Example usage for org.apache.hadoop.mapred.jobcontrol JobControl JobControl

Introduction

In this page you can find the example usage for org.apache.hadoop.mapred.jobcontrol JobControl JobControl.

Prototype

public JobControl(String groupName) 

Source Link

Document

Construct a job control for a group of jobs.

Usage

From source file:org.apache.pig.test.pigmix.mapreduce.L14.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);//from   www .  j  a va 2s .c o  m
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    String user = System.getProperty("user.name");
    JobConf lp = new JobConf(L14.class);
    lp.setJobName("L14 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views_sorted"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/indexed_pages_14"));
    lp.setNumReduceTasks(0);
    Job loadPages = new Job(lp);

    JobConf lu = new JobConf(L14.class);
    lu.setJobName("L14 Load Users");
    lu.setInputFormat(TextInputFormat.class);
    lu.setOutputKeyClass(Text.class);
    lu.setOutputValueClass(Text.class);
    lu.setMapperClass(ReadUsers.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lu.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lu, new Path(inputDir + "/users_sorted"));
    FileOutputFormat.setOutputPath(lu, new Path(outputDir + "/indexed_users_14"));
    lu.setNumReduceTasks(0);
    Job loadUsers = new Job(lu);

    JobConf join = new JobConf(L14.class);
    join.setJobName("L14 Join Users and Pages");
    join.setInputFormat(KeyValueTextInputFormat.class);
    join.setOutputKeyClass(Text.class);
    join.setOutputValueClass(Text.class);
    join.setMapperClass(IdentityMapper.class);
    join.setReducerClass(Join.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        join.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_pages_14"));
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_users_14"));
    FileOutputFormat.setOutputPath(join, new Path(outputDir + "/L14out"));
    join.setNumReduceTasks(Integer.parseInt(parallel));
    Job joinJob = new Job(join);
    joinJob.addDependingJob(loadPages);
    joinJob.addDependingJob(loadUsers);

    JobControl jc = new JobControl("L14 join");
    jc.addJob(loadPages);
    jc.addJob(loadUsers);
    jc.addJob(joinJob);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L15.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);//from www  . j av a 2  s.c  om
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L15.class);
    lp.setJobName("L15 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setCombinerClass(Combiner.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L15out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L15 join");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L16.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);// w  w w .  j  av a 2 s .  co m
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L16.class);
    lp.setJobName("L16 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L16out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L16 join");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L17.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);/*from   w  w  w  . j  av  a 2  s  .c o  m*/
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L17.class);
    lp.setJobName("L17 Wide group by");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setCombinerClass(Combiner.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/widegroupbydata"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L17out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L17 group by");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L2.java

License:Apache License

public static void main(String[] args) throws IOException, URISyntaxException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);/*from   w w  w. jav a2  s  . c o m*/
    }
    String inputDir = args[0];
    String outputDir = args[1];
    JobConf lp = new JobConf(L2.class);
    lp.setJobName("L2 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(Join.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    DistributedCache.addCacheFile(new URI(inputDir + "/power_users"), lp);
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L2out"));
    lp.setNumReduceTasks(0);
    Job loadPages = new Job(lp);

    JobControl jc = new JobControl("L2 join");
    jc.addJob(loadPages);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L3.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);//  w w w  .j a va 2s . c o  m
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L3.class);
    lp.setJobName("L3 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/indexed_pages_3"));
    lp.setNumReduceTasks(0);
    Job loadPages = new Job(lp);

    JobConf lu = new JobConf(L3.class);
    lu.setJobName("L3 Load Users");
    lu.setInputFormat(TextInputFormat.class);
    lu.setOutputKeyClass(Text.class);
    lu.setOutputValueClass(Text.class);
    lu.setMapperClass(ReadUsers.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lu.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lu, new Path(inputDir + "/users"));
    FileOutputFormat.setOutputPath(lu, new Path(outputDir + "/indexed_users_3"));
    lu.setNumReduceTasks(0);
    Job loadUsers = new Job(lu);

    JobConf join = new JobConf(L3.class);
    join.setJobName("L3 Join Users and Pages");
    join.setInputFormat(KeyValueTextInputFormat.class);
    join.setOutputKeyClass(Text.class);
    join.setOutputValueClass(Text.class);
    join.setMapperClass(IdentityMapper.class);
    join.setReducerClass(Join.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        join.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_pages_3"));
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_users_3"));
    FileOutputFormat.setOutputPath(join, new Path(outputDir + "/L3out"));
    join.setNumReduceTasks(Integer.parseInt(parallel));
    Job joinJob = new Job(join);
    joinJob.addDependingJob(loadPages);
    joinJob.addDependingJob(loadUsers);

    JobControl jc = new JobControl("L3 join");
    jc.addJob(loadPages);
    jc.addJob(loadUsers);
    jc.addJob(joinJob);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L4.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);/*  w w  w.j ava 2  s.c o  m*/
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L4.class);
    lp.setJobName("L4 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setCombinerClass(Combiner.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L4out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L4 join");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L5.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);// w  w w. j a v  a 2 s  . com
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    String user = System.getProperty("user.name");
    JobConf lp = new JobConf(L5.class);
    lp.setJobName("L5 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/indexed_pages_5"));
    lp.setNumReduceTasks(0);
    Job loadPages = new Job(lp);

    JobConf lu = new JobConf(L5.class);
    lu.setJobName("L5 Load Users");
    lu.setInputFormat(TextInputFormat.class);
    lu.setOutputKeyClass(Text.class);
    lu.setOutputValueClass(Text.class);
    lu.setMapperClass(ReadUsers.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lu.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lu, new Path(inputDir + "/users"));
    FileOutputFormat.setOutputPath(lu, new Path(outputDir + "/indexed_users_5"));
    lu.setNumReduceTasks(0);
    Job loadUsers = new Job(lu);

    JobConf join = new JobConf(L5.class);
    join.setJobName("L5 Join Users and Pages");
    join.setInputFormat(KeyValueTextInputFormat.class);
    join.setOutputKeyClass(Text.class);
    join.setOutputValueClass(Text.class);
    join.setMapperClass(IdentityMapper.class);
    join.setReducerClass(Join.class);
    props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        join.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_pages_5"));
    FileInputFormat.addInputPath(join, new Path(outputDir + "/indexed_users_5"));
    FileOutputFormat.setOutputPath(join, new Path(outputDir + "/L5out"));
    join.setNumReduceTasks(Integer.parseInt(parallel));
    Job joinJob = new Job(join);
    joinJob.addDependingJob(loadPages);
    joinJob.addDependingJob(loadUsers);

    JobControl jc = new JobControl("L5 join");
    jc.addJob(loadPages);
    jc.addJob(loadUsers);
    jc.addJob(joinJob);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L6.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);/*from   w w  w. ja  va2 s .co  m*/
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L6.class);
    lp.setJobName("L6 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(IntWritable.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setCombinerClass(Group.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L6out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L6 join");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

From source file:org.apache.pig.test.pigmix.mapreduce.L7.java

License:Apache License

public static void main(String[] args) throws IOException {

    if (args.length != 3) {
        System.out.println("Parameters: inputDir outputDir parallel");
        System.exit(1);/*w  w  w  .jav a  2  s.  c o  m*/
    }
    String inputDir = args[0];
    String outputDir = args[1];
    String parallel = args[2];
    JobConf lp = new JobConf(L7.class);
    lp.setJobName("L7 Load Page Views");
    lp.setInputFormat(TextInputFormat.class);
    lp.setOutputKeyClass(Text.class);
    lp.setOutputValueClass(Text.class);
    lp.setMapperClass(ReadPageViews.class);
    lp.setCombinerClass(Combiner.class);
    lp.setReducerClass(Group.class);
    Properties props = System.getProperties();
    for (Map.Entry<Object, Object> entry : props.entrySet()) {
        lp.set((String) entry.getKey(), (String) entry.getValue());
    }
    FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
    FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L7out"));
    lp.setNumReduceTasks(Integer.parseInt(parallel));
    Job group = new Job(lp);

    JobControl jc = new JobControl("L7 join");
    jc.addJob(group);

    new Thread(jc).start();

    int i = 0;
    while (!jc.allFinished()) {
        ArrayList<Job> failures = jc.getFailedJobs();
        if (failures != null && failures.size() > 0) {
            for (Job failure : failures) {
                System.err.println(failure.getMessage());
            }
            break;
        }

        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
        }

        if (i % 10000 == 0) {
            System.out.println("Running jobs");
            ArrayList<Job> running = jc.getRunningJobs();
            if (running != null && running.size() > 0) {
                for (Job r : running) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Ready jobs");
            ArrayList<Job> ready = jc.getReadyJobs();
            if (ready != null && ready.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Waiting jobs");
            ArrayList<Job> waiting = jc.getWaitingJobs();
            if (waiting != null && waiting.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
            System.out.println("Successful jobs");
            ArrayList<Job> success = jc.getSuccessfulJobs();
            if (success != null && success.size() > 0) {
                for (Job r : ready) {
                    System.out.println(r.getJobName());
                }
            }
        }
        i++;
    }
    ArrayList<Job> failures = jc.getFailedJobs();
    if (failures != null && failures.size() > 0) {
        for (Job failure : failures) {
            System.err.println(failure.getMessage());
        }
    }
    jc.stop();
}

---

• «
• 1
• 2
• 3
• »