source: sample/HBaseRecord.java @ 7

/*
 *  NCHC Hbase with map reduce sample code 
 *  DemoHBaseSlink.java
 */

package tw.org.nchc.demo;

import java.io.IOException;
import java.util.Iterator;

import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
import org.apache.hadoop.hbase.mapred.TableReduce;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.MapWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.JobClient;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.mapred.lib.IdentityMapper;
import org.apache.hadoop.mapred.lib.IdentityReducer;

/**
 * This sample code will put the indicate data to Hbase.
 * 1. put test.txt in t1 directory which content is 
---------------
name:locate:years 
waue:taiwan:1981
shellon:taiwan:1981
---------------
 * 2. hadoop_root/$ bin/hadoop dfs -put t1 t1
 * 3. hbase_root/$ bin/hbase shell
 * 4. hql > create table t1_table("person");
 * 5. Come to Eclipse and run this code, and we will let database as that 
 t1_table -> person
  ----------------
  |  name | locate | years |
  ----------------
  | waue  | taiwan | 1981 |
  ----------------
  | shellon | taiwan | 1981 |
  * 6. Go to hbase console, type : hql > select * from t1_table;
     
08/06/06 12:20:48 INFO hbase.HTable: Creating scanner over t1_table starting at key 
+-------------------------+-------------------------+-------------------------+
| Row                     | Column                  | Cell                    |
+-------------------------+-------------------------+-------------------------+
| 0                       | person:locate           | locate                  |
+-------------------------+-------------------------+-------------------------+
| 0                       | person:name             | name                    |
+-------------------------+-------------------------+-------------------------+
| 0                       | person:years            | years                   |
+-------------------------+-------------------------+-------------------------+
| 19                      | person:locate           | taiwan                  |
+-------------------------+-------------------------+-------------------------+
| 19                      | person:name             | waue                    |
+-------------------------+-------------------------+-------------------------+
| 19                      | person:years            | 1981                    |
+-------------------------+-------------------------+-------------------------+
| 36                      | person:locate           | taiwan                  |
+-------------------------+-------------------------+-------------------------+
| 36                      | person:name             | shellon                 |
+-------------------------+-------------------------+-------------------------+
| 36                      | person:years            | 1981                    |
+-------------------------+-------------------------+-------------------------+
3 row(s) in set. (0.04 sec)
 **/
public class HBaseRecord {

  /* Denify parameter */
  // one column family: person; three column qualifier: name,locate,years
  static private String  baseId1 = "person:name";
  static private String  baseId2 ="person:locate";
  static private String  baseId3 ="person:years";
  //split character
  static private String sp = ":";
  // file path in hadoop file system (not phisical file system)
  String file_path = "/user/waue/t1";
  // Hbase table name
  String table_name = "t1_table";
  // setup MapTask and Reduce Task
  int mapTasks = 1;
  int reduceTasks = 1;
  
  private static class ReduceClass extends TableReduce<LongWritable, Text> {

    // Column id is created dymanically, 
    private static final Text col_name = new Text(baseId1);
    private static final Text col_local = new Text(baseId2);
    private static final Text col_year = new Text(baseId3);
    
    // this map holds the columns per row
    private MapWritable map = new MapWritable();  
    
    // on this sample, map is nonuse, we use reduce to handle
    public void reduce(LongWritable key, Iterator<Text> values,
        OutputCollector<Text, MapWritable> output, Reporter reporter)
        throws IOException {

      // values.next().getByte() can get value and transfer to byte form, there is an other way that let decode()
      // to substitude getByte() 
      String stro = new String(values.next().getBytes());
      String str[] = stro.split(sp);
      byte b_local[] = str[0].getBytes();
      byte b_name[] = str[1].getBytes();
      byte b_year[] = str[2].getBytes();
      
      // contents must be ImmutableBytesWritable
      ImmutableBytesWritable w_local = new ImmutableBytesWritable( b_local);
      ImmutableBytesWritable w_name = new ImmutableBytesWritable( b_name );
      ImmutableBytesWritable w_year = new ImmutableBytesWritable( b_year );

      // populate the current row
      map.clear();
      map.put(col_name, w_local);
      map.put(col_local, w_name);
      map.put(col_year, w_year);

      // add the row with the key as the row id
      output.collect(new Text(key.toString()), map);
    }
  }

  private HBaseRecord() {
  }

  /**
   * Runs the demo.
   */
  public static void main(String[] args) throws IOException {
    // which path of input files in Hadoop file system  

    
    HBaseRecord setup = new HBaseRecord();
    JobConf conf = new JobConf(HBaseRecord.class);

    //Job name; you can modify to any you like  
    conf.setJobName("NCHC_PersonDataBase");

    // Hbase table name must be correct , in our profile is t1_table
    TableReduce.initJob(setup.table_name, ReduceClass.class, conf);
    
    // below are map-reduce profile
    conf.setNumMapTasks(setup.mapTasks);
    conf.setNumReduceTasks(setup.reduceTasks);
    conf.setInputPath(new Path(setup.file_path));
    conf.setMapperClass(IdentityMapper.class);
    conf.setCombinerClass(IdentityReducer.class);
    conf.setReducerClass(ReduceClass.class);
    JobClient.runJob(conf);
  }
}