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
019package org.apache.hadoop.mapred.lib;
020
021import java.io.IOException;
022import java.util.ArrayList;
023
024import org.apache.hadoop.classification.InterfaceAudience;
025import org.apache.hadoop.classification.InterfaceStability;
026import org.apache.hadoop.fs.FileStatus;
027import org.apache.hadoop.fs.Path;
028import org.apache.hadoop.io.LongWritable;
029import org.apache.hadoop.io.Text;
030import org.apache.hadoop.mapred.FileInputFormat;
031import org.apache.hadoop.mapred.FileSplit;
032import org.apache.hadoop.mapred.InputSplit;
033import org.apache.hadoop.mapred.JobConf;
034import org.apache.hadoop.mapred.JobConfigurable;
035import org.apache.hadoop.mapred.LineRecordReader;
036import org.apache.hadoop.mapred.RecordReader;
037import org.apache.hadoop.mapred.Reporter;
038
039/**
040 * NLineInputFormat which splits N lines of input as one split.
041 *
042 * In many "pleasantly" parallel applications, each process/mapper 
043 * processes the same input file (s), but with computations are 
044 * controlled by different parameters.(Referred to as "parameter sweeps").
045 * One way to achieve this, is to specify a set of parameters 
046 * (one set per line) as input in a control file 
047 * (which is the input path to the map-reduce application,
048 * where as the input dataset is specified 
049 * via a config variable in JobConf.).
050 * 
051 * The NLineInputFormat can be used in such applications, that splits 
052 * the input file such that by default, one line is fed as
053 * a value to one map task, and key is the offset.
054 * i.e. (k,v) is (LongWritable, Text).
055 * The location hints will span the whole mapred cluster.
056 */
057@InterfaceAudience.Public
058@InterfaceStability.Stable
059public class NLineInputFormat extends FileInputFormat<LongWritable, Text> 
060                              implements JobConfigurable { 
061  private int N = 1;
062
063  public RecordReader<LongWritable, Text> getRecordReader(
064                                            InputSplit genericSplit,
065                                            JobConf job,
066                                            Reporter reporter) 
067  throws IOException {
068    reporter.setStatus(genericSplit.toString());
069    return new LineRecordReader(job, (FileSplit) genericSplit);
070  }
071
072  /** 
073   * Logically splits the set of input files for the job, splits N lines
074   * of the input as one split.
075   * 
076   * @see org.apache.hadoop.mapred.FileInputFormat#getSplits(JobConf, int)
077   */
078  public InputSplit[] getSplits(JobConf job, int numSplits)
079  throws IOException {
080    ArrayList<FileSplit> splits = new ArrayList<FileSplit>();
081    for (FileStatus status : listStatus(job)) {
082      for (org.apache.hadoop.mapreduce.lib.input.FileSplit split : 
083          org.apache.hadoop.mapreduce.lib.input.
084          NLineInputFormat.getSplitsForFile(status, job, N)) {
085        splits.add(new FileSplit(split));
086      }
087    }
088    return splits.toArray(new FileSplit[splits.size()]);
089  }
090
091  public void configure(JobConf conf) {
092    N = conf.getInt("mapreduce.input.lineinputformat.linespermap", 1);
093  }
094  
095  /**
096   * NLineInputFormat uses LineRecordReader, which always reads
097   * (and consumes) at least one character out of its upper split
098   * boundary. So to make sure that each mapper gets N lines, we
099   * move back the upper split limits of each split 
100   * by one character here.
101   * @param fileName  Path of file
102   * @param begin  the position of the first byte in the file to process
103   * @param length  number of bytes in InputSplit
104   * @return  FileSplit
105   */
106  protected static FileSplit createFileSplit(Path fileName, long begin, long length) {
107    return (begin == 0) 
108    ? new FileSplit(fileName, begin, length - 1, new String[] {})
109    : new FileSplit(fileName, begin - 1, length, new String[] {});
110  }
111}