Example usage for org.apache.hadoop.mapreduce MRJobConfig IO_SORT_MB

Introduction

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

Prototype

String IO_SORT_MB

To view the source code for org.apache.hadoop.mapreduce MRJobConfig IO_SORT_MB.

Click Source Link

Usage

From source file:com.msd.gin.halyard.tools.HalyardBulkLoad.java

License:Apache License

@Override
public int run(String[] args) throws Exception {
    if (args.length != 3) {
        System.err.println("Usage: bulkload [-D" + MRJobConfig.QUEUE_NAME + "=proofofconcepts] [-D"
                + SKIP_INVALID_PROPERTY + "=true] [-D" + SPLIT_BITS_PROPERTY + "=8] [-D"
                + DEFAULT_CONTEXT_PROPERTY + "=http://new_context] [-D" + OVERRIDE_CONTEXT_PROPERTY
                + "=true] <input_path(s)> <output_path> <table_name>");
        return -1;
    }//w ww  .j a  v a 2 s  . co m
    TableMapReduceUtil.addDependencyJars(getConf(), NTriplesUtil.class, Rio.class, AbstractRDFHandler.class,
            RDFFormat.class, RDFParser.class);
    HBaseConfiguration.addHbaseResources(getConf());
    if (SnappyCodec.isNativeCodeLoaded()) {
        getConf().setBoolean(MRJobConfig.MAP_OUTPUT_COMPRESS, true);
        getConf().setClass(MRJobConfig.MAP_OUTPUT_COMPRESS_CODEC, SnappyCodec.class, CompressionCodec.class);
    }
    getConf().setDouble(MRJobConfig.COMPLETED_MAPS_FOR_REDUCE_SLOWSTART, 1.0);
    getConf().setLong(MRJobConfig.TASK_TIMEOUT, 3600000l);
    getConf().setInt(MRJobConfig.IO_SORT_FACTOR, 100);
    getConf().setInt(MRJobConfig.IO_SORT_MB, 1000);
    getConf().setInt(FileInputFormat.SPLIT_MAXSIZE, 1000000000);
    getConf().setInt(LoadIncrementalHFiles.MAX_FILES_PER_REGION_PER_FAMILY, 2048);
    Job job = Job.getInstance(getConf(), "HalyardBulkLoad -> " + args[1] + " -> " + args[2]);
    job.setJarByClass(HalyardBulkLoad.class);
    job.setMapperClass(RDFMapper.class);
    job.setMapOutputKeyClass(ImmutableBytesWritable.class);
    job.setMapOutputValueClass(KeyValue.class);
    job.setInputFormatClass(RioFileInputFormat.class);
    job.setSpeculativeExecution(false);
    job.setReduceSpeculativeExecution(false);
    Map<String, Integer> contextSplitsMap = new HashMap<>();
    for (Map.Entry<String, String> me : getConf().getValByRegex(CONTEXT_SPLIT_REGEXP).entrySet()) {
        int splits = Integer.parseInt(me.getKey().substring(me.getKey().lastIndexOf('.') + 1));
        StringTokenizer stk = new StringTokenizer(me.getValue(), ",");
        while (stk.hasMoreTokens()) {
            contextSplitsMap.put(stk.nextToken(), splits);
        }
    }
    try (HTable hTable = HalyardTableUtils.getTable(getConf(), args[2], true,
            getConf().getInt(SPLIT_BITS_PROPERTY, 3), contextSplitsMap)) {
        HFileOutputFormat2.configureIncrementalLoad(job, hTable.getTableDescriptor(),
                hTable.getRegionLocator());
        FileInputFormat.setInputDirRecursive(job, true);
        FileInputFormat.setInputPaths(job, args[0]);
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        TableMapReduceUtil.addDependencyJars(job);
        TableMapReduceUtil.initCredentials(job);
        if (job.waitForCompletion(true)) {
            new LoadIncrementalHFiles(getConf()).doBulkLoad(new Path(args[1]), hTable);
            LOG.info("Bulk Load Completed..");
            return 0;
        }
    }
    return -1;
}

---

From source file:com.msd.gin.halyard.tools.HalyardBulkUpdate.java

License:Apache License

@Override
public int run(String[] args) throws Exception {
    if (args.length != 3) {
        System.err.println("Usage: bulkupdate [-D" + MRJobConfig.QUEUE_NAME + "=proofofconcepts] [-D"
                + DEFAULT_CONTEXT_PROPERTY + "=http://new_context] [-D" + OVERRIDE_CONTEXT_PROPERTY
                + "=true] <input_file_with_SPARQL_queries> <output_path> <table_name>");
        return -1;
    }//from   www .j av a  2 s. c  o m
    TableMapReduceUtil.addDependencyJars(getConf(), NTriplesUtil.class, Rio.class, RDFFormat.class,
            RDFParser.class);
    HBaseConfiguration.addHbaseResources(getConf());
    if (SnappyCodec.isNativeCodeLoaded()) {
        getConf().setBoolean(MRJobConfig.MAP_OUTPUT_COMPRESS, true);
        getConf().setClass(MRJobConfig.MAP_OUTPUT_COMPRESS_CODEC, SnappyCodec.class, CompressionCodec.class);
    }
    getConf().setDouble(MRJobConfig.COMPLETED_MAPS_FOR_REDUCE_SLOWSTART, 1.0);
    getConf().setLong(MRJobConfig.TASK_TIMEOUT, 3600000l);
    getConf().setInt(MRJobConfig.IO_SORT_FACTOR, 100);
    getConf().setInt(MRJobConfig.IO_SORT_MB, 1000);
    getConf().setInt(FileInputFormat.SPLIT_MAXSIZE, 1000000000);
    getConf().setInt(LoadIncrementalHFiles.MAX_FILES_PER_REGION_PER_FAMILY, 2048);
    getConf().setStrings(TABLE_NAME_PROPERTY, args[2]);
    Job job = Job.getInstance(getConf(), "HalyardBulkUpdate -> " + args[1] + " -> " + args[2]);
    NLineInputFormat.setNumLinesPerSplit(job, 1);
    job.setJarByClass(HalyardBulkUpdate.class);
    job.setMapperClass(SPARQLMapper.class);
    job.setMapOutputKeyClass(ImmutableBytesWritable.class);
    job.setMapOutputValueClass(KeyValue.class);
    job.setInputFormatClass(NLineInputFormat.class);
    job.setSpeculativeExecution(false);
    job.setReduceSpeculativeExecution(false);
    try (HTable hTable = HalyardTableUtils.getTable(getConf(), args[2], false, 0, null)) {
        HFileOutputFormat2.configureIncrementalLoad(job, hTable.getTableDescriptor(),
                hTable.getRegionLocator());
        FileInputFormat.setInputPaths(job, args[0]);
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        TableMapReduceUtil.addDependencyJars(job);
        TableMapReduceUtil.initCredentials(job);
        if (job.waitForCompletion(true)) {
            new LoadIncrementalHFiles(getConf()).doBulkLoad(new Path(args[1]), hTable);
            LOG.info("Bulk Update Completed..");
            return 0;
        }
    }
    return -1;
}

---

From source file:com.msd.gin.halyard.tools.HalyardHiveLoad.java

License:Apache License

@Override
public int run(String[] args) throws Exception {
    if (args.length != 3) {
        System.err.println("Usage: hiveload -D" + RDF_MIME_TYPE_PROPERTY + "='application/ld+json' [-D"
                + MRJobConfig.QUEUE_NAME + "=proofofconcepts] [-D" + HIVE_DATA_COLUMN_INDEX_PROPERTY + "=3] [-D"
                + BASE_URI_PROPERTY + "='http://my_base_uri/'] [-D" + HalyardBulkLoad.SPLIT_BITS_PROPERTY
                + "=8] [-D" + HalyardBulkLoad.DEFAULT_CONTEXT_PROPERTY + "=http://new_context] [-D"
                + HalyardBulkLoad.OVERRIDE_CONTEXT_PROPERTY
                + "=true] <hive_table_name> <output_path> <hbase_table_name>");
        return -1;
    }//w w w  .j  av  a2s  .  c o m
    TableMapReduceUtil.addDependencyJars(getConf(), NTriplesUtil.class, Rio.class, AbstractRDFHandler.class,
            RDFFormat.class, RDFParser.class);
    HBaseConfiguration.addHbaseResources(getConf());
    if (SnappyCodec.isNativeCodeLoaded()) {
        getConf().setBoolean(MRJobConfig.MAP_OUTPUT_COMPRESS, true);
        getConf().setClass(MRJobConfig.MAP_OUTPUT_COMPRESS_CODEC, SnappyCodec.class, CompressionCodec.class);
    }
    getConf().setDouble(MRJobConfig.COMPLETED_MAPS_FOR_REDUCE_SLOWSTART, 1.0);
    getConf().setLong(MRJobConfig.TASK_TIMEOUT, 3600000l);
    getConf().setInt(MRJobConfig.IO_SORT_FACTOR, 100);
    getConf().setInt(MRJobConfig.IO_SORT_MB, 1000);
    getConf().setInt(FileInputFormat.SPLIT_MAXSIZE, 1000000000);
    getConf().setInt(LoadIncrementalHFiles.MAX_FILES_PER_REGION_PER_FAMILY, 2048);
    Job job = Job.getInstance(getConf(), "HalyardHiveLoad -> " + args[1] + " -> " + args[2]);
    int i = args[0].indexOf('.');
    HCatInputFormat.setInput(job, i > 0 ? args[0].substring(0, i) : null, args[0].substring(i + 1));
    job.setJarByClass(HalyardHiveLoad.class);
    job.setMapperClass(HiveMapper.class);
    job.setMapOutputKeyClass(ImmutableBytesWritable.class);
    job.setMapOutputValueClass(KeyValue.class);
    job.setInputFormatClass(HCatInputFormat.class);
    job.setSpeculativeExecution(false);
    job.setReduceSpeculativeExecution(false);
    Map<String, Integer> contextSplitsMap = new HashMap<>();
    for (Map.Entry<String, String> me : getConf().getValByRegex(HalyardBulkLoad.CONTEXT_SPLIT_REGEXP)
            .entrySet()) {
        int splits = Integer.parseInt(me.getKey().substring(me.getKey().lastIndexOf('.') + 1));
        StringTokenizer stk = new StringTokenizer(me.getValue(), ",");
        while (stk.hasMoreTokens()) {
            contextSplitsMap.put(stk.nextToken(), splits);
        }
    }
    try (HTable hTable = HalyardTableUtils.getTable(getConf(), args[2], true,
            getConf().getInt(HalyardBulkLoad.SPLIT_BITS_PROPERTY, 3), contextSplitsMap)) {
        HFileOutputFormat2.configureIncrementalLoad(job, hTable.getTableDescriptor(),
                hTable.getRegionLocator());
        FileInputFormat.setInputDirRecursive(job, true);
        FileInputFormat.setInputPaths(job, args[0]);
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        TableMapReduceUtil.addDependencyJars(job);
        TableMapReduceUtil.initCredentials(job);
        if (job.waitForCompletion(true)) {
            new LoadIncrementalHFiles(getConf()).doBulkLoad(new Path(args[1]), hTable);
            LOG.info("Bulk Load Completed..");
            return 0;
        }
    }
    return -1;
}

---