Example usage for org.apache.hadoop.io Text toString

List of usage examples for org.apache.hadoop.io Text toString

Introduction

In this page you can find the example usage for org.apache.hadoop.io Text toString.

Prototype

@Override
public String toString() 

Source Link

Document

Convert text back to string

Usage

From source file:Assignment4_P4_MemoryConscious.MovieRatingMemConscious_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] movieInfo = value.toString().split("::");

    if (!movieInfo[1].trim().toLowerCase().equals("movieid")) {
        // extract movieID and rating
        movieID = new IntWritable(Integer.parseInt(movieInfo[1]));
        rating = new FloatWritable(Float.parseFloat(movieInfo[2]));

        // push rating in sorted hashmap
        outSortMap.put(rating, ONE);//w  ww  .  j a v a 2  s .  com

        // send this fellow to combiner now, come on, do it bleeeeeedy
        context.write(movieID, outSortMap);
    }
}

From source file:Assignment5_P2_DistinctIPAddress.DistinctIPAddress_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] ipInfo = value.toString().split("-");
    ipaddr = new Text(ipInfo[0].trim());
    context.write(ipaddr, NullWritable.get());
}

From source file:Assignment5_P3_PartitionPattern.Partition_IPAddress_By_Month_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] ipInfo = value.toString().split("-");
    String timing = ipInfo[2].trim();
    String[] cDate = timing.split(":");

    String createDate = cDate[0].substring(1, cDate[0].length()).trim();//System.out.println("Date found : '" + createDate + "'");

    Calendar cal = Calendar.getInstance();
    try {/*from  ww  w. ja v a  2 s. c  o  m*/
        cal.setTime(fmt.parse(createDate)); //System.out.println("Month evaluated by format : " + fmt.parse(createDate)); //System.out.println("Month evaluated by cal : " + cal.get(Calendar.MONTH));
        createMonth.set(cal.get(Calendar.MONTH));
    } catch (ParseException e) {
        e.printStackTrace();
    }

    //System.out.println("CREATE MONTH " + createMonth);
    context.write(createMonth, value);
}

From source file:Assignment5_P4_BinningPattern.Binning_IPAddress_By_Day_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] ipInfo = value.toString().split("-");
    String timing = ipInfo[2].trim();

    String createDate = timing.substring(1, timing.length()).trim();//System.out.println("Date found : '" + createDate + "'");

    Calendar cal = Calendar.getInstance();
    try {//www . j av  a 2 s. c om
        cal.setTime(fmt.parse(createDate)); //System.out.println("Month evaluated by format : " + fmt.parse(createDate)); //System.out.println("Month evaluated by cal : " + cal.get(Calendar.MONTH));
        createHour.set(cal.get(Calendar.HOUR_OF_DAY));
        multipleOutputs.write("textualBins", NullWritable.get(), value, createHour + "-hour");
        multipleOutputs.write("massaBins", NullWritable.get(), value, createHour + "-trial-by-witnessing");
    } catch (ParseException e) {
        e.printStackTrace();
    }
}

From source file:Assignment5_P6_StructureToHierarchyPattern.Structure_Hierarchy_Movie_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] movieInfo = value.toString().split(",");
    String movieId = movieInfo[0];

    //append 'T' for title identification
    String movieTitle = "T" + movieInfo[1];

    // movieID, movieTitle
    context.write(new Text(movieId), new Text(movieTitle));
}

From source file:Assignment5_P6_StructureToHierarchyPattern.Structure_Hierarchy_Reducer.java

public void reduce(Text key, Iterable<Text> values, Context context) throws IOException, InterruptedException {
    sb.append("<Movie>");

    for (Text value : values) {
        if (value.toString().charAt(0) == 'T' && count < 1) {
            String title = value.toString().substring(1, value.toString().length()).trim();
            // put it in a tag
            sb.append("<Title>").append(title).append("</Title>");
            // in count so next time it doesnt come in
            count++;//from w  w w  . j a  v a 2 s.  c o m
        } else {
            String tag = value.toString().substring(1, value.toString().length()).trim();

            // call genre fellow
            constructPropertyXml(tag);
        }
    }

    sb.append("</Movie>");
    context.write(null, new Text(sb.toString()));
}

From source file:Assignment5_P6_StructureToHierarchyPattern.Structure_Hierarchy_Tag_Mapper.java

public void map(Object key, Text value, Context context) throws IOException, InterruptedException {
    String[] movieInfo = value.toString().split(",");
    String movieId = movieInfo[1];

    //append 'A' for tag identification
    String movieTag = "A" + movieInfo[2];

    // movieID, movieTag
    context.write(new Text(movieId), new Text(movieTag));
}

From source file:at.illecker.hama.hybrid.examples.onlinecf.OnlineCF.java

License:Apache License

@Override
public boolean load(String path, boolean lazy) {
    this.m_isLazyLoadModel = lazy;
    this.m_modelPath = path;

    if (lazy == false) {
        Path dataPath = new Path(m_modelPath);
        Configuration conf = new Configuration();
        try {/*  w  ww  .  ja  v  a 2 s  .  c om*/
            FileSystem fs = dataPath.getFileSystem(conf);
            LinkedList<Path> files = new LinkedList<Path>();

            if (!fs.exists(dataPath)) {
                this.m_isLazyLoadModel = false;
                this.m_modelPath = null;
                return false;
            }

            if (!fs.isFile(dataPath)) {
                for (int i = 0; i < 100000; i++) {
                    Path partFile = new Path(
                            m_modelPath + "/part-" + String.valueOf(100000 + i).substring(1, 6));
                    if (fs.exists(partFile)) {
                        files.add(partFile);
                    } else {
                        break;
                    }
                }
            } else {
                files.add(dataPath);
            }

            LOG.info("loading model from " + path);
            for (Path file : files) {
                SequenceFile.Reader reader = new SequenceFile.Reader(fs, file, conf);

                Text key = new Text();
                PipesVectorWritable value = new PipesVectorWritable();
                String strKey = null;
                Long actualKey = null;
                String firstSymbol = null;

                while (reader.next(key, value) != false) {
                    strKey = key.toString();
                    firstSymbol = strKey.substring(0, 1);
                    try {
                        actualKey = Long.valueOf(strKey.substring(1));
                    } catch (Exception e) {
                        actualKey = new Long(0);
                    }

                    if (firstSymbol.equals(OnlineCF.DFLT_MODEL_ITEM_DELIM)) {
                        // LOG.info("loaded itemId: " + actualKey + " itemVector: "
                        // + value.getVector());
                        m_modelItemFactorizedValues.put(actualKey, new PipesVectorWritable(value));
                    } else if (firstSymbol.equals(OnlineCF.DFLT_MODEL_USER_DELIM)) {
                        // LOG.info("loaded userId: " + actualKey + " userVector: "
                        // + value.getVector());
                        m_modelUserFactorizedValues.put(actualKey, new PipesVectorWritable(value));
                    } else {
                        // unknown
                        continue;
                    }
                }
                reader.close();
            }

            LOG.info("loaded: " + m_modelUserFactorizedValues.size() + " users, "
                    + m_modelItemFactorizedValues.size() + " items");
            // for (Long user : m_modelUserFactorizedValues.keySet()) {
            // LOG.info("userId: " + user + " userVector: "
            // + m_modelUserFactorizedValues.get(user));
            // }
            // for (Long item : m_modelItemFactorizedValues.keySet()) {
            // LOG.info("itemId: " + item + " itemVector: "
            // + m_modelItemFactorizedValues.get(item));
            // }

        } catch (Exception e) {
            e.printStackTrace();
            this.m_isLazyLoadModel = false;
            this.m_modelPath = null;
            return false;
        }
    }
    return true;
}

From source file:at.illecker.hama.hybrid.examples.summation.SummationBSP.java

License:Apache License

@Override
public void bsp(BSPPeer<Text, Text, Text, DoubleWritable, DoubleWritable> peer)
        throws IOException, SyncException, InterruptedException {

    BSPJob job = new BSPJob((HamaConfiguration) peer.getConfiguration());
    FileSystem fs = FileSystem.get(peer.getConfiguration());
    FSDataOutputStream outStream = fs/*from w  w w. ja va  2  s  .c  o m*/
            .create(new Path(FileOutputFormat.getOutputPath(job), peer.getTaskId() + ".log"));

    outStream.writeChars("SummationBSP.bsp executed on CPU!\n");

    double intermediateSum = 0.0;
    Text key = new Text();
    Text value = new Text();

    while (peer.readNext(key, value)) {
        outStream.writeChars("SummationBSP.bsp key: " + key + " value: " + value + "\n");
        intermediateSum += Double.parseDouble(value.toString());
    }

    outStream.writeChars("SummationBSP.bsp send intermediateSum: " + intermediateSum + "\n");

    peer.send(m_masterTask, new DoubleWritable(intermediateSum));
    peer.sync();

    // Consume messages
    if (peer.getPeerName().equals(m_masterTask)) {
        outStream.writeChars("SummationBSP.bsp consume messages...\n");

        double sum = 0.0;
        int msg_count = peer.getNumCurrentMessages();

        for (int i = 0; i < msg_count; i++) {
            DoubleWritable msg = peer.getCurrentMessage();
            outStream.writeChars("SummationBSP.bsp message: " + msg.get() + "\n");
            sum += msg.get();
        }

        outStream.writeChars("SummationBSP.bsp write Sum: " + sum + "\n");
        peer.write(new Text("Sum"), new DoubleWritable(sum));
    }
    outStream.close();
}

From source file:authordetect.input.SingleBookReader.java

private boolean containsTitleOrAuthor(Text line, int option) throws IOException {
    String lineString = line.toString();
    String target;/* w  w w.  j  a va 2s.  com*/

    if (option == 0) {
        target = "Author";
        if (lineString.startsWith(target)) {
            title = lineString.split(":")[1].substring(1);
            return true;
        } else {
            return false;
        }
    } else {
        target = "Title";
        if (lineString.startsWith(target)) {
            title = lineString.split(":")[1].substring(1);
        }
        if (lineString.startsWith("Author")) {
            String author = lineString.split(":")[1].substring(1);
            title = title.concat("_" + author);
            return true;
        }
        return false;
    }

}