat.class);conf.setInputFormat(DBInputFormat.class);// // mysql to hdfs// conf.setReducerClass(IdentityReducer.class);// Path outPath = new Path("/tmp/1");// FileSystem.get(conf).delete(outPath, true);// FileOutputFormat.setOutputPath(conf, outPath);DBConfiguration.configureDB(conf, "com.mysql.jdbc.Driver", "jdbc:mysql://192.168.1.101:3306/test","root", "root");String[] fields = { "id", "name" };// 从 t 表读数据DBInputFormat.setInput(conf, StudentinfoRecord.class, "t", null, "id", fields);// mapreduce 将数据输出到 t2 表DBOutputFormat.setOutput(conf, "t2", "id", "name");// conf.setMapperClass(org.apache.hadoop.mapred.lib.IdentityMapper.class);conf.setMapperClass(DBInputMapper.class);conf.setReducerClass(MyReducer.class);JobClient.runJob(conf);}}

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>

7、日志：

13/07/29 02:33:03 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same.13/07/29 02:33:03 INFO filecache.TrackerDistributedCacheManager: Creating mysql-connector-java-5.0.8-bin.jar in /tmp/hadoop-june/mapred/local/archive/-8943686319031389138_-1232673160_640840668/192.168.1.101/tmp-work--8372797484204470322 with rwxr-xr-x13/07/29 02:33:03 INFO filecache.TrackerDistributedCacheManager: Cached hdfs://192.168.1.101:9000/tmp/mysql-connector-java-5.0.8-bin.jar as /tmp/hadoop-june/mapred/local/archive/-8943686319031389138_-1232673160_640840668/192.168.1.101/tmp/mysql-connector-java-5.0.8-bin.jar13/07/29 02:33:03 INFO filecache.TrackerDistributedCacheManager: Cached hdfs://192.168.1.101:9000/tmp/mysql-connector-java-5.0.8-bin.jar as /tmp/hadoop-june/mapred/local/archive/-8943686319031389138_-1232673160_640840668/192.168.1.101/tmp/mysql-connector-java-5.0.8-bin.jar13/07/29 02:33:03 INFO mapred.JobClient: Running job: job_local_000113/07/29 02:33:03 INFO mapred.MapTask: numReduceTasks: 113/07/29 02:33:03 INFO mapred.MapTask: io.sort.mb = 10013/07/29 02:33:03 INFO mapred.MapTask: data buffer = 79691776/9961472013/07/29 02:33:03 INFO mapred.MapTask: record buffer = 262144/32768013/07/29 02:33:03 INFO mapred.MapTask: Starting flush of map output13/07/29 02:33:03 INFO mapred.MapTask: Finished spill 013/07/29 02:33:03 INFO mapred.Task: Task:attempt_local_0001_m_000000_0 is done. And is in the process of commiting13/07/29 02:33:04 INFO mapred.JobClient:  map 0% reduce 0%13/07/29 02:33:06 INFO mapred.LocalJobRunner: 13/07/29 02:33:06 INFO mapred.Task: Task 'attempt_local_0001_m_000000_0' done.13/07/29 02:33:06 INFO mapred.LocalJobRunner: 13/07/29 02:33:06 INFO mapred.Merger: Merging 1 sorted segments13/07/29 02:33:06 INFO mapred.Merger: Down to the last merge-pass, with 1 segments left of total size: 235 bytes13/07/29 02:33:06 INFO mapred.LocalJobRunner: 13/07/29 02:33:06 INFO mapred.Task: Task:attempt_local_0001_r_000000_0 is done. And is in the process of commiting13/07/29 02:33:07 INFO mapred.JobClient:  map 100% reduce 0%13/07/29 02:33:09 INFO mapred.LocalJobRunner: reduce > reduce13/07/29 02:33:09 INFO mapred.Task: Task 'attempt_local_0001_r_000000_0' done.13/07/29 02:33:09 WARN mapred.FileOutputCommitter: Output path is null in cleanup13/07/29 02:33:10 INFO mapred.JobClient:  map 100% reduce 100%13/07/29 02:33:10 INFO mapred.JobClient: Job complete: job_local_000113/07/29 02:33:10 INFO mapred.JobClient: Counters: 1813/07/29 02:33:10 INFO mapred.JobClient:   File Input Format Counters 13/07/29 02:33:10 INFO mapred.JobClient:     Bytes Read=013/07/29 02:33:10 INFO mapred.JobClient:   File Output Format Counters 13/07/29 02:33:10 INFO mapred.JobClient:     Bytes Written=013/07/29 02:33:10 INFO mapred.JobClient:   FileSystemCounters13/07/29 02:33:10 INFO mapred.JobClient:     FILE_BYTES_READ=121169113/07/29 02:33:10 INFO mapred.JobClient:     HDFS_BYTES_READ=108170413/07/29 02:33:10 INFO mapred.JobClient:     FILE_BYTES_WRITTEN=239284413/07/29 02:33:10 INFO mapred.JobClient:   Map-Reduce Framework13/07/29 02:33:10 INFO mapred.JobClient:     Map output materialized bytes=23913/07/29 02:33:10 INFO mapred.JobClient:     Map input records=1313/07/29 02:33:10 INFO mapred.JobClient:     Reduce shuffle bytes=013/07/29 02:33:10 INFO mapred.JobClient:     Spilled Records=2613/07/29 02:33:10 INFO mapred.JobClient:     Map output bytes=20713/07/29 02:33:10 INFO mapred.JobClient:     Map input bytes=1313/07/29 02:33:10 INFO mapred.JobClient:     SPLIT_RAW_BYTES=7513/07/29 02:33:10 INFO mapred.JobClient:     Combine input records=013/07/29 02:33:10 INFO mapred.JobClient:     Reduce input records=1313/07/29 02:33:10 INFO mapred.JobClient:     Reduce input groups=1313/07/29 02:33:10 INFO mapred.JobClient:     Combine output records=013/07/29 02:33:10 INFO mapred.JobClient:     Reduce output records=1313/07/29 02:33:10 INFO mapred.JobClient:     Map output records=13