Hadoop 中利用 mapreduce 讀寫 mysql 數據
有時候我們在項目中會遇到輸入結果集很大，但是輸出結果很小，比如一些 pv、uv 數據，然后為了實時查詢的需求，或者一些 OLAP 的需求，我們需要 mapreduce 與 mysql 進行數據的交互，而這些是 hbase 或者 hive 目前亟待改進的地方。
好了言歸正傳，簡單的說說背景、原理以及需要注意的地方：
1、為了方便 MapReduce 直接訪問關系型數據庫（Mysql,Oracle），Hadoop提供了DBInputFormat和DBOutputFormat兩個類。通過DBInputFormat類把數據庫表數據讀入到HDFS，根據DBOutputFormat類把MapReduce產生的結果集導入到數據庫表中。
2、由于0.20版本對DBInputFormat和DBOutputFormat支持不是很好，該例用了0.19版本來說明這兩個類的用法。
至少在我的 0.20.203 中的 org.apache.hadoop.mapreduce.lib 下是沒見到 db 包，所以本文也是以老版的 API 來為例說明的。
3、運行MapReduce時候報錯：java.io.IOException: com.mysql.jdbc.Driver，一般是由于程序找不到mysql驅動包。解決方法是讓每個tasktracker運行MapReduce程序時都可以找到該驅動包。
添加包有兩種方式：
（1）在每個節點下的${HADOOP_HOME}/lib下添加該包。重啟集群，一般是比較原始的方法。
（2）a)把包傳到集群上： hadoop fs -put mysql-connector-java-5.1.0- bin.jar /hdfsPath/
b)在mr程序提交job前，添加語句：DistributedCache.addFileToClassPath(new Path(“/hdfsPath/mysql- connector-java- 5.1.0-bin.jar”), conf);
（3）雖然API用的是0.19的，但是使用0.20的API一樣可用，只是會提示方法已過時而已。
4、測試數據：
CREATE TABLEt(idint DEFAULT NULL,namevarchar(10) DEFAULT NULL ) ENGINE=InnoDB DEFAULT CHARSET=utf8;CREATE TABLE
t2(idint DEFAULT NULL,namevarchar(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");</pre>
5、代碼：
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: 測試 mr 與 mysql 的數據交互，此測試用例將一個表中的數據復制到另一張表中
- 實際當中，可能只需要從 mysql 讀，或者寫到 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); } @Override public void write(PreparedStatement stmt) throws SQLException { stmt.setInt(1, this.id); stmt.setString(2, this.name); } @Override public void readFields(ResultSet result) throws SQLException { this.id = result.getInt(1); this.name = result.getString(2); } @Override public void write(DataOutput out) throws IOException { out.writeInt(this.id); Text.writeString(out, this.name); }}
// 記住此處是靜態內部類，要不然你自己實現無參構造器，或者等著拋異常： // Caused by: java.lang.NoSuchMethodException: DBInputMapper.<init>() // http://stackoverflow.com/questions/7154125/custom-mapreduce-input-format-cant-find-constructor // 網上腦殘式的轉帖，沒見到一個寫對的。。。 public static class DBInputMapper extends MapReduceBase implements
Mapper<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 implements
Reducer<LongWritable, Text, StudentinfoRecord, Text> { @Override public 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" }; // 從 t 表讀數據 DBInputFormat.setInput(conf, StudentinfoRecord.class, "t", null, "id", fields); // mapreduce 將數據輸出到 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);} }</pre>
6、結果：
執行兩次后，你可以看到mysql結果：
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></pre>
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-x 13/07/29 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 13/07/29 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 13/07/29 02:33:03 INFO mapred.JobClient: Running job: job_local_0001 13/07/29 02:33:03 INFO mapred.MapTask: numReduceTasks: 1 13/07/29 02:33:03 INFO mapred.MapTask: io.sort.mb = 100 13/07/29 02:33:03 INFO mapred.MapTask: data buffer = 79691776/99614720 13/07/29 02:33:03 INFO mapred.MapTask: record buffer = 262144/327680 13/07/29 02:33:03 INFO mapred.MapTask: Starting flush of map output 13/07/29 02:33:03 INFO mapred.MapTask: Finished spill 0 13/07/29 02:33:03 INFO mapred.Task: Task:attempt_local_0001_m_000000_0 is done. And is in the process of commiting 13/07/29 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 segments 13/07/29 02:33:06 INFO mapred.Merger: Down to the last merge-pass, with 1 segments left of total size: 235 bytes 13/07/29 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 commiting 13/07/29 02:33:07 INFO mapred.JobClient: map 100% reduce 0% 13/07/29 02:33:09 INFO mapred.LocalJobRunner: reduce > reduce 13/07/29 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 cleanup 13/07/29 02:33:10 INFO mapred.JobClient: map 100% reduce 100% 13/07/29 02:33:10 INFO mapred.JobClient: Job complete: job_local_0001 13/07/29 02:33:10 INFO mapred.JobClient: Counters: 18 13/07/29 02:33:10 INFO mapred.JobClient: File Input Format Counters 13/07/29 02:33:10 INFO mapred.JobClient: Bytes Read=0 13/07/29 02:33:10 INFO mapred.JobClient: File Output Format Counters 13/07/29 02:33:10 INFO mapred.JobClient: Bytes Written=0 13/07/29 02:33:10 INFO mapred.JobClient: FileSystemCounters 13/07/29 02:33:10 INFO mapred.JobClient: FILE_BYTES_READ=1211691 13/07/29 02:33:10 INFO mapred.JobClient: HDFS_BYTES_READ=1081704 13/07/29 02:33:10 INFO mapred.JobClient: FILE_BYTES_WRITTEN=2392844 13/07/29 02:33:10 INFO mapred.JobClient: Map-Reduce Framework 13/07/29 02:33:10 INFO mapred.JobClient: Map output materialized bytes=239 13/07/29 02:33:10 INFO mapred.JobClient: Map input records=13 13/07/29 02:33:10 INFO mapred.JobClient: Reduce shuffle bytes=0 13/07/29 02:33:10 INFO mapred.JobClient: Spilled Records=26 13/07/29 02:33:10 INFO mapred.JobClient: Map output bytes=207 13/07/29 02:33:10 INFO mapred.JobClient: Map input bytes=13 13/07/29 02:33:10 INFO mapred.JobClient: SPLIT_RAW_BYTES=75 13/07/29 02:33:10 INFO mapred.JobClient: Combine input records=0 13/07/29 02:33:10 INFO mapred.JobClient: Reduce input records=13 13/07/29 02:33:10 INFO mapred.JobClient: Reduce input groups=13 13/07/29 02:33:10 INFO mapred.JobClient: Combine output records=0 13/07/29 02:33:10 INFO mapred.JobClient: Reduce output records=13 13/07/29 02:33:10 INFO mapred.JobClient: Map output records=13
8、REF：
新版 API 寫法：
http://superlxw1234.iteye.com/blog/1880712
老版：
http://blog.csdn.net/dajuezhao/article/details/5799371
http://www.zhengmenbb.com/archives/583.htm