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

Introduction

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

Prototype

public void setNumMapTasks(int n) 

Source Link

Document

Set the number of map tasks for this job.

Usage

From source file:net.sf.katta.indexing.IndexerJob.java

License:Apache License

public void startIndexer(String path, String finalDestination, int numOfShards) throws IOException {
    // create job conf with class pointing into job jar.
    JobConf jobConf = new JobConf(IndexerJob.class);
    jobConf.setJobName("indexer");
    jobConf.setMapRunnerClass(Indexer.class);
    // alternative use a text file and a TextInputFormat
    jobConf.setInputFormat(SequenceFileInputFormat.class);

    Path input = new Path(path);
    FileInputFormat.setInputPaths(jobConf, input);
    // we just set the output path to make hadoop happy.
    FileOutputFormat.setOutputPath(jobConf, new Path(finalDestination));
    // setting the folder where lucene indexes will be copied when finished.
    jobConf.set("finalDestination", finalDestination);
    // important to switch spec exec off.
    // We dont want to have something duplicated.
    jobConf.setSpeculativeExecution(false);

    // The num of map tasks is equal to the num of input splits.
    // The num of input splits by default is equal to the num of hdf blocks
    // for the input file(s). To get the right num of shards we need to
    // calculate the best input split size.

    FileSystem fs = FileSystem.get(input.toUri(), jobConf);
    FileStatus[] status = fs.globStatus(input);
    long size = 0;
    for (FileStatus fileStatus : status) {
        size += fileStatus.getLen();//from w  w w  .j  a v  a 2 s  .c  om
    }
    long optimalSplisize = size / numOfShards;
    jobConf.set("mapred.min.split.size", "" + optimalSplisize);

    // give more mem to lucene tasks.
    jobConf.set("mapred.child.java.opts", "-Xmx2G");
    jobConf.setNumMapTasks(1);
    jobConf.setNumReduceTasks(0);
    JobClient.runJob(jobConf);
}

---

From source file:nl.tudelft.graphalytics.mapreducev2.MapReduceJob.java

License:Apache License

@Override
public int run(String[] args) throws Exception {
    FileSystem dfs = FileSystem.get(getConf());
    String inPath = inputPath;/*w  w  w  . j a  v  a 2s.  c om*/

    while (!isFinished()) {
        iteration++;

        // Prepare job configuration
        JobConf jobConfiguration = new JobConf(this.getConf());
        jobConfiguration.setJarByClass(this.getClass());

        jobConfiguration.setMapOutputKeyClass(getMapOutputKeyClass());
        jobConfiguration.setMapOutputValueClass(getMapOutputValueClass());

        jobConfiguration.setMapperClass(getMapperClass());
        if (getCombinerClass() != null)
            jobConfiguration.setCombinerClass(getCombinerClass());
        jobConfiguration.setReducerClass(getReducerClass());

        jobConfiguration.setOutputKeyClass(getOutputKeyClass());
        jobConfiguration.setOutputValueClass(getOutputValueClass());

        jobConfiguration.setInputFormat(getInputFormatClass());
        jobConfiguration.setOutputFormat(getOutputFormatClass());

        if (getNumMappers() != -1)
            jobConfiguration.setNumMapTasks(getNumMappers());
        if (getNumReducers() != -1)
            jobConfiguration.setNumReduceTasks(getNumReducers());

        setConfigurationParameters(jobConfiguration);

        // Set the input and output paths
        String outPath = intermediatePath + "/iteration-" + iteration;
        FileInputFormat.addInputPath(jobConfiguration, new Path(inPath));
        FileOutputFormat.setOutputPath(jobConfiguration, new Path(outPath));

        // Execute the current iteration
        RunningJob jobExecution = JobClient.runJob(jobConfiguration);
        jobExecution.waitForCompletion();

        // Remove the output of the previous job (unless it is the input graph)
        if (iteration != 1) {
            dfs.delete(new Path(inPath), true);
        }
        inPath = outPath;

        processJobOutput(jobExecution);
    }

    // Rename the last job output to the specified output path
    try {
        dfs.mkdirs(new Path(outputPath).getParent());
        dfs.rename(new Path(inPath), new Path(outputPath));
    } catch (Exception e) {
        LOG.warn("Failed to rename MapReduce job output.", e);
    }

    return 0;
}

---

From source file:nlp.com.knowledgebooks.mapreduce.NameFinder.java

License:Open Source License

/**
 * The main driver for name finder map/reduce program.
 * <p/>//from   w w w  .  ja  v  a 2  s .  c om
 * NOTE: copied with modifications from Hadoppjava example programs
 * <p/>
 * Invoke this method to submit the map/reduce job.
 *
 * @throws IOException When there is communication problems with the
 *                     job tracker.
 */
public int run(String[] args) throws Exception {
    JobConf conf = new JobConf(getConf(), NameFinder.class);
    conf.setJobName("namefinder");
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);
    conf.setMapperClass(MapClass.class);
    //conf.setCombinerClass(Reduce.class);
    conf.setReducerClass(Reduce.class);
    List<String> other_args = new ArrayList<String>();
    for (int i = 0; i < args.length; ++i) {
        try {
            if ("-m".equals(args[i])) {
                conf.setNumMapTasks(Integer.parseInt(args[++i]));
            } else if ("-r".equals(args[i])) {
                conf.setNumReduceTasks(Integer.parseInt(args[++i]));
            } else {
                other_args.add(args[i]);
            }
        } catch (Exception ex) {
            System.err.println("ERROR: " + ex);
        }
    }
    FileInputFormat.setInputPaths(conf, other_args.get(0));
    FileOutputFormat.setOutputPath(conf, new Path(other_args.get(1)));
    JobClient.runJob(conf);
    return 0;
}

---

From source file:nthu.scopelab.tsqr.ssvd.ABtDenseOutJob.java

License:Apache License

public static void run(Configuration conf, Path[] inputPath, Path inputBt, Path outputPath, int k, int p,
        int reduceTasks, int mis) throws Exception {

    JobConf job = new JobConf(conf, ABtDenseOutJob.class);
    job.setInputFormat(SequenceFileInputFormat.class);
    job.setOutputFormat(SequenceFileOutputFormat.class);

    job.setInt(QJob.PROP_K, k);/* w w w  .  j  av  a  2 s.  c o m*/
    job.setInt(QJob.PROP_P, p);
    job.set(PROP_BT_PATH, inputBt.toString());

    FileOutputFormat.setOutputPath(job, outputPath);
    job.setJobName("ABtDenseOutJob");

    job.setMapOutputKeyClass(IntWritable.class);
    job.setMapOutputValueClass(LMatrixWritable.class);
    job.setOutputKeyClass(IntWritable.class);
    job.setOutputValueClass(LMatrixWritable.class);

    job.setMapperClass(ABtMapper.class);

    fileGather fgather = new fileGather(inputPath, "", FileSystem.get(job));
    mis = Checker.checkMis(mis, fgather.getInputSize(), FileSystem.get(job));
    job.setNumMapTasks(fgather.recNumMapTasks(mis));

    job.setNumReduceTasks(0);

    FileInputFormat.setInputPaths(job, inputPath);

    RunningJob rj = JobClient.runJob(job);
}

---

From source file:nthu.scopelab.tsqr.ssvd.BtJob.java

License:Apache License

public static void run(Configuration conf, Path[] inputPath, Path btPath, String qrfPath, int k, int p,
        int outerBlockHeight, int reduceTasks, boolean outputBBtProducts, String reduceSchedule, int mis)
        throws Exception {
    boolean outputQ = true;

    String stages[] = reduceSchedule.split(",");

    JobConf job = new JobConf(conf, BtJob.class);
    job.setInputFormat(SequenceFileInputFormat.class);
    job.setOutputFormat(SequenceFileOutputFormat.class);
    job.setInt(SCHEDULE_NUM, stages.length);
    job.setInt(PROP_OUTER_PROD_BLOCK_HEIGHT, outerBlockHeight);
    job.setInt(QJob.PROP_K, k);//w ww . j  av  a2 s. c o  m
    job.setInt(QJob.PROP_P, p);
    job.setBoolean(QmultiplyJob.OUTPUT_Q, outputQ);
    job.setBoolean(PROP_OUPTUT_BBT_PRODUCTS, outputBBtProducts);
    job.set(QmultiplyJob.QRF_DIR, qrfPath);
    FileSystem.get(job).delete(btPath, true);

    FileOutputFormat.setOutputPath(job, btPath);

    FileOutputFormat.setCompressOutput(job, true);
    FileOutputFormat.setOutputCompressorClass(job, DefaultCodec.class);
    SequenceFileOutputFormat.setOutputCompressionType(job, CompressionType.BLOCK);

    job.setJobName("BtJob");

    job.setMapOutputKeyClass(IntWritable.class);
    job.setMapOutputValueClass(SparseRowBlockWritable.class);
    job.setOutputKeyClass(IntWritable.class);
    //job.setOutputValueClass(SparseRowBlockWritable.class);
    job.setOutputValueClass(VectorWritable.class);

    job.setMapperClass(BtMapper.class);
    job.setCombinerClass(OuterProductCombiner.class);
    job.setReducerClass(OuterProductReducer.class);

    fileGather fgather = new fileGather(inputPath, "", FileSystem.get(job));
    mis = Checker.checkMis(mis, fgather.getInputSize(), FileSystem.get(job));
    job.setNumMapTasks(fgather.recNumMapTasks(mis));

    //job.setNumReduceTasks(0);
    job.setNumReduceTasks(reduceTasks);

    FileInputFormat.setInputPaths(job, inputPath);

    if (outputQ) {
        MultipleOutputs.addNamedOutput(job, QmultiplyJob.Q_MAT, SequenceFileOutputFormat.class,
                IntWritable.class, LMatrixWritable.class);
    }
    if (outputBBtProducts) {
        MultipleOutputs.addNamedOutput(job, OUTPUT_BBT, SequenceFileOutputFormat.class, IntWritable.class,
                VectorWritable.class);
    }
    RunningJob rj = JobClient.runJob(job);
    System.out.println("Btjob Job ID: " + rj.getJobID().toString());
}

---

From source file:nthu.scopelab.tsqr.ssvd.itBtJob.java

License:Apache License

public static void run(Configuration conf, Path[] inputPath, Path btPath, String qrfPath, int k, int p,
        int outerBlockHeight, int reduceTasks, boolean outputBBtProducts, String reduceSchedule, int mis)
        throws Exception {
    boolean outputQ = true;

    String stages[] = reduceSchedule.split(",");

    JobConf job = new JobConf(conf, itBtJob.class);
    job.setInputFormat(SequenceFileInputFormat.class);
    job.setOutputFormat(SequenceFileOutputFormat.class);
    job.setInt(SCHEDULE_NUM, stages.length);
    job.setInt(PROP_OUTER_PROD_BLOCK_HEIGHT, outerBlockHeight);
    job.setInt(QJob.PROP_K, k);/*from   www .j a va  2  s . com*/
    job.setInt(QJob.PROP_P, p);
    job.setBoolean(QmultiplyJob.OUTPUT_Q, outputQ);
    job.setBoolean(PROP_OUPTUT_BBT_PRODUCTS, outputBBtProducts);
    job.set(QmultiplyJob.QRF_DIR, qrfPath);
    FileSystem.get(job).delete(btPath, true);

    FileOutputFormat.setOutputPath(job, btPath);

    FileOutputFormat.setCompressOutput(job, true);
    FileOutputFormat.setOutputCompressorClass(job, DefaultCodec.class);
    SequenceFileOutputFormat.setOutputCompressionType(job, CompressionType.BLOCK);

    job.setJobName("itBtJob");

    job.setMapOutputKeyClass(IntWritable.class);
    job.setMapOutputValueClass(SparseRowBlockWritable.class);
    job.setOutputKeyClass(IntWritable.class);
    //job.setOutputValueClass(SparseRowBlockWritable.class);
    job.setOutputValueClass(VectorWritable.class);

    job.setMapperClass(BtMapper.class);
    job.setCombinerClass(OuterProductCombiner.class);
    job.setReducerClass(OuterProductReducer.class);

    fileGather fgather = new fileGather(inputPath, "", FileSystem.get(job));
    mis = Checker.checkMis(mis, fgather.getInputSize(), FileSystem.get(job));
    job.setNumMapTasks(fgather.recNumMapTasks(mis));

    //job.setNumReduceTasks(0);
    job.setNumReduceTasks(reduceTasks);

    FileInputFormat.setInputPaths(job, inputPath);

    if (outputQ) {
        MultipleOutputs.addNamedOutput(job, QmultiplyJob.Q_MAT, SequenceFileOutputFormat.class,
                IntWritable.class, LMatrixWritable.class);
    }
    if (outputBBtProducts) {
        MultipleOutputs.addNamedOutput(job, OUTPUT_BBT, SequenceFileOutputFormat.class, IntWritable.class,
                VectorWritable.class);
    }
    RunningJob rj = JobClient.runJob(job);
    System.out.println("itBtJob Job ID: " + rj.getJobID().toString());
}

---

From source file:nthu.scopelab.tsqr.ssvd.itQJob.java

License:Apache License

public static void run(Configuration conf, Path[] inputPaths, String outputPath, String reduceSchedule, int k,
        int p, long seed, int mis) throws ClassNotFoundException, InterruptedException, IOException {

    String stages[] = reduceSchedule.split(",");
    String rinput = "";
    String routput = outputPath + "/iter-r-";

    for (int i = 0; i < stages.length; i++) {
        String thenumber = Integer.toString(i + 1);
        JobConf job = new JobConf(conf, itQJob.class);
        job.setJobName("itQ-job-" + thenumber);
        job.setInputFormat(SequenceFileInputFormat.class);
        job.setOutputFormat(SequenceFileOutputFormat.class);

        if (i == 0)
            job.setMapperClass(QMapper.class);
        else//from  www  . j  a v  a2 s . c  om
            job.setMapperClass(IdentityMapper.class);

        job.setReducerClass(QReducer.class);
        job.setOutputKeyClass(IntWritable.class);
        job.setOutputValueClass(LMatrixWritable.class);

        FileSystem fs = FileSystem.get(job);
        Path Paths[];
        fileGather fgather = null;
        if (i == 0)
            fgather = new fileGather(inputPaths, "part", fs);
        else
            fgather = new fileGather(new Path(rinput), "part", fs);
        Paths = fgather.getPaths();
        mis = Checker.checkMis(mis, fgather.getInputSize(), fs);
        job.setNumMapTasks(fgather.recNumMapTasks(mis));

        job.setNumReduceTasks(Integer.parseInt(stages[i]));

        job.setInt(QRFirstJob.COLUMN_SIZE, k + p);
        job.setLong(PROP_OMEGA_SEED, seed);
        job.setInt(PROP_K, k);
        job.setInt(PROP_P, p);

        fs.delete(new Path(routput + thenumber), true);

        FileInputFormat.setInputPaths(job, Paths);

        FileOutputFormat.setOutputPath(job, new Path(routput + thenumber));

        //FileOutputFormat.setCompressOutput(job, true);
        //FileOutputFormat.setOutputCompressorClass(job, DefaultCodec.class);
        //SequenceFileOutputFormat.setOutputCompressionType(job,CompressionType.BLOCK);
        //output first level Q
        MultipleOutputs.addNamedOutput(job, QF_MAT, SequenceFileOutputFormat.class, IntWritable.class,
                LMatrixWritable.class);

        RunningJob rj = JobClient.runJob(job);
        System.out.println("itQJob Job ID: " + rj.getJobID().toString());
        rinput = routput + thenumber;
    }
}

---

From source file:nthu.scopelab.tsqr.ssvd.QJob.java

License:Apache License

public static void run(Configuration conf, Path[] inputPaths, String outputPath, String reduceSchedule, int k,
        int p, long seed, int mis) throws ClassNotFoundException, InterruptedException, IOException {

    String stages[] = reduceSchedule.split(",");
    String rinput = "";
    String routput = outputPath + "/iter-r-";

    for (int i = 0; i < stages.length; i++) {
        String thenumber = Integer.toString(i + 1);
        JobConf job = new JobConf(conf, QJob.class);
        job.setJobName("Q-job-" + thenumber);
        job.setInputFormat(SequenceFileInputFormat.class);
        job.setOutputFormat(SequenceFileOutputFormat.class);

        if (i == 0)
            job.setMapperClass(QMapper.class);
        else/*from  w w w .ja va  2 s  .c  om*/
            job.setMapperClass(IdentityMapper.class);

        job.setReducerClass(QReducer.class);
        job.setOutputKeyClass(IntWritable.class);
        job.setOutputValueClass(LMatrixWritable.class);

        FileSystem fs = FileSystem.get(job);
        Path Paths[];
        fileGather fgather = null;
        if (i == 0)
            fgather = new fileGather(inputPaths, "part", fs);
        else
            fgather = new fileGather(new Path(rinput), "part", fs);
        Paths = fgather.getPaths();
        mis = Checker.checkMis(mis, fgather.getInputSize(), fs);
        job.setNumMapTasks(fgather.recNumMapTasks(mis));

        job.setNumReduceTasks(Integer.parseInt(stages[i]));

        job.setInt(QRFirstJob.COLUMN_SIZE, k + p);
        job.setLong(PROP_OMEGA_SEED, seed);
        job.setInt(PROP_K, k);
        job.setInt(PROP_P, p);

        fs.delete(new Path(routput + thenumber), true);

        FileInputFormat.setInputPaths(job, Paths);

        FileOutputFormat.setOutputPath(job, new Path(routput + thenumber));

        //FileOutputFormat.setCompressOutput(job, true);
        //FileOutputFormat.setOutputCompressorClass(job, DefaultCodec.class);
        //SequenceFileOutputFormat.setOutputCompressionType(job,CompressionType.BLOCK);
        //output first level Q
        MultipleOutputs.addNamedOutput(job, QF_MAT, SequenceFileOutputFormat.class, IntWritable.class,
                LMatrixWritable.class);

        RunningJob rj = JobClient.runJob(job);
        System.out.println("QJob Job ID: " + rj.getJobID().toString());
        rinput = routput + thenumber;
    }
}

---

From source file:nthu.scopelab.tsqr.ssvd.UJob.java

License:Apache License

public void start(Configuration conf, Path inputPathQ, Path inputUHatPath, Path sigmaPath, Path outputPath,
        int k, boolean uHalfSigma, int mis) throws ClassNotFoundException, InterruptedException, IOException {
    String input = "";

    JobConf job = new JobConf(conf, UJob.class);
    jobclient = new JobClient(job);
    job.setJobName("UJob");
    job.setInputFormat(SequenceFileInputFormat.class);
    job.setOutputFormat(SequenceFileOutputFormat.class);

    job.setMapperClass(MultiplyMapper.class);

    job.setOutputKeyClass(LongWritable.class);
    job.setOutputValueClass(LMatrixWritable.class);
    job.setMapOutputKeyClass(LongWritable.class);
    job.setMapOutputValueClass(LMatrixWritable.class);

    FileSystem fs = FileSystem.get(job);
    fileGather fgather = new fileGather(
            new Path(inputPathQ.toString().substring(0, inputPathQ.toString().lastIndexOf("/") - 1)), "Q-", fs);
    mis = Checker.checkMis(mis, fgather.getInputSize(), fs);
    job.setNumMapTasks(fgather.recNumMapTasks(mis));

    job.setNumReduceTasks(0);/*from  w w  w .  ja  v  a2 s  . co m*/
    job.set("mapreduce.output.basename", OUTPUT_U);
    job.set(PROP_UHAT_PATH, inputUHatPath.toString());
    job.set(PROP_SIGMA_PATH, sigmaPath.toString());
    if (uHalfSigma) {
        job.set(PROP_U_HALFSIGMA, "y");
    }
    job.setInt(QJob.PROP_K, k);
    FileSystem.get(job).delete(outputPath, true);
    FileOutputFormat.setOutputPath(job, outputPath);
    FileOutputFormat.setCompressOutput(job, true);
    FileOutputFormat.setOutputCompressorClass(job, DefaultCodec.class);
    SequenceFileOutputFormat.setOutputCompressionType(job, CompressionType.BLOCK);

    FileInputFormat.setInputPaths(job, inputPathQ);
    //JobClient.runJob(job);
    jobid = jobclient.submitJob(job).getID();

}

---

From source file:org.acacia.csr.java.CSRConverter.java

License:Apache License

public static void main(String[] args) throws Exception {
    if (!validArgs(args)) {
        printUsage();//  www.j ava2s .c o  m
        return;
    }
    //These are the temp paths that are created on HDFS
    String dir1 = "/user/miyuru/csrconverter-output";
    String dir2 = "/user/miyuru/csrconverter-output-sorted";

    //We first delete the temporary directories if they exist on the HDFS
    FileSystem fs1 = FileSystem.get(new JobConf());

    System.out.println("Deleting the dir : " + dir1);

    if (fs1.exists(new Path(dir1))) {
        fs1.delete(new Path(dir1), true);
    }

    System.out.println("Done deleting the dir : " + dir1);
    System.out.println("Deleting the dir : " + dir2);
    if (fs1.exists(new Path(dir2))) {
        fs1.delete(new Path(dir2), true);
    }

    Path notinPath = new Path("/user/miyuru/notinverts/notinverts");

    if (!fs1.exists(notinPath)) {
        fs1.create(notinPath);
    }

    System.out.println("Done deleting the dir : " + dir2);

    //Note on Aug 23 2014: Sometimes after this the mapReduce job hangs. need to see why.

    VertexCounterClient.setDefaultGraphID(args[3], args[2]);

    //First job creates the inverted index

    JobConf conf = new JobConf(CSRConverter.class);
    conf.set("org.acacia.partitioner.hbase.zookeeper.quorum", args[1]);
    conf.set("org.acacia.partitioner.hbase.table", args[2]);
    conf.set("org.acacia.partitioner.hbase.contacthost", args[3]);
    conf.setOutputKeyClass(LongWritable.class);
    conf.setOutputValueClass(Text.class);
    //conf.setMapperClass(InvertedMapper.class);
    conf.setReducerClass(InvertedReducer.class);
    //conf.setInputFormat(TextInputFormat.class);
    conf.setInputFormat(NLinesInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    //FileInputFormat.setInputPaths(conf, new Path(args[0]));
    MultipleInputs.addInputPath(conf, new Path(args[0]), NLinesInputFormat.class, InvertedMapper.class);
    MultipleInputs.addInputPath(conf, new Path("/user/miyuru/notinverts/notinverts"), TextInputFormat.class,
            InvertedMapper.class);
    FileOutputFormat.setOutputPath(conf, new Path(dir1));

    //Also for the moment we turn-off the speculative execution
    conf.setBoolean("mapred.map.tasks.speculative.execution", false);
    conf.setBoolean("mapred.reduce.tasks.speculative.execution", false);
    conf.setNumMapTasks(96);
    conf.setNumReduceTasks(96);
    conf.setPartitionerClass(VertexPartitioner.class);
    conf.set("vertex-count", args[4]);
    conf.set("zero-flag", args[5]);
    Job job = new Job(conf, "csr_inverter");
    job.setSortComparatorClass(SortComparator.class);
    job.waitForCompletion(true);
}

---

• «
• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• »