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：
1CREATE TABLE `t` ( 2`id` int DEFAULT NULL, 3`name` varchar(10) DEFAULT NULL 4) ENGINE=InnoDB DEFAULT CHARSET=utf8; 5 6CREATE TABLE `t2` ( 7`id` int DEFAULT NULL, 8`name` varchar(10) DEFAULT NULL 9) ENGINE=InnoDB DEFAULT CHARSET=utf8; 10 11insert into t values (1,"june"),(2,"decli"),(3,"hello"), 12 (4,"june"),(5,"decli"),(6,"hello"),(7,"june"), 13 (8,"decli"),(9,"hello"),(10,"june"), 14 (11,"june"),(12,"decli"),(13,"hello");
5、代码：
1import java.io.DataInput; 2import java.io.DataOutput; 3import java.io.IOException; 4import java.sql.PreparedStatement; 5import java.sql.ResultSet; 6import java.sql.SQLException; 7import java.util.Iterator; 8 9import org.apache.hadoop.filecache.DistributedCache; 10import org.apache.hadoop.fs.Path; 11import org.apache.hadoop.io.LongWritable; 12import org.apache.hadoop.io.Text; 13import org.apache.hadoop.io.Writable; 14import org.apache.hadoop.mapred.JobClient; 15import org.apache.hadoop.mapred.JobConf; 16import org.apache.hadoop.mapred.MapReduceBase; 17import org.apache.hadoop.mapred.Mapper; 18import org.apache.hadoop.mapred.OutputCollector; 19import org.apache.hadoop.mapred.Reducer; 20import org.apache.hadoop.mapred.Reporter; 21import org.apache.hadoop.mapred.lib.IdentityReducer; 22import org.apache.hadoop.mapred.lib.db.DBConfiguration; 23import org.apache.hadoop.mapred.lib.db.DBInputFormat; 24import org.apache.hadoop.mapred.lib.db.DBOutputFormat; 25import org.apache.hadoop.mapred.lib.db.DBWritable; 26 27/** 28 * 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 29 * 实际当中，可能只需要从 mysql 读，或者写到 mysql 中。 30 * date: 2013-7-29 上午2:34:04 <br/> 31 * @author june 32 */ 33public class Mysql2Mr { 34 // DROP TABLE IF EXISTS `hadoop`.`studentinfo`; 35 // CREATE TABLE studentinfo ( 36 // id INTEGER NOT NULL PRIMARY KEY, 37 // name VARCHAR(32) NOT NULL); 38 39 public static class StudentinfoRecord implements Writable, DBWritable { 40 int id; 41 String name; 42 43 public StudentinfoRecord() { 44 45 } 46 47 public void readFields(DataInput in) throws IOException { 48 this.id = in.readInt(); 49 this.name = Text.readString(in); 50 } 51 52 public String toString() { 53 return new String(this.id + " " + this.name); 54 } 55 56 @Override 57 public void write(PreparedStatement stmt) throws SQLException { 58 stmt.setInt(1, this.id); 59 stmt.setString(2, this.name); 60 } 61 62 @Override 63 public void readFields(ResultSet result) throws SQLException { 64 this.id = result.getInt(1); 65 this.name = result.getString(2); 66 } 67 68 @Override 69 public void write(DataOutput out) throws IOException { 70 out.writeInt(this.id); 71 Text.writeString(out, this.name); 72 } 73 } 74 75 // 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： 76 // Caused by: java.lang.NoSuchMethodException: DBInputMapper.<init>() 77 // http://stackoverflow.com/questions/7154125/custom-mapreduce-input-format-cant-find-constructor 78 // 网上脑残式的转帖，没见到一个写对的。。。 79 public static class DBInputMapper extends MapReduceBase implements 80 Mapper<LongWritable, StudentinfoRecord, LongWritable, Text> { 81 public void map(LongWritable key, StudentinfoRecord value, 82 OutputCollector<LongWritable, Text> collector, Reporter reporter) throws IOException { 83 collector.collect(new LongWritable(value.id), new Text(value.toString())); 84 } 85 } 86 87 public static class MyReducer extends MapReduceBase implements 88 Reducer<LongWritable, Text, StudentinfoRecord, Text> { 89 @Override 90 public void reduce(LongWritable key, Iterator<Text> values, 91 OutputCollector<StudentinfoRecord, Text> output, Reporter reporter) throws IOException { 92 String[] splits = values.next().toString().split(" "); 93 StudentinfoRecord r = new StudentinfoRecord(); 94 r.id = Integer.parseInt(splits[0]); 95 r.name = splits[1]; 96 output.collect(r, new Text(r.name)); 97 } 98 } 99 100 public static void main(String[] args) throws IOException { 101 JobConf conf = new JobConf(Mysql2Mr.class); 102 DistributedCache.addFileToClassPath(new Path("/tmp/mysql-connector-java-5.0.8-bin.jar"), conf); 103 104 conf.setMapOutputKeyClass(LongWritable.class); 105 conf.setMapOutputValueClass(Text.class); 106 conf.setOutputKeyClass(LongWritable.class); 107 conf.setOutputValueClass(Text.class); 108 109 conf.setOutputFormat(DBOutputFormat.class); 110 conf.setInputFormat(DBInputFormat.class); 111 // // mysql to hdfs 112 // conf.setReducerClass(IdentityReducer.class); 113 // Path outPath = new Path("/tmp/1"); 114 // FileSystem.get(conf).delete(outPath, true); 115 // FileOutputFormat.setOutputPath(conf, outPath); 116 117 DBConfiguration.configureDB(conf, "com.mysql.jdbc.Driver", "jdbc:mysql://192.168.1.101:3306/test", 118 "root", "root"); 119 String[] fields = { "id", "name" }; 120 // 从 t 表读数据 121 DBInputFormat.setInput(conf, StudentinfoRecord.class, "t", null, "id", fields); 122 // mapreduce 将数据输出到 t2 表 123 DBOutputFormat.setOutput(conf, "t2", "id", "name"); 124 // conf.setMapperClass(org.apache.hadoop.mapred.lib.IdentityMapper.class); 125 conf.setMapperClass(DBInputMapper.class); 126 conf.setReducerClass(MyReducer.class); 127 128 JobClient.runJob(conf); 129 } 130}
6、结果：
执行两次后，你可以看到mysql结果：
1mysql> select * from t2; 2+------+-------+ 3| id | name | 4+------+-------+ 5| 1 | june | 6| 2 | decli | 7| 3 | hello | 8| 4 | june | 9| 5 | decli | 10| 6 | hello | 11| 7 | june | 12| 8 | decli | 13| 9 | hello | 14| 10 | june | 15| 11 | june | 16| 12 | decli | 17| 13 | hello | 18| 1 | june | 19| 2 | decli | 20| 3 | hello | 21| 4 | june | 22| 5 | decli | 23| 6 | hello | 24| 7 | june | 25| 8 | decli | 26| 9 | hello | 27| 10 | june | 28| 11 | june | 29| 12 | decli | 30| 13 | hello | 31+------+-------+ 3226 rows in set (0.00 sec) 33 34mysql>
7、日志：
113/07/29 02:33:03 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same. 213/07/29 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-x 313/07/29 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.jar 413/07/29 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.jar 513/07/29 02:33:03 INFO mapred.JobClient: Running job: job_local_0001 613/07/29 02:33:03 INFO mapred.MapTask: numReduceTasks: 1 713/07/29 02:33:03 INFO mapred.MapTask: io.sort.mb = 100 813/07/29 02:33:03 INFO mapred.MapTask: data buffer = 79691776/99614720 913/07/29 02:33:03 INFO mapred.MapTask: record buffer = 262144/327680 1013/07/29 02:33:03 INFO mapred.MapTask: Starting flush of map output 1113/07/29 02:33:03 INFO mapred.MapTask: Finished spill 0 1213/07/29 02:33:03 INFO mapred.Task: Task:attempt_local_0001_m_000000_0 is done. And is in the process of commiting 1313/07/29 02:33:04 INFO mapred.JobClient: map 0% reduce 0% 1413/07/29 02:33:06 INFO mapred.LocalJobRunner: 1513/07/29 02:33:06 INFO mapred.Task: Task 'attempt_local_0001_m_000000_0' done. 1613/07/29 02:33:06 INFO mapred.LocalJobRunner: 1713/07/29 02:33:06 INFO mapred.Merger: Merging 1 sorted segments 1813/07/29 02:33:06 INFO mapred.Merger: Down to the last merge-pass, with 1 segments left of total size: 235 bytes 1913/07/29 02:33:06 INFO mapred.LocalJobRunner: 2013/07/29 02:33:06 INFO mapred.Task: Task:attempt_local_0001_r_000000_0 is done. And is in the process of commiting 2113/07/29 02:33:07 INFO mapred.JobClient: map 100% reduce 0% 2213/07/29 02:33:09 INFO mapred.LocalJobRunner: reduce > reduce 2313/07/29 02:33:09 INFO mapred.Task: Task 'attempt_local_0001_r_000000_0' done. 2413/07/29 02:33:09 WARN mapred.FileOutputCommitter: Output path is null in cleanup 2513/07/29 02:33:10 INFO mapred.JobClient: map 100% reduce 100% 2613/07/29 02:33:10 INFO mapred.JobClient: Job complete: job_local_0001 2713/07/29 02:33:10 INFO mapred.JobClient: Counters: 18 2813/07/29 02:33:10 INFO mapred.JobClient: File Input Format Counters 2913/07/29 02:33:10 INFO mapred.JobClient: Bytes Read=0 3013/07/29 02:33:10 INFO mapred.JobClient: File Output Format Counters 3113/07/29 02:33:10 INFO mapred.JobClient: Bytes Written=0 3213/07/29 02:33:10 INFO mapred.JobClient: FileSystemCounters 3313/07/29 02:33:10 INFO mapred.JobClient: FILE_BYTES_READ=1211691 3413/07/29 02:33:10 INFO mapred.JobClient: HDFS_BYTES_READ=1081704 3513/07/29 02:33:10 INFO mapred.JobClient: FILE_BYTES_WRITTEN=2392844 3613/07/29 02:33:10 INFO mapred.JobClient: Map-Reduce Framework 3713/07/29 02:33:10 INFO mapred.JobClient: Map output materialized bytes=239 3813/07/29 02:33:10 INFO mapred.JobClient: Map input records=13 3913/07/29 02:33:10 INFO mapred.JobClient: Reduce shuffle bytes=0 4013/07/29 02:33:10 INFO mapred.JobClient: Spilled Records=26 4113/07/29 02:33:10 INFO mapred.JobClient: Map output bytes=207 4213/07/29 02:33:10 INFO mapred.JobClient: Map input bytes=13 4313/07/29 02:33:10 INFO mapred.JobClient: SPLIT_RAW_BYTES=75 4413/07/29 02:33:10 INFO mapred.JobClient: Combine input records=0 4513/07/29 02:33:10 INFO mapred.JobClient: Reduce input records=13 4613/07/29 02:33:10 INFO mapred.JobClient: Reduce input groups=13 4713/07/29 02:33:10 INFO mapred.JobClient: Combine output records=0 4813/07/29 02:33:10 INFO mapred.JobClient: Reduce output records=13 4913/07/29 02:33:10 INFO mapred.JobClient: Map output records=13
8、REF：
新版 API 写法：
http://superlxw1234.iteye.com/blog/1880712
老版：