View Javadoc

1   /**
2    * Copyright 2008 The Apache Software Foundation
3    *
4    * Licensed to the Apache Software Foundation (ASF) under one
5    * or more contributor license agreements.  See the NOTICE file
6    * distributed with this work for additional information
7    * regarding copyright ownership.  The ASF licenses this file
8    * to you under the Apache License, Version 2.0 (the
9    * "License"); you may not use this file except in compliance
10   * with the License.  You may obtain a copy of the License at
11   *
12   *     http://www.apache.org/licenses/LICENSE-2.0
13   *
14   * Unless required by applicable law or agreed to in writing, software
15   * distributed under the License is distributed on an "AS IS" BASIS,
16   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17   * See the License for the specific language governing permissions and
18   * limitations under the License.
19   */
20  package org.apache.hadoop.hbase.mapreduce;
21  
22  import java.io.IOException;
23  
24  import org.apache.commons.logging.Log;
25  import org.apache.commons.logging.LogFactory;
26  import org.apache.hadoop.conf.Configurable;
27  import org.apache.hadoop.conf.Configuration;
28  import org.apache.hadoop.hbase.HBaseConfiguration;
29  import org.apache.hadoop.hbase.client.HTable;
30  import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
31  import org.apache.hadoop.hbase.util.Bytes;
32  import org.apache.hadoop.mapreduce.Partitioner;
33  
34  /**
35   * This is used to partition the output keys into groups of keys.
36   * Keys are grouped according to the regions that currently exist
37   * so that each reducer fills a single region so load is distributed.
38   *
39   * @param <KEY>  The type of the key.
40   * @param <VALUE>  The type of the value.
41   */
42  public class HRegionPartitioner<KEY, VALUE>
43  extends Partitioner<ImmutableBytesWritable, VALUE>
44  implements Configurable {
45  
46    private final Log LOG = LogFactory.getLog(TableInputFormat.class);
47    private Configuration conf = null;
48    private HTable table;
49    private byte[][] startKeys;
50  
51    /**
52     * Gets the partition number for a given key (hence record) given the total
53     * number of partitions i.e. number of reduce-tasks for the job.
54     *
55     * <p>Typically a hash function on a all or a subset of the key.</p>
56     *
57     * @param key  The key to be partitioned.
58     * @param value  The entry value.
59     * @param numPartitions  The total number of partitions.
60     * @return The partition number for the <code>key</code>.
61     * @see org.apache.hadoop.mapreduce.Partitioner#getPartition(
62     *   java.lang.Object, java.lang.Object, int)
63     */
64    @Override
65    public int getPartition(ImmutableBytesWritable key,
66        VALUE value, int numPartitions) {
67      byte[] region = null;
68      // Only one region return 0
69      if (this.startKeys.length == 1){
70        return 0;
71      }
72      try {
73        // Not sure if this is cached after a split so we could have problems
74        // here if a region splits while mapping
75        region = table.getRegionLocation(key.get()).getRegionInfo().getStartKey();
76      } catch (IOException e) {
77        LOG.error(e);
78      }
79      for (int i = 0; i < this.startKeys.length; i++){
80        if (Bytes.compareTo(region, this.startKeys[i]) == 0 ){
81          if (i >= numPartitions-1){
82            // cover if we have less reduces then regions.
83            return (Integer.toString(i).hashCode()
84                & Integer.MAX_VALUE) % numPartitions;
85          }
86          return i;
87        }
88      }
89      // if above fails to find start key that match we need to return something
90      return 0;
91    }
92  
93    /**
94     * Returns the current configuration.
95     *
96     * @return The current configuration.
97     * @see org.apache.hadoop.conf.Configurable#getConf()
98     */
99    @Override
100   public Configuration getConf() {
101     return conf;
102   }
103 
104   /**
105    * Sets the configuration. This is used to determine the start keys for the
106    * given table.
107    *
108    * @param configuration  The configuration to set.
109    * @see org.apache.hadoop.conf.Configurable#setConf(
110    *   org.apache.hadoop.conf.Configuration)
111    */
112   @Override
113   public void setConf(Configuration configuration) {
114     this.conf = configuration;
115     try {
116       HBaseConfiguration.addHbaseResources(conf);
117       this.table = new HTable(this.conf,
118         configuration.get(TableOutputFormat.OUTPUT_TABLE));
119     } catch (IOException e) {
120       LOG.error(e);
121     }
122     try {
123       this.startKeys = this.table.getStartKeys();
124     } catch (IOException e) {
125       LOG.error(e);
126     }
127   }
128 }