Example usage for org.apache.hadoop.mapreduce Job Job

List of usage examples for org.apache.hadoop.mapreduce Job Job

  1. HOME
  2. Java
  3. org.apache.hadoop
  4. org.apache.hadoop.mapreduce.*
  5. Job
  6. Job

Introduction

In this page you can find the example usage for org.apache.hadoop.mapreduce Job Job.

Prototype

Job(JobConf conf) throws IOException

Source Link

Usage

From source file:co.cask.cdap.internal.app.runtime.batch.dataset.partitioned.DynamicPartitioningOutputFormat.java

License:Apache License

private TaskAttemptContext getTaskAttemptContext(TaskAttemptContext context, String newOutputName)
        throws IOException {
    Job job = new Job(context.getConfiguration());
    FileOutputFormat.setOutputName(job, newOutputName);
    // CDAP-4806 We must set this parameter in addition to calling FileOutputFormat#setOutputName, because
    // AvroKeyOutputFormat/AvroKeyValueOutputFormat use a different parameter for the output name than FileOutputFormat.
    if (isAvroOutputFormat(getFileOutputFormat(context))) {
        job.getConfiguration().set("avro.mo.config.namedOutput", newOutputName);
    }/*from w ww . j a va  2 s . c  o  m*/

    Path jobOutputPath = createJobSpecificPath(FileOutputFormat.getOutputPath(job), context);
    FileOutputFormat.setOutputPath(job, jobOutputPath);

    return new TaskAttemptContextImpl(job.getConfiguration(), context.getTaskAttemptID());
}

From source file:co.nubetech.hiho.dedup.DedupJob.java

License:Apache License

@Override
public int run(String[] args) throws Exception {
    Configuration conf = getConf();
    populateConfiguration(args);/*from   w w  w  .java  2 s . c  o m*/
    try {
        checkMandatoryConfs();
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new Exception(e1);
    }
    Job job = new Job(conf);
    job.setJobName("Dedup job");
    job.setJarByClass(DedupJob.class);

    Class inputFormatClass = Class.forName(inputFormat);
    Class outputFormatClass = Class.forName(outputFormat);
    Class inputKeyClass = Class.forName(inputKeyClassName);
    Class inputValueClass = Class.forName(inputValueClassName);

    if (dedupBy.equals("key")) {
        job.setMapperClass(DedupKeyMapper.class);
        job.setReducerClass(DedupKeyReducer.class);
        job.setMapOutputValueClass(inputValueClass);
    } else if (dedupBy.equals("value")) {
        job.setMapperClass(DedupValueMapper.class);
        job.setReducerClass(DedupValueReducer.class);
        job.setMapOutputValueClass(inputKeyClass);
    }

    job.setInputFormatClass(inputFormatClass);
    if (inputFormat.equals("co.nubetech.hiho.dedup.DelimitedTextInputFormat")) {
        DelimitedTextInputFormat.setProperties(job, delimiter, column);
    }

    job.setMapOutputKeyClass(HihoTuple.class);

    job.setOutputKeyClass(inputKeyClass);
    job.setOutputValueClass(inputValueClass);
    job.setPartitionerClass(HihoHashPartitioner.class);
    FileInputFormat.setInputPaths(job, inputPath);
    job.setOutputFormatClass(outputFormatClass);
    FileOutputFormat.setOutputPath(job, new Path(outputPath));

    try {
        logger.debug("Output format class is " + job.getOutputFormatClass());
        logger.debug("Class is " + ReflectionUtils
                .newInstance(job.getOutputFormatClass(), job.getConfiguration()).getClass().getName());
        job.waitForCompletion(false);
        if (job.isComplete()) {
            Counters counters = job.getCounters();
            totalRecordsRead = counters.findCounter(DedupRecordCounter.TOTAL_RECORDS_READ).getValue();
            badRecords = counters.findCounter(DedupRecordCounter.BAD_RECORD).getValue();
            output = counters.findCounter(DedupRecordCounter.OUTPUT).getValue();
            duplicateRecords = totalRecordsRead - output;
            logger.info("Total records read are: " + totalRecordsRead);
            logger.info("Bad Records are: " + badRecords);
            logger.info("Output records are: " + output);
            logger.info("Duplicate records are: " + duplicateRecords);
        }

    } catch (Exception e) {
        e.printStackTrace();
    }
    return 0;
}

From source file:co.nubetech.hiho.dedup.TestDelimitedTextInputFormat.java

License:Apache License

@Test
public void testSetProperties() throws IOException {
    Configuration conf = new Configuration();
    Job job = new Job(conf);
    DelimitedTextInputFormat.setProperties(job, ",", 1);
    assertEquals(",", job.getConfiguration().get(DelimitedTextInputFormat.DELIMITER_CONF));
    assertEquals("1", job.getConfiguration().get(DelimitedTextInputFormat.COLUMN_CONF));
}

From source file:co.nubetech.hiho.job.DBQueryInputJob.java

License:Apache License

public void runJobs(Configuration conf, int jobCounter) throws IOException {

    try {//  www .j  av  a2s  . com
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new IOException(e1);
    }

    Job job = new Job(conf);
    for (Entry<String, String> entry : conf) {
        logger.warn("key, value " + entry.getKey() + "=" + entry.getValue());
    }

    // logger.debug("Number of maps " +
    // conf.getInt("mapred.map.tasks", 1));
    // conf.setInt(JobContext.NUM_MAPS,
    // conf.getInt("mapreduce.job.maps", 1));
    // job.getConfiguration().setInt("mapred.map.tasks", 4);
    job.getConfiguration().setInt(MRJobConfig.NUM_MAPS, conf.getInt(HIHOConf.NUMBER_MAPPERS, 1));
    logger.warn("Number of maps " + conf.getInt(MRJobConfig.NUM_MAPS, 1));

    job.setJobName("Import job");
    job.setJarByClass(DBQueryInputJob.class);

    String strategy = conf.get(HIHOConf.INPUT_OUTPUT_STRATEGY);
    OutputStrategyEnum os = OutputStrategyEnum.value(strategy);
    if (os == null) {
        throw new IllegalArgumentException("Wrong value of output strategy. Please correct");
    }
    if (os != OutputStrategyEnum.AVRO) {
        switch (os) {

        case DUMP: {
            // job.setMapperClass(DBImportMapper.class);
            break;
        }
        /*
         * case AVRO: { job.setMapperClass(DBInputAvroMapper.class); //
         * need avro in cp // job.setJarByClass(Schema.class); // need
         * jackson which is needed by avro - ugly! //
         * job.setJarByClass(ObjectMapper.class);
         * job.setMapOutputKeyClass(NullWritable.class);
         * job.setMapOutputValueClass(AvroValue.class);
         * job.setOutputKeyClass(NullWritable.class);
         * job.setOutputValueClass(AvroValue.class);
         * job.setOutputFormatClass(AvroOutputFormat.class);
         * 
         * AvroOutputFormat.setOutputPath(job, new
         * Path(getConf().get(HIHOConf.INPUT_OUTPUT_PATH))); break; }
         */
        case DELIMITED: {
            job.setMapperClass(DBInputDelimMapper.class);
            job.setMapOutputKeyClass(Text.class);
            job.setMapOutputValueClass(Text.class);
            job.setOutputKeyClass(Text.class);
            job.setOutputValueClass(Text.class);
            job.setOutputFormatClass(NoKeyOnlyValueOutputFormat.class);

            NoKeyOnlyValueOutputFormat.setOutputPath(job, new Path(getConf().get(HIHOConf.INPUT_OUTPUT_PATH)));
        }
        case JSON: {
            // job.setMapperClass(DBImportJsonMapper.class);
            // job.setJarByClass(ObjectMapper.class);
            break;
        }
        default: {
            job.setMapperClass(DBInputDelimMapper.class);
            job.setMapOutputKeyClass(Text.class);
            job.setMapOutputValueClass(Text.class);
            job.setOutputKeyClass(Text.class);
            job.setOutputValueClass(Text.class);
            job.setOutputFormatClass(NoKeyOnlyValueOutputFormat.class);

            NoKeyOnlyValueOutputFormat.setOutputPath(job, new Path(getConf().get(HIHOConf.INPUT_OUTPUT_PATH)));
            break;
        }
        }

        String inputQuery = conf.get(DBConfiguration.INPUT_QUERY);
        String inputBoundingQuery = conf.get(DBConfiguration.INPUT_BOUNDING_QUERY);
        logger.debug("About to set the params");
        DBQueryInputFormat.setInput(job, inputQuery, inputBoundingQuery, params);
        logger.debug("Set the params");

        job.setNumReduceTasks(0);

        try {
            // job.setJarByClass(Class.forName(conf.get(
            // org.apache.hadoop.mapred.lib.db.DBConfiguration.DRIVER_CLASS_PROPERTY)));
            logger.debug("OUTPUT format class is " + job.getOutputFormatClass());

            /*
             * org.apache.hadoop.mapreduce.OutputFormat<?, ?> output =
             * ReflectionUtils.newInstance(job.getOutputFormatClass(),
             * job.getConfiguration()); output.checkOutputSpecs(job);
             */
            logger.debug("Class is " + ReflectionUtils
                    .newInstance(job.getOutputFormatClass(), job.getConfiguration()).getClass().getName());
            job.waitForCompletion(false);
            if (conf.get(HIHOConf.INPUT_OUTPUT_LOADTO) != null) {
                generateHiveScript(conf, job, jobCounter);
                generatePigScript(conf, job);
            }

        }
        /*
         * catch (HIHOException h) { h.printStackTrace(); }
         */
        catch (Exception e) {
            e.printStackTrace();
        } catch (HIHOException e) {
            e.printStackTrace();
        }
    }
    // avro to be handled differently, thanks to all the incompatibilities
    // in the apis.
    else {
        String inputQuery = conf.get(DBConfiguration.INPUT_QUERY);
        String inputBoundingQuery = conf.get(DBConfiguration.INPUT_BOUNDING_QUERY);
        logger.debug("About to set the params");
        // co.nubetech.apache.hadoop.mapred.DBQueryInputFormat.setInput(job,
        // inputQuery, inputBoundingQuery, params);
        logger.debug("Set the params");

        JobConf jobConf = new JobConf(conf);

        try {
            GenericDBWritable queryWritable = getDBWritable(jobConf);
            Schema pair = DBMapper.getPairSchema(queryWritable.getColumns());

            AvroJob.setMapOutputSchema(jobConf, pair);
            GenericRecordAvroOutputFormat.setOutputPath(jobConf,
                    new Path(getConf().get(HIHOConf.INPUT_OUTPUT_PATH)));

            co.nubetech.apache.hadoop.mapred.DBQueryInputFormat.setInput(jobConf, inputQuery,
                    inputBoundingQuery, params);
            jobConf.setInputFormat(co.nubetech.apache.hadoop.mapred.DBQueryInputFormat.class);
            jobConf.setMapperClass(DBInputAvroMapper.class);
            jobConf.setMapOutputKeyClass(NullWritable.class);
            jobConf.setMapOutputValueClass(AvroValue.class);
            jobConf.setOutputKeyClass(NullWritable.class);
            jobConf.setOutputValueClass(Text.class);
            jobConf.setOutputFormat(GenericRecordAvroOutputFormat.class);
            jobConf.setJarByClass(DBQueryInputJob.class);
            jobConf.setStrings("io.serializations",
                    "org.apache.hadoop.io.serializer.JavaSerialization,org.apache.hadoop.io.serializer.WritableSerialization,org.apache.avro.mapred.AvroSerialization");
            jobConf.setNumReduceTasks(0);
            /*
             * jobConf.setOutputFormat(org.apache.hadoop.mapred.
             * SequenceFileOutputFormat.class);
             * org.apache.hadoop.mapred.SequenceFileOutputFormat
             * .setOutputPath(jobConf, new
             * Path(getConf().get(HIHOConf.INPUT_OUTPUT_PATH)));
             */
            JobClient.runJob(jobConf);
        } catch (Throwable e) {
            e.printStackTrace();
        }

    }

}

From source file:co.nubetech.hiho.job.ExportDelimitedToDB.java

License:Apache License

public int run(String[] args) throws IOException {
    Configuration conf = getConf();
    Job job = new Job(conf);
    job.setJobName("MySQLBulkLoading");
    job.setMapperClass(DelimitedLoadMapper.class);
    job.setJarByClass(DelimitedLoadMapper.class);
    job.setNumReduceTasks(0);//from ww w  .  ja  v a  2  s  .com

    job.setInputFormatClass(TextInputFormat.class);
    TextInputFormat.addInputPath(job, new Path(args[0]));

    job.setMapOutputKeyClass(NullWritable.class);
    job.setMapOutputValueClass(NullWritable.class);

    job.setOutputFormatClass(DBOutputFormat.class);
    int ret = 0;

    try {
        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    return ret;

}

From source file:co.nubetech.hiho.job.ExportToDB.java

License:Apache License

public int run(String[] args) throws Exception {
    Configuration conf = getConf();
    populateConfiguration(args, conf);//from w  w w.  ja v  a  2 s  . co  m
    try {
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new Exception(e1);
    }
    Job job = new Job(conf);
    job.getConfiguration().setInt(MRJobConfig.NUM_MAPS, conf.getInt(HIHOConf.NUMBER_MAPPERS, 1));
    job.setJobName("HihoDBExport");

    job.setMapperClass(GenericDBLoadDataMapper.class);
    job.setJarByClass(ExportToDB.class);
    job.setNumReduceTasks(0);
    job.setInputFormatClass(TextInputFormat.class);
    TextInputFormat.addInputPath(job, new Path(inputPath));
    GenericDBOutputFormat.setOutput(job, tableName, columnNames);

    int ret = 0;
    try {
        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    return ret;

}

From source file:co.nubetech.hiho.job.ExportToFTPServer.java

License:Apache License

@Override
public int run(String[] args) throws IOException {
    Configuration conf = getConf();
    populateConfiguration(args, conf);/*from  w ww .  j a v  a2  s.  c  o  m*/
    try {
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new IOException(e1);
    }

    for (Entry<String, String> entry : conf) {
        logger.debug("key, value " + entry.getKey() + "=" + entry.getValue());
    }
    Job job = new Job(conf);
    job.setMapperClass(TokenCounterMapper.class);
    job.setInputFormatClass(TextInputFormat.class);
    TextInputFormat.addInputPath(job, new Path(inputPath));
    job.setReducerClass(IntSumReducer.class);
    job.setOutputFormatClass(FTPTextOutputFormat.class);
    FTPTextOutputFormat.setOutputPath(job, new Path(outputPath));
    job.setJarByClass(ExportToFTPServer.class);
    job.setMapOutputKeyClass(Text.class);
    job.setMapOutputValueClass(IntWritable.class);
    job.setNumReduceTasks(2);

    int ret = 0;

    try {
        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    return ret;

}

From source file:co.nubetech.hiho.job.ExportToMySQLDB.java

License:Apache License

@Override
public int run(String[] args) throws IOException {

    Configuration conf = getConf();
    populateConfiguration(args, conf);/* www .  jav a 2  s  . com*/
    try {
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new IOException(e1);
    }
    Job job = new Job(conf);
    job.setJobName("MySQLBulkLoading");
    job.setMapperClass(MySQLLoadDataMapper.class);
    job.setJarByClass(MySQLLoadDataMapper.class);
    for (Entry<String, String> entry : conf) {
        logger.debug("key, value " + entry.getKey() + "=" + entry.getValue());
    }
    // verify required properties are loaded
    logger.debug(conf.get(DBConfiguration.URL_PROPERTY));
    logger.debug(conf.get(DBConfiguration.USERNAME_PROPERTY));
    logger.debug(conf.get(DBConfiguration.PASSWORD_PROPERTY));

    job.setNumReduceTasks(0);
    job.setInputFormatClass(FileStreamInputFormat.class);
    FileStreamInputFormat.addInputPath(job, new Path(inputPath));
    job.setMapOutputKeyClass(NullWritable.class);
    job.setMapOutputValueClass(NullWritable.class);
    // job.setJarByClass(com.mysql.jdbc.Driver.class);
    job.setOutputFormatClass(NullOutputFormat.class);

    int ret = 0;
    try {
        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    return ret;
}

From source file:co.nubetech.hiho.job.ExportToOracleDb.java

License:Apache License

@Override
public int run(String[] args) throws IOException {
    Configuration conf = getConf();
    for (Entry<String, String> entry : conf) {
        logger.debug("key, value " + entry.getKey() + "=" + entry.getValue());
    }//  w w w . jav  a  2 s  . c om

    for (int i = 0; i < args.length; i++) {
        logger.debug("Remaining arguments are" + " " + args[i]);
    }
    populateConfiguration(args, conf);

    try {
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new IOException(e1);
    }

    Job job = new Job(conf);
    job.setJobName("OracleLoading");
    job.setMapperClass(OracleLoadMapper.class);
    job.setJarByClass(ExportToOracleDb.class);
    job.getConfiguration().setInt(MRJobConfig.NUM_MAPS, conf.getInt(HIHOConf.NUMBER_MAPPERS, 1));

    try {
        // we first create the external table definition
        String query = conf.get(HIHOConf.EXTERNAL_TABLE_DML);
        // create table if user has specified
        if (query != null) {
            this.runQuery(query, conf);
        }
    } catch (HIHOException e1) {

        e1.printStackTrace();
    }

    // verify required properties are loaded

    job.setNumReduceTasks(0);
    job.setInputFormatClass(FileStreamInputFormat.class);
    FileStreamInputFormat.addInputPath(job, new Path(inputPath));
    job.setMapOutputKeyClass(NullWritable.class);
    job.setMapOutputValueClass(NullWritable.class);
    // job.setJarByClass(com.mysql.jdbc.Driver.class);
    job.setOutputFormatClass(NullOutputFormat.class);

    int ret = 0;
    try {

        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    // run alter table query and add locations
    try {
        this.runQuery(getAlterTableDML(new Path(inputPath), conf), conf);
    } catch (HIHOException e1) {

        e1.printStackTrace();
    }
    return ret;
}

From source file:co.nubetech.hiho.job.sf.ExportSalesForceJob.java

License:Apache License

@Override
public int run(String[] arg0) throws Exception {
    Configuration conf = getConf();
    populateConfiguration(arg0, conf);/*from   ww w  .ja va 2  s  . co m*/
    try {
        checkMandatoryConfs(conf);
    } catch (HIHOException e1) {
        e1.printStackTrace();
        throw new Exception(e1);
    }

    Job job = new Job(conf);
    job.setJobName("SaleForceLoading");
    job.setMapperClass(SalesForceLoadMapper.class);
    job.setJarByClass(SalesForceLoadMapper.class);
    job.setNumReduceTasks(0);

    job.setInputFormatClass(TextInputFormat.class);
    TextInputFormat.addInputPath(job, new Path(inputPath));
    // NLineInputFormat.setNumLinesPerSplit(job, 10);

    job.setMapOutputKeyClass(NullWritable.class);
    job.setMapOutputValueClass(NullWritable.class);

    job.setOutputFormatClass(NullOutputFormat.class);

    int ret = 0;

    try {
        ret = job.waitForCompletion(true) ? 0 : 1;
    } catch (Exception e) {
        e.printStackTrace();
    }
    return ret;

}