001    /**
002     * Licensed to the Apache Software Foundation (ASF) under one
003     * or more contributor license agreements.  See the NOTICE file
004     * distributed with this work for additional information
005     * regarding copyright ownership.  The ASF licenses this file
006     * to you under the Apache License, Version 2.0 (the
007     * "License"); you may not use this file except in compliance
008     * with the License.  You may obtain a copy of the License at
009     *
010     *     http://www.apache.org/licenses/LICENSE-2.0
011     *
012     * Unless required by applicable law or agreed to in writing, software
013     * distributed under the License is distributed on an "AS IS" BASIS,
014     * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
015     * See the License for the specific language governing permissions and
016     * limitations under the License.
017     */
018    
019    package org.apache.hadoop.mapreduce.lib.partition;
020    
021    import org.apache.hadoop.classification.InterfaceAudience;
022    import org.apache.hadoop.classification.InterfaceStability;
023    import org.apache.hadoop.conf.Configurable;
024    import org.apache.hadoop.conf.Configuration;
025    import org.apache.hadoop.io.BinaryComparable;
026    import org.apache.hadoop.io.WritableComparator;
027    import org.apache.hadoop.mapreduce.Partitioner;
028    
029    /**
030     * <p>Partition {@link BinaryComparable} keys using a configurable part of 
031     * the bytes array returned by {@link BinaryComparable#getBytes()}.</p>
032     * 
033     * <p>The subarray to be used for the partitioning can be defined by means
034     * of the following properties:
035     * <ul>
036     *   <li>
037     *     <i>mapreduce.partition.binarypartitioner.left.offset</i>:
038     *     left offset in array (0 by default)
039     *   </li>
040     *   <li>
041     *     <i>mapreduce.partition.binarypartitioner.right.offset</i>: 
042     *     right offset in array (-1 by default)
043     *   </li>
044     * </ul>
045     * Like in Python, both negative and positive offsets are allowed, but
046     * the meaning is slightly different. In case of an array of length 5,
047     * for instance, the possible offsets are:
048     * <pre><code>
049     *  +---+---+---+---+---+
050     *  | B | B | B | B | B |
051     *  +---+---+---+---+---+
052     *    0   1   2   3   4
053     *   -5  -4  -3  -2  -1
054     * </code></pre>
055     * The first row of numbers gives the position of the offsets 0...5 in 
056     * the array; the second row gives the corresponding negative offsets. 
057     * Contrary to Python, the specified subarray has byte <code>i</code> 
058     * and <code>j</code> as first and last element, repectively, when 
059     * <code>i</code> and <code>j</code> are the left and right offset.
060     * 
061     * <p>For Hadoop programs written in Java, it is advisable to use one of 
062     * the following static convenience methods for setting the offsets:
063     * <ul>
064     *   <li>{@link #setOffsets}</li>
065     *   <li>{@link #setLeftOffset}</li>
066     *   <li>{@link #setRightOffset}</li>
067     * </ul></p>
068     */
069    @InterfaceAudience.Public
070    @InterfaceStability.Evolving
071    public class BinaryPartitioner<V> extends Partitioner<BinaryComparable, V> 
072      implements Configurable {
073    
074      public static final String LEFT_OFFSET_PROPERTY_NAME = 
075        "mapreduce.partition.binarypartitioner.left.offset";
076      public static final String RIGHT_OFFSET_PROPERTY_NAME = 
077        "mapreduce.partition.binarypartitioner.right.offset";
078      
079      /**
080       * Set the subarray to be used for partitioning to 
081       * <code>bytes[left:(right+1)]</code> in Python syntax.
082       * 
083       * @param conf configuration object
084       * @param left left Python-style offset
085       * @param right right Python-style offset
086       */
087      public static void setOffsets(Configuration conf, int left, int right) {
088        conf.setInt(LEFT_OFFSET_PROPERTY_NAME, left);
089        conf.setInt(RIGHT_OFFSET_PROPERTY_NAME, right);
090      }
091      
092      /**
093       * Set the subarray to be used for partitioning to 
094       * <code>bytes[offset:]</code> in Python syntax.
095       * 
096       * @param conf configuration object
097       * @param offset left Python-style offset
098       */
099      public static void setLeftOffset(Configuration conf, int offset) {
100        conf.setInt(LEFT_OFFSET_PROPERTY_NAME, offset);
101      }
102      
103      /**
104       * Set the subarray to be used for partitioning to 
105       * <code>bytes[:(offset+1)]</code> in Python syntax.
106       * 
107       * @param conf configuration object
108       * @param offset right Python-style offset
109       */
110      public static void setRightOffset(Configuration conf, int offset) {
111        conf.setInt(RIGHT_OFFSET_PROPERTY_NAME, offset);
112      }
113      
114      
115      private Configuration conf;
116      private int leftOffset, rightOffset;
117      
118      public void setConf(Configuration conf) {
119        this.conf = conf;
120        leftOffset = conf.getInt(LEFT_OFFSET_PROPERTY_NAME, 0);
121        rightOffset = conf.getInt(RIGHT_OFFSET_PROPERTY_NAME, -1);
122      }
123      
124      public Configuration getConf() {
125        return conf;
126      }
127      
128      /** 
129       * Use (the specified slice of the array returned by) 
130       * {@link BinaryComparable#getBytes()} to partition. 
131       */
132      @Override
133      public int getPartition(BinaryComparable key, V value, int numPartitions) {
134        int length = key.getLength();
135        int leftIndex = (leftOffset + length) % length;
136        int rightIndex = (rightOffset + length) % length;
137        int hash = WritableComparator.hashBytes(key.getBytes(), 
138          leftIndex, rightIndex - leftIndex + 1);
139        return (hash & Integer.MAX_VALUE) % numPartitions;
140      }
141      
142    }