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   * <p>This class is not suitable as partitioner creating hfiles
40   * for incremental bulk loads as region spread will likely change between time of
41   * hfile creation and load time. See {@link LoadIncrementalHFiles}
42   * and <a href="http://hbase.apache.org/docs/current/bulk-loads.html">Bulk Load</a>.
43   *
44   * @param <KEY>  The type of the key.
45   * @param <VALUE>  The type of the value.
46   */
47  public class HRegionPartitioner<KEY, VALUE>
48  extends Partitioner<ImmutableBytesWritable, VALUE>
49  implements Configurable {
50  
51    private final Log LOG = LogFactory.getLog(TableInputFormat.class);
52    private Configuration conf = null;
53    private HTable table;
54    private byte[][] startKeys;
55  
56    /**
57     * Gets the partition number for a given key (hence record) given the total
58     * number of partitions i.e. number of reduce-tasks for the job.
59     *
60     * <p>Typically a hash function on a all or a subset of the key.</p>
61     *
62     * @param key  The key to be partitioned.
63     * @param value  The entry value.
64     * @param numPartitions  The total number of partitions.
65     * @return The partition number for the <code>key</code>.
66     * @see org.apache.hadoop.mapreduce.Partitioner#getPartition(
67     *   java.lang.Object, java.lang.Object, int)
68     */
69    @Override
70    public int getPartition(ImmutableBytesWritable key,
71        VALUE value, int numPartitions) {
72      byte[] region = null;
73      // Only one region return 0
74      if (this.startKeys.length == 1){
75        return 0;
76      }
77      try {
78        // Not sure if this is cached after a split so we could have problems
79        // here if a region splits while mapping
80        region = table.getRegionLocation(key.get()).getRegionInfo().getStartKey();
81      } catch (IOException e) {
82        LOG.error(e);
83      }
84      for (int i = 0; i < this.startKeys.length; i++){
85        if (Bytes.compareTo(region, this.startKeys[i]) == 0 ){
86          if (i >= numPartitions-1){
87            // cover if we have less reduces then regions.
88            return (Integer.toString(i).hashCode()
89                & Integer.MAX_VALUE) % numPartitions;
90          }
91          return i;
92        }
93      }
94      // if above fails to find start key that match we need to return something
95      return 0;
96    }
97  
98    /**
99     * Returns the current configuration.
100    *
101    * @return The current configuration.
102    * @see org.apache.hadoop.conf.Configurable#getConf()
103    */
104   @Override
105   public Configuration getConf() {
106     return conf;
107   }
108 
109   /**
110    * Sets the configuration. This is used to determine the start keys for the
111    * given table.
112    *
113    * @param configuration  The configuration to set.
114    * @see org.apache.hadoop.conf.Configurable#setConf(
115    *   org.apache.hadoop.conf.Configuration)
116    */
117   @Override
118   public void setConf(Configuration configuration) {
119     this.conf = HBaseConfiguration.create(configuration);
120     try {
121       this.table = new HTable(this.conf,
122         configuration.get(TableOutputFormat.OUTPUT_TABLE));
123     } catch (IOException e) {
124       LOG.error(e);
125     }
126     try {
127       this.startKeys = this.table.getStartKeys();
128     } catch (IOException e) {
129       LOG.error(e);
130     }
131   }
132 }