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

Introduction

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

Prototype

public void setMapOutputKeyClass(Class<?> theClass)

Source Link

Document

Set the key class for the map output data.

Usage

From source file:at.illecker.hadoop.rootbeer.examples.matrixmultiplication.gpu.MatrixMultiplicationGpu.java

License:Apache License

public static Configuration createMatrixMultiplicationGpuConf(Configuration initialConf, Path aPath, Path bPath,
        Path outPath, int outCardinality, int tileWidth, boolean isDebugging) {

    JobConf conf = new JobConf(initialConf, MatrixMultiplicationGpu.class);
    conf.setJobName("MatrixMultiplicationGPU: " + aPath + " x " + bPath + " = " + outPath);

    conf.setInt(CONF_OUT_CARD, outCardinality);
    conf.setInt(CONF_TILE_WIDTH, tileWidth);
    conf.setBoolean(CONF_DEBUG, isDebugging);

    conf.setInputFormat(CompositeInputFormat.class);
    conf.set("mapred.join.expr",
            CompositeInputFormat.compose("inner", SequenceFileInputFormat.class, aPath, bPath));

    conf.setOutputFormat(SequenceFileOutputFormat.class);
    FileOutputFormat.setOutputPath(conf, outPath);

    conf.setMapperClass(MatrixMultiplyGpuMapper.class);

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

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

    // Increase client heap size for GPU Rootbeer execution
    conf.set("mapred.child.java.opts", "-Xms8G -Xmx8G");

    // No Reduce step is needed
    // -> 0 reducer means reduce step will be skipped and
    // mapper output will be the final out
    // -> Identity reducer means then shuffling/sorting will still take place
    conf.setNumReduceTasks(0);/*from   ww  w.jav  a2  s .  c o m*/

    return conf;
}

From source file:Brush.AdjustMateEdge.java

License:Apache License

public RunningJob run(String inputPath, String outputPath, long reads, long ctg_sum) throws Exception {
    sLogger.info("Tool name: AdjustMateEdge");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    //JobConf conf = new JobConf(Stats.class);
    JobConf conf = new JobConf(AdjustMateEdge.class);
    conf.setJobName("AdjustMateEdge " + inputPath);

    conf.setLong("READS", reads);
    conf.setLong("CTG_SUM", ctg_sum);
    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(AdjustMateEdgeMapper.class);
    conf.setReducerClass(AdjustMateEdgeReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.Compressible.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: Compressible");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    //JobConf conf = new JobConf(Stats.class);
    JobConf conf = new JobConf(Compressible.class);
    conf.setJobName("Compressible " + inputPath);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(CompressibleMapper.class);
    conf.setReducerClass(CompressibleReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.CountBraid.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: CountBraid");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(CountBraid.class);
    conf.setJobName("CountBraid " + inputPath + " " + BrushConfig.K);
    //conf.setFloat("Error_Rate", ErrorRate);
    //conf.setFloat("Exp_Cov", Exp_Cov);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(CountBraidMapper.class);
    conf.setReducerClass(CountBraidReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.CountKmer.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: CountKmer");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(CountKmer.class);
    conf.setJobName("CountKmer " + inputPath + " " + BrushConfig.K);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    //conf.setMapOutputValueClass(IntWritable.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    //conf.setBoolean("mapred.output.compress", true);

    conf.setMapperClass(CountKmerMapper.class);
    conf.setReducerClass(CountKmerReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.CountReads.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: CountReads");
    sLogger.info(" - input: " + inputPath);

    JobConf conf = new JobConf(CountReads.class);
    conf.setJobName("CountReads " + inputPath + " " + BrushConfig.K);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(CountReadsMapper.class);
    //conf.setReducerClass(CountReadsReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.CutChimericLinks.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: CutChimericLinks");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(CutChimericLinks.class);
    conf.setJobName("CutChimericLinks " + inputPath + " " + BrushConfig.K);
    //conf.setFloat("Error_Rate", ErrorRate);
    //conf.setFloat("Exp_Cov", Exp_Cov);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(CutChimericLinksMapper.class);
    conf.setReducerClass(CutChimericLinksReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.CutRepeatBoundary.java

License:Apache License

public RunningJob run(String inputPath, String outputPath, long reads, long ctg_sum) throws Exception {
    sLogger.info("Tool name: CutRepeatBoundary");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(CutRepeatBoundary.class);
    conf.setJobName("CutRepeatBoundary " + inputPath + " " + BrushConfig.K);
    //conf.setFloat("Error_Rate", ErrorRate);
    //conf.setFloat("Exp_Cov", Exp_Cov);
    conf.setLong("READS", reads);
    conf.setLong("CTG_SUM", ctg_sum);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(CutRepeatBoundaryMapper.class);
    conf.setReducerClass(CutRepeatBoundaryReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.DefineConsensus.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: DefineConsensus");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(DefineConsensus.class);
    conf.setJobName("DefineConsensus " + inputPath + " " + BrushConfig.K);
    //conf.setFloat("Error_Rate", ErrorRate);
    //conf.setFloat("Exp_Cov", Exp_Cov);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(DefineConsensusMapper.class);
    conf.setReducerClass(DefineConsensusReducer.class);

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

    return JobClient.runJob(conf);
}

From source file:Brush.EdgeRemoval.java

License:Apache License

public RunningJob run(String inputPath, String outputPath) throws Exception {
    sLogger.info("Tool name: EdgeRemoval");
    sLogger.info(" - input: " + inputPath);
    sLogger.info(" - output: " + outputPath);

    JobConf conf = new JobConf(EdgeRemoval.class);
    conf.setJobName("EdgeRemoval " + inputPath + " " + BrushConfig.K);

    BrushConfig.initializeConfiguration(conf);

    FileInputFormat.addInputPath(conf, new Path(inputPath));
    FileOutputFormat.setOutputPath(conf, new Path(outputPath));

    conf.setInputFormat(TextInputFormat.class);
    conf.setOutputFormat(TextOutputFormat.class);

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(Text.class);
    //conf.setBoolean("mapred.output.compress", true);

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);

    conf.setMapperClass(EdgeRemovalMapper.class);
    conf.setReducerClass(EdgeRemovalReducer.class);

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

    return JobClient.runJob(conf);
}