Java FileInputFormat.getInputPaths方法代码示例

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
public static void cut(String name, String in) throws Exception {
	Job job = createJob(name, in);
	job.setNumReduceTasks(0);
	boolean success = job.waitForCompletion(true);

	if (success) {
		// MapReduce reads an input and writes the result to an new location.
		// Hence it can not modify data in place, and we have to replace the input
		// with the output
		Path output = FileOutputFormat.getOutputPath(job);
		FileSystem fs = output.getFileSystem(job.getConfiguration());
		Path[] inputs = FileInputFormat.getInputPaths(job);
		for (int i = 0; i < inputs.length; i++) {
			fs.delete(inputs[i], true);
		}
		fs.rename(output, inputs[0]);
	}
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
/**
 * Submit this job to mapred. The state becomes RUNNING if submission 
 * is successful, FAILED otherwise.  
 */
protected synchronized void submit() {
  try {
    Configuration conf = job.getConfiguration();
    if (conf.getBoolean(CREATE_DIR, false)) {
      FileSystem fs = FileSystem.get(conf);
      Path inputPaths[] = FileInputFormat.getInputPaths(job);
      for (int i = 0; i < inputPaths.length; i++) {
        if (!fs.exists(inputPaths[i])) {
          try {
            fs.mkdirs(inputPaths[i]);
          } catch (IOException e) {

          }
        }
      }
    }
    job.submit();
    this.state = State.RUNNING;
  } catch (Exception ioe) {
    LOG.info(getJobName()+" got an error while submitting ",ioe);
    this.state = State.FAILED;
    this.message = StringUtils.stringifyException(ioe);
  }
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
@Override
public List<InputSplit> getSplits(JobContext context)
        throws IOException {
    Path[] paths = FileInputFormat.getInputPaths(context);

    return FluentIterable.from(BaseInputFormat.getSplits(context.getConfiguration(), paths))
            .transform(_fromSplit)
            .toList();
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
@Override
public BaseStatistics getStatistics(BaseStatistics cachedStats) throws IOException {
	// only gather base statistics for FileInputFormats
	if (!(mapreduceInputFormat instanceof FileInputFormat)) {
		return null;
	}

	JobContext jobContext = new JobContextImpl(configuration, null);

	final FileBaseStatistics cachedFileStats = (cachedStats instanceof FileBaseStatistics) ?
			(FileBaseStatistics) cachedStats : null;

	try {
		final org.apache.hadoop.fs.Path[] paths = FileInputFormat.getInputPaths(jobContext);
		return getFileStats(cachedFileStats, paths, new ArrayList<FileStatus>(1));
	} catch (IOException ioex) {
		if (LOG.isWarnEnabled()) {
			LOG.warn("Could not determine statistics due to an io error: "
					+ ioex.getMessage());
		}
	} catch (Throwable t) {
		if (LOG.isErrorEnabled()) {
			LOG.error("Unexpected problem while getting the file statistics: "
					+ t.getMessage(), t);
		}
	}

	// no statistics available
	return null;
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
@Override
public BaseStatistics getStatistics(BaseStatistics cachedStats) throws IOException {
	// only gather base statistics for FileInputFormats
	if (!(mapreduceInputFormat instanceof FileInputFormat)) {
		return null;
	}

	JobContext jobContext = new JobContextImpl(configuration, null);

	final FileBaseStatistics cachedFileStats = (cachedStats != null && cachedStats instanceof FileBaseStatistics) ?
			(FileBaseStatistics) cachedStats : null;

	try {
		final org.apache.hadoop.fs.Path[] paths = FileInputFormat.getInputPaths(jobContext);
		return getFileStats(cachedFileStats, paths, new ArrayList<FileStatus>(1));
	} catch (IOException ioex) {
		if (LOG.isWarnEnabled()) {
			LOG.warn("Could not determine statistics due to an io error: "
					+ ioex.getMessage());
		}
	} catch (Throwable t) {
		if (LOG.isErrorEnabled()) {
			LOG.error("Unexpected problem while getting the file statistics: "
					+ t.getMessage(), t);
		}
	}

	// no statistics available
	return null;
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
public int run(String [] argv) throws Exception {
  Job job = Job.getInstance(getConf());
  job.setJarByClass(GenericMRLoadGenerator.class);
  job.setMapperClass(SampleMapper.class);
  job.setReducerClass(SampleReducer.class);
  if (!parseArgs(argv, job)) {
    return -1;
  }

  Configuration conf = job.getConfiguration();
  if (null == FileOutputFormat.getOutputPath(job)) {
    // No output dir? No writes
    job.setOutputFormatClass(NullOutputFormat.class);
  }

  if (0 == FileInputFormat.getInputPaths(job).length) {
    // No input dir? Generate random data
    System.err.println("No input path; ignoring InputFormat");
    confRandom(job);
  } else if (null != conf.getClass(INDIRECT_INPUT_FORMAT, null)) {
    // specified IndirectInputFormat? Build src list
    JobClient jClient = new JobClient(conf);  
    Path tmpDir = new Path("/tmp");
    Random r = new Random();
    Path indirInputFile = new Path(tmpDir,
        Integer.toString(r.nextInt(Integer.MAX_VALUE), 36) + "_files");
    conf.set(INDIRECT_INPUT_FILE, indirInputFile.toString());
    SequenceFile.Writer writer = SequenceFile.createWriter(
        tmpDir.getFileSystem(conf), conf, indirInputFile,
        LongWritable.class, Text.class,
        SequenceFile.CompressionType.NONE);
    try {
      for (Path p : FileInputFormat.getInputPaths(job)) {
        FileSystem fs = p.getFileSystem(conf);
        Stack<Path> pathstack = new Stack<Path>();
        pathstack.push(p);
        while (!pathstack.empty()) {
          for (FileStatus stat : fs.listStatus(pathstack.pop())) {
            if (stat.isDirectory()) {
              if (!stat.getPath().getName().startsWith("_")) {
                pathstack.push(stat.getPath());
              }
            } else {
              writer.sync();
              writer.append(new LongWritable(stat.getLen()),
                  new Text(stat.getPath().toUri().toString()));
            }
          }
        }
      }
    } finally {
      writer.close();
    }
  }

  Date startTime = new Date();
  System.out.println("Job started: " + startTime);
  int ret = job.waitForCompletion(true) ? 0 : 1;
  Date endTime = new Date();
  System.out.println("Job ended: " + endTime);
  System.out.println("The job took " +
                     (endTime.getTime() - startTime.getTime()) /1000 +
                     " seconds.");

  return ret;
}
 

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; //导入方法依赖的package包/类
@Override
public int run(String[] strings) throws Exception {
  Configuration conf = getConf();
  String dir = conf.get(LindenJobConfig.INPUT_DIR, null);
  logger.info("input dir:" + dir);
  Path inputPath = new Path(StringUtils.unEscapeString(dir));
  Path outputPath = new Path(conf.get(LindenJobConfig.OUTPUT_DIR));
  String indexPath = conf.get(LindenJobConfig.INDEX_PATH);

  FileSystem fs = FileSystem.get(conf);
  if (fs.exists(outputPath)) {
    fs.delete(outputPath, true);
  }
  if (fs.exists(new Path(indexPath))) {
    fs.delete(new Path(indexPath), true);
  }

  int numShards = conf.getInt(LindenJobConfig.NUM_SHARDS, 1);
  Shard[] shards = createShards(indexPath, numShards);

  Shard.setIndexShards(conf, shards);

  //empty trash;
  (new Trash(conf)).expunge();

  Job job = Job.getInstance(conf, "linden-hadoop-indexing");
  job.setJarByClass(LindenJob.class);
  job.setMapperClass(LindenMapper.class);
  job.setCombinerClass(LindenCombiner.class);
  job.setReducerClass(LindenReducer.class);
  job.setMapOutputKeyClass(Shard.class);
  job.setMapOutputValueClass(IntermediateForm.class);
  job.setOutputKeyClass(Shard.class);
  job.setOutputValueClass(Text.class);
  job.setInputFormatClass(TextInputFormat.class);
  job.setOutputFormatClass(IndexUpdateOutputFormat.class);
  job.setReduceSpeculativeExecution(false);
  job.setNumReduceTasks(numShards);

  String lindenSchemaFile = conf.get(LindenJobConfig.SCHEMA_FILE_URL);
  if (lindenSchemaFile == null) {
    throw new IOException("no schema file is found");
  }
  logger.info("Adding schema file: " + lindenSchemaFile);
  job.addCacheFile(new URI(lindenSchemaFile + "#lindenSchema"));
  String lindenPropertiesFile = conf.get(LindenJobConfig.LINDEN_PROPERTIES_FILE_URL);
  if (lindenPropertiesFile == null) {
    throw new IOException("no linden properties file is found");
  }
  logger.info("Adding linden properties file: " + lindenPropertiesFile);
  job.addCacheFile(new URI(lindenPropertiesFile + "#lindenProperties"));

  FileInputFormat.setInputPaths(job, inputPath);
  FileOutputFormat.setOutputPath(job, outputPath);

  Path[] inputs = FileInputFormat.getInputPaths(job);
  StringBuilder buffer = new StringBuilder(inputs[0].toString());
  for (int i = 1; i < inputs.length; i++) {
    buffer.append(",");
    buffer.append(inputs[i].toString());
  }
  logger.info("mapreduce.input.dir = " + buffer.toString());
  logger.info("mapreduce.output.dir = " + FileOutputFormat.getOutputPath(job).toString());
  logger.info("mapreduce.job.num.reduce.tasks = " + job.getNumReduceTasks());
  logger.info(shards.length + " shards = " + conf.get(LindenJobConfig.INDEX_SHARDS));
  logger.info("mapreduce.input.format.class = " + job.getInputFormatClass());
  logger.info("mapreduce.output.format.class = " + job.getOutputFormatClass());
  logger.info("mapreduce.cluster.temp.dir = " + conf.get(MRJobConfig.TEMP_DIR));

  job.waitForCompletion(true);
  if (!job.isSuccessful()) {
    throw new RuntimeException("Job failed");
  }
  return 0;
}