Example usage for org.apache.hadoop.mapred JobConf set

List of usage examples for org.apache.hadoop.mapred JobConf set

  1. HOME
  2. Java
  3. org.apache.hadoop
  4. org.apache.hadoop.mapred.*
  5. JobConf
  6. set

Introduction

In this page you can find the example usage for org.apache.hadoop.mapred JobConf set.

Prototype

public void set(String name, String value)

Source Link

Document

Set the value of the name property.

Usage

From source file:edu.umd.cloud9.pagerank.PartitionGraph.java

License:Apache License

public int run(String[] args) throws IOException {
    if (args.length != 5) {
        printUsage();/*from  w  ww .  j  av  a2s.co m*/
        return -1;
    }

    String inPath = args[0];
    String outPath = args[1];
    int numParts = Integer.parseInt(args[2]);
    boolean useRange = Integer.parseInt(args[3]) != 0;
    int nodeCount = Integer.parseInt(args[4]);

    sLogger.info("Tool name: PartitionGraph");
    sLogger.info(" - inputDir: " + inPath);
    sLogger.info(" - outputDir: " + outPath);
    sLogger.info(" - numPartitions: " + numParts);
    sLogger.info(" - useRange?: " + useRange);
    sLogger.info(" - nodeCnt: " + nodeCount);

    JobConf conf = new JobConf(PartitionGraph.class);

    conf.setJobName("Partition Graph " + numParts);
    conf.setNumReduceTasks(numParts);

    conf.setInt("mapred.min.split.size", 1024 * 1024 * 1024);
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setInt("NodeCount", nodeCount);

    FileInputFormat.setInputPaths(conf, new Path(inPath));
    FileOutputFormat.setOutputPath(conf, new Path(outPath));

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(PageRankNode.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(PageRankNode.class);

    conf.setMapperClass(MapClass.class);
    conf.setReducerClass(ReduceClass.class);

    conf.setSpeculativeExecution(false);

    if (useRange) {
        conf.setPartitionerClass(RangePartitioner.class);
    }

    FileSystem.get(conf).delete(new Path(outPath), true);

    JobClient.runJob(conf);

    return 0;
}

From source file:edu.umd.cloud9.pagerank.RunPageRankBasic.java

License:Apache License

private float phase1(String path, int i, int j, int n, boolean useCombiner, boolean useInmapCombiner,
        boolean useRange) throws IOException {
    JobConf conf = new JobConf(RunPageRankBasic.class);

    String in = path + "/iter" + sFormat.format(i);
    String out = path + "/iter" + sFormat.format(j) + "t";
    String outm = out + "-mass";

    // we need to actually count the number of part files to get the number
    // of partitions (because the directory might contain _log)
    int numPartitions = 0;
    for (FileStatus s : FileSystem.get(conf).listStatus(new Path(in))) {
        if (s.getPath().getName().contains("part-"))
            numPartitions++;/*from ww w  . jav  a2 s  .c  om*/
    }

    sLogger.info("PageRank: iteration " + j + ": Phase1");
    sLogger.info(" - input: " + in);
    sLogger.info(" - output: " + out);
    sLogger.info(" - nodeCnt: " + n);
    sLogger.info(" - useCombiner: " + useCombiner);
    sLogger.info(" - useInmapCombiner: " + useInmapCombiner);
    sLogger.info(" - useRange: " + useRange);
    sLogger.info("computed number of partitions: " + numPartitions);

    int numMapTasks = numPartitions;
    int numReduceTasks = numPartitions;

    conf.setJobName("PageRank:Basic:iteration" + j + ":Phase1");
    conf.setInt("NodeCount", n);

    conf.setNumMapTasks(numMapTasks);
    conf.setNumReduceTasks(numReduceTasks);

    conf.setInt("mapred.min.split.size", 1024 * 1024 * 1024);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.set("PageRankMassPath", outm);

    FileInputFormat.setInputPaths(conf, new Path(in));
    FileOutputFormat.setOutputPath(conf, new Path(out));

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(PageRankNode.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(PageRankNode.class);

    if (useInmapCombiner) {
        conf.setMapperClass(MapWithInMapperCombiningClass.class);
    } else {
        conf.setMapperClass(MapClass.class);
    }

    if (useCombiner) {
        conf.setCombinerClass(CombineClass.class);
    }

    if (useRange) {
        conf.setPartitionerClass(RangePartitioner.class);
    }

    conf.setReducerClass(ReduceClass.class);

    conf.setSpeculativeExecution(false);

    FileSystem.get(conf).delete(new Path(out), true);
    FileSystem.get(conf).delete(new Path(outm), true);

    JobClient.runJob(conf);

    float mass = Float.NEGATIVE_INFINITY;
    FileSystem fs = FileSystem.get(conf);
    for (FileStatus f : fs.listStatus(new Path(outm))) {
        FSDataInputStream fin = fs.open(f.getPath());
        mass = sumLogProbs(mass, fin.readFloat());
        fin.close();
    }

    return mass;
}

From source file:edu.umd.cloud9.pagerank.RunPageRankSchimmy.java

License:Apache License

private float phase1(String path, int i, int j, int n, boolean useCombiner, boolean useInmapCombiner,
        boolean useRange) throws IOException {
    JobConf conf = new JobConf(RunPageRankBasic.class);

    String in = path + "/iter" + sFormat.format(i);
    String out = path + "/iter" + sFormat.format(j) + "t";
    String outm = out + "-mass";

    FileSystem fs = FileSystem.get(conf);

    // we need to actually count the number of part files to get the number
    // of partitions (because the directory might contain _log)
    int numPartitions = 0;
    for (FileStatus s : FileSystem.get(conf).listStatus(new Path(in))) {
        if (s.getPath().getName().contains("part-"))
            numPartitions++;//from   w w  w.ja  va2s  . co  m
    }

    conf.setInt("NodeCount", n);

    Partitioner p = null;

    if (useRange) {
        p = new RangePartitioner<IntWritable, Writable>();
        p.configure(conf);
    } else {
        p = new HashPartitioner<WritableComparable, Writable>();
    }

    // this is really annoying: the mapping between the partition numbers on
    // disk (i.e., part-XXXX) and what partition the file contains (i.e.,
    // key.hash % #reducer) is arbitrary... so this means that we need to
    // open up each partition, peek inside to find out.
    IntWritable key = new IntWritable();
    PageRankNode value = new PageRankNode();
    FileStatus[] status = fs.listStatus(new Path(in));

    StringBuilder sb = new StringBuilder();

    for (FileStatus f : status) {
        if (f.getPath().getName().contains("_logs"))
            continue;

        SequenceFile.Reader reader = new SequenceFile.Reader(fs, f.getPath(), conf);

        reader.next(key, value);
        int np = p.getPartition(key, value, numPartitions);
        reader.close();

        sLogger.info(f.getPath() + "\t" + np);
        sb.append(np + "=" + f.getPath() + "\t");
    }

    sLogger.info(sb.toString().trim());

    sLogger.info("PageRankSchimmy: iteration " + j + ": Phase1");
    sLogger.info(" - input: " + in);
    sLogger.info(" - output: " + out);
    sLogger.info(" - nodeCnt: " + n);
    sLogger.info(" - useCombiner: " + useCombiner);
    sLogger.info(" - useInmapCombiner: " + useInmapCombiner);
    sLogger.info(" - numPartitions: " + numPartitions);
    sLogger.info(" - useRange: " + useRange);
    sLogger.info("computed number of partitions: " + numPartitions);

    int numMapTasks = numPartitions;
    int numReduceTasks = numPartitions;

    conf.setJobName("PageRankSchimmy:iteration" + j + ":Phase1");

    conf.setNumMapTasks(numMapTasks);
    conf.setNumReduceTasks(numReduceTasks);

    conf.setInt("mapred.min.split.size", 1024 * 1024 * 1024);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.set("PageRankMassPath", outm);
    conf.set("BasePath", in);
    conf.set("PartitionMapping", sb.toString().trim());

    FileInputFormat.setInputPaths(conf, new Path(in));
    FileOutputFormat.setOutputPath(conf, new Path(out));

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(FloatWritable.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(PageRankNode.class);

    if (useInmapCombiner) {
        conf.setMapperClass(MapWithInMapperCombiningClass.class);
    } else {
        conf.setMapperClass(MapClass.class);
    }

    if (useCombiner) {
        conf.setCombinerClass(CombineClass.class);
    }

    if (useRange) {
        conf.setPartitionerClass(RangePartitioner.class);
    }

    conf.setReducerClass(ReduceClass.class);

    conf.setSpeculativeExecution(false);

    FileSystem.get(conf).delete(new Path(out), true);
    FileSystem.get(conf).delete(new Path(outm), true);

    JobClient.runJob(conf);

    float mass = Float.NEGATIVE_INFINITY;
    for (FileStatus f : fs.listStatus(new Path(outm))) {
        FSDataInputStream fin = fs.open(f.getPath());
        mass = sumLogProbs(mass, fin.readFloat());
        fin.close();
    }

    return mass;
}

From source file:edu.umd.cloud9.webgraph.BuildReverseWebGraph.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildReverseWebGraph.class);
    FileSystem fs = FileSystem.get(conf);

    int numMappers = conf.getInt("Cloud9.Mappers", 1);
    int numReducers = conf.getInt("Cloud9.Reducers", 200);

    String inputPath = conf.get("Cloud9.InputPath");
    String outputPath = conf.get("Cloud9.OutputPath");

    conf.setJobName("ReverseWebGraph");
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setInt("mapred.task.timeout", 60000000);
    conf.set("mapreduce.map.memory.mb", "2048");
    conf.set("mapreduce.map.java.opts", "-Xmx2048m");
    conf.set("mapreduce.reduce.memory.mb", "2048");
    conf.set("mapreduce.reduce.java.opts", "-Xmx2048m");
    conf.set("mapreduce.task.timeout", "60000000");

    conf.setNumMapTasks(numMappers);/*from w ww .  j  a v a2  s. c  o  m*/
    conf.setNumReduceTasks(numReducers);
    conf.setMapperClass(IdentityMapper.class);
    conf.setReducerClass(Reduce.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(ArrayListWritable.class);
    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(ArrayListWritable.class);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);

    SequenceFileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    LOG.info("BuildReverseWebGraph");
    LOG.info(" - input path: " + inputPath);
    LOG.info(" - output path: " + outputPath);

    if (!fs.exists(new Path(outputPath))) {
        JobClient.runJob(conf);
    } else {
        LOG.info(outputPath + " already exists! Skipping this step...");
    }

    return 0;
}

From source file:edu.umd.cloud9.webgraph.BuildWebGraph.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildWebGraph.class);
    FileSystem fs = FileSystem.get(conf);

    int numMappers = conf.getInt("Cloud9.Mappers", 1);
    int numReducers = conf.getInt("Cloud9.Reducers", 200);

    String inputPath = conf.get("Cloud9.InputPath");
    String outputPath = conf.get("Cloud9.OutputPath");

    conf.setJobName("ConstructWebGraph");
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setInt("mapred.task.timeout", 60000000);
    conf.set("mapreduce.map.memory.mb", "2048");
    conf.set("mapreduce.map.java.opts", "-Xmx2048m");
    conf.set("mapreduce.reduce.memory.mb", "2048");
    conf.set("mapreduce.reduce.java.opts", "-Xmx2048m");
    conf.set("mapreduce.task.timeout", "60000000");

    conf.setNumMapTasks(numMappers);//from  w w w  . j  av a  2 s.co  m
    conf.setNumReduceTasks(numReducers);

    conf.setMapperClass(Map.class);
    conf.setReducerClass(Reduce.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(ArrayListWritable.class);

    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(ArrayListWritable.class);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);

    SequenceFileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    LOG.info("BuildWebGraph");
    LOG.info(" - input path: " + inputPath);
    LOG.info(" - output path: " + outputPath);

    if (!fs.exists(new Path(outputPath))) {
        JobClient.runJob(conf);
    } else {
        LOG.info(outputPath + " already exists! Skipping this step...");
    }

    return 0;
}

From source file:edu.umd.cloud9.webgraph.ClueExtractLinks.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), ClueExtractLinks.class);
    FileSystem fs = FileSystem.get(conf);

    int numMappers = conf.getInt("Cloud9.Mappers", 1);
    int numReducers = conf.getInt("Cloud9.Reducers", 200);

    String inputPath = conf.get("Cloud9.InputPath");
    String outputPath = conf.get("Cloud9.OutputPath");
    String mappingFile = conf.get("Cloud9.DocnoMappingFile");

    if (!fs.exists(new Path(mappingFile))) {
        throw new RuntimeException("Error: Docno mapping data file " + mappingFile + " doesn't exist!");
    }// w ww .jav  a  2s . c  o  m

    DistributedCache.addCacheFile(new URI(mappingFile), conf);

    conf.setJobName("ClueExtractLinks");
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setInt("mapred.task.timeout", 60000000);
    conf.set("mapreduce.map.memory.mb", "2048");
    conf.set("mapreduce.map.java.opts", "-Xmx2048m");
    conf.set("mapreduce.reduce.memory.mb", "2048");
    conf.set("mapreduce.reduce.java.opts", "-Xmx2048m");
    conf.set("mapreduce.task.timeout", "60000000");

    conf.setNumMapTasks(numMappers);
    conf.setNumReduceTasks(numReducers);
    // TODO: to read!!
    conf.setMapperClass(Map.class);
    conf.setCombinerClass(Reduce.class);
    conf.setReducerClass(Reduce.class);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(ArrayListWritable.class);
    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);
    SequenceFileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    LOG.info("ClueExtractLinks");
    LOG.info(" - input path: " + inputPath);
    LOG.info(" - output path: " + outputPath);
    LOG.info(" - mapping file: " + mappingFile);
    LOG.info(" - include internal links? " + conf.getBoolean("Cloud9.IncludeInternalLinks", false));

    if (!fs.exists(new Path(outputPath))) {
        JobClient.runJob(conf);
    } else {
        LOG.info(outputPath + " already exists! Skipping this step...");
    }
    return 0;
}

From source file:edu.umd.cloud9.webgraph.CollectHostnames.java

License:Apache License

public int runTool() throws Exception {

    JobConf conf = new JobConf(getConf(), CollectHostnames.class);
    FileSystem fs = FileSystem.get(conf);

    int numMappers = conf.getInt("Cloud9.Mappers", 1);
    int numReducers = conf.getInt("Cloud9.Reducers", 200);

    String inputPath = conf.get("Cloud9.InputPath");
    String outputPath = conf.get("Cloud9.OutputPath");

    conf.setJobName("CollectHostnames");
    conf.set("mapred.child.java.opts", "-Xmx4096m");
    conf.setInt("mapred.task.timeout", 60000000);

    conf.setNumMapTasks(numMappers);/*from  w w  w .  ja  va2s  .  c  o m*/
    conf.setNumReduceTasks(numReducers);

    conf.setMapperClass(Map.class);
    conf.setPartitionerClass(Partition.class);
    conf.setReducerClass(Reduce.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(ArrayListWritable.class);

    conf.setMapOutputKeyClass(PairOfIntString.class);
    conf.setMapOutputValueClass(IntWritable.class);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);

    SequenceFileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    sLogger.info("PropagateHostname");
    sLogger.info(" - input path: " + inputPath);
    sLogger.info(" - output path: " + outputPath);

    if (!fs.exists(new Path(outputPath))) {
        JobClient.runJob(conf);
    } else {
        sLogger.info(outputPath + " already exists! Skipping this step...");
    }

    return 0;
}

From source file:edu.umd.cloud9.webgraph.ComputeWeight.java

License:Apache License

public int runTool() throws Exception {

    JobConf conf = new JobConf(getConf(), ComputeWeight.class);
    FileSystem fs = FileSystem.get(conf);

    int numMappers = conf.getInt("Cloud9.Mappers", 1);
    int numReducers = conf.getInt("Cloud9.Reducers", 200);

    String inputPath = conf.get("Cloud9.InputPath");
    String outputPath = conf.get("Cloud9.OutputPath");

    conf.setJobName("ComputeWeights");
    conf.set("mapred.child.java.opts", "-Xmx4096m");
    conf.setInt("mapred.task.timeout", 60000000);

    conf.setNumMapTasks(numMappers);/*  w ww .  j  a  v  a  2 s .c om*/
    conf.setNumReduceTasks(numReducers);

    conf.setMapperClass(Map.class);
    conf.setPartitionerClass(Partition.class);
    conf.setReducerClass(Reduce.class);

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(ArrayListWritable.class);

    conf.setMapOutputKeyClass(PairOfInts.class);
    conf.setMapOutputValueClass(ArrayListWritable.class);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);

    SequenceFileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    LOG.info("ComputeWeight");
    LOG.info(" - input path: " + inputPath);
    LOG.info(" - output path: " + outputPath);

    if (!fs.exists(new Path(outputPath))) {
        JobClient.runJob(conf);
    } else {
        LOG.info(outputPath + " already exists! Skipping this step...");
    }

    return 0;
}

From source file:edu.umd.cloud9.webgraph.driver.BuildAnchorTextForwardIndex.java

License:Apache License

/**
 * Runs this tool./*from www .jav a2  s.  c om*/
 */
public int run(String[] args) throws Exception {
    if (args.length != 3) {
        printUsage();
        return -1;
    }

    JobConf conf = new JobConf(getConf());
    FileSystem fs = FileSystem.get(conf);

    String collectionPath = args[0];
    String outputPath = args[1];
    String indexFile = args[2];

    LOG.info("Tool name: BuildAnchorTextForwardIndex");
    LOG.info(" - collection path: " + collectionPath);
    LOG.info(" - output path: " + outputPath);
    LOG.info(" - index file: " + indexFile);
    LOG.info("Note: This tool only works on block-compressed SequenceFiles!");

    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setJobName("BuildAnchorTextForwardIndex");

    conf.setNumMapTasks(100);
    conf.setNumReduceTasks(1);

    FileInputFormat.setInputPaths(conf, new Path(collectionPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));
    FileOutputFormat.setCompressOutput(conf, false);

    conf.setInputFormat(NoSplitSequenceFileInputFormat.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapRunnerClass(MyMapRunner.class);
    conf.setReducerClass(IdentityReducer.class);

    // delete the output directory if it exists already
    fs.delete(new Path(outputPath), true);

    RunningJob job = JobClient.runJob(conf);

    Counters counters = job.getCounters();
    int blocks = (int) counters.findCounter(Blocks.Total).getCounter();

    LOG.info("number of blocks: " + blocks);

    LOG.info("Writing index file...");
    LineReader reader = new LineReader(fs.open(new Path(outputPath + "/part-00000")));
    FSDataOutputStream out = fs.create(new Path(indexFile), true);

    out.writeUTF(IndexableAnchorTextForwardIndex.class.getName());
    out.writeUTF(collectionPath);
    out.writeInt(blocks);

    int cnt = 0;
    Text line = new Text();
    while (reader.readLine(line) > 0) {
        String[] arr = line.toString().split("\\s+");

        int docno = Integer.parseInt(arr[0]);
        int offset = Integer.parseInt(arr[1]);
        short fileno = Short.parseShort(arr[2]);

        out.writeInt(docno);
        out.writeInt(offset);
        out.writeShort(fileno);

        cnt++;

        if (cnt % 1000 == 0) {
            LOG.info(cnt + " blocks written");
        }

    }

    reader.close();
    out.close();

    if (cnt != blocks) {
        throw new RuntimeException("Error: mismatch in block count!");
    }

    return 0;
}

From source file:edu.umd.cloud9.webgraph.driver.BuildIndexableAnchorCollection.java

License:Apache License

/**
 * Runs this tool./*from  w  ww .j  av  a 2s.com*/
 */
public int run(String[] args) throws Exception {
    if (args.length < 5) {
        printUsage();
        return -1;
    }

    JobConf conf = new JobConf(getConf());
    FileSystem fs = FileSystem.get(conf);

    String collectionPath = DriverUtil.argValue(args, DriverUtil.CL_INPUT);
    String outputPath = DriverUtil.argValue(args, DriverUtil.CL_OUTPUT);
    String docnoMappingClass = DriverUtil.argValue(args, DriverUtil.CL_DOCNO_MAPPING_CLASS);
    String docnoMapping = DriverUtil.argValue(args, DriverUtil.CL_DOCNO_MAPPING);
    int numReducers = Integer.parseInt(DriverUtil.argValue(args, DriverUtil.CL_NUMBER_OF_REDUCERS));
    if (DriverUtil.argExists(args, DriverUtil.CL_MAX_LENGTH)) {
        conf.setInt("Cloud9.maxContentLength",
                Integer.parseInt(DriverUtil.argValue(args, DriverUtil.CL_MAX_LENGTH)));
    }
    conf.set("Cloud9.DocnoMappingClass", docnoMappingClass);

    LOG.info("Tool name: BuildAnchorTextForwardIndex");
    LOG.info(" - collection path: " + collectionPath);
    LOG.info(" - output path: " + outputPath);
    LOG.info(" - docno-mapping class: " + docnoMappingClass);
    LOG.info(" - docno-mapping file: " + docnoMapping);
    if (args.length == 6) {
        LOG.info(" - maximum content length: " + conf.getInt("Cloud9.maxContentLength", 0));
    }

    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setJobName("BuildIndexableAnchorCollection");
    conf.setJarByClass(BuildIndexableAnchorCollection.class);

    conf.setNumMapTasks(100);
    conf.setNumReduceTasks(numReducers);
    DistributedCache.addCacheFile(new URI(docnoMapping), conf);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    SequenceFileOutputFormat.setCompressOutput(conf, true);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.BLOCK);
    SequenceFileInputFormat.setInputPaths(conf, new Path(collectionPath));
    SequenceFileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(IndexableAnchorText.class);

    conf.setMapperClass(MyMapper.class);
    conf.setReducerClass(IdentityReducer.class);

    // delete the output directory if it exists already
    fs.delete(new Path(outputPath), true);
    RunningJob job = JobClient.runJob(conf);

    return 0;
}