Java Code Examples for org.apache.hadoop.fs.FileSystem#get()

@SuppressWarnings("deprecation")
private void testDelegationTokenWithFS(Class fileSystemClass)
  throws Exception {
  createHttpFSServer();
  Configuration conf = new Configuration();
  conf.set("fs.webhdfs.impl", fileSystemClass.getName());
  conf.set("fs.hdfs.impl.disable.cache", "true");
  URI uri = new URI( "webhdfs://" +
                     TestJettyHelper.getJettyURL().toURI().getAuthority());
  FileSystem fs = FileSystem.get(uri, conf);
  Token<?> tokens[] = fs.addDelegationTokens("foo", null);
  fs.close();
  Assert.assertEquals(1, tokens.length);
  fs = FileSystem.get(uri, conf);
  ((DelegationTokenRenewer.Renewable) fs).setDelegationToken(tokens[0]);
  fs.listStatus(new Path("/"));
  fs.close();
}
 

@Test
public void testCopyFromLocalFileSystem() throws Exception {
  Path localFilePath = new Path(System.getProperty("test.build.data",
      "azure_test"));
  FileSystem localFs = FileSystem.get(new Configuration());
  localFs.delete(localFilePath, true);
  try {
    writeString(localFs, localFilePath, "Testing");
    Path dstPath = new Path("copiedFromLocal");
    assertTrue(FileUtil.copy(localFs, localFilePath, fs, dstPath, false,
        fs.getConf()));
    assertTrue(fs.exists(dstPath));
    assertEquals("Testing", readString(fs, dstPath));
    fs.delete(dstPath, true);
  } finally {
    localFs.delete(localFilePath, true);
  }
}
 

private void codecTestMapFile(Class<? extends CompressionCodec> clazz,
    CompressionType type, int records) throws Exception {
  
  FileSystem fs = FileSystem.get(conf);
  LOG.info("Creating MapFiles with " + records  + 
          " records using codec " + clazz.getSimpleName());
  Path path = new Path(new Path(
      System.getProperty("test.build.data", "/tmp")),
    clazz.getSimpleName() + "-" + type + "-" + records);

  LOG.info("Writing " + path);
  createMapFile(conf, fs, path, clazz.newInstance(), type, records);
  MapFile.Reader reader = new MapFile.Reader(path, conf);
  Text key1 = new Text("002");
  assertNotNull(reader.get(key1, new Text()));
  Text key2 = new Text("004");
  assertNotNull(reader.get(key2, new Text()));
}
 

@Test
public void testCleanup() throws IOException
{
  RandomAccessFile r = new RandomAccessFile(testMeta.testFile, "r");
  r.seek(0);
  byte[] b = r.readLine().getBytes();
  storage.store(new Slice(b, 0, b.length));
  byte[] val = storage.store(new Slice(b, 0, b.length));
  storage.flush();
  storage.clean(val);
  Configuration conf = new Configuration();
  FileSystem fs = FileSystem.get(conf);
  boolean exists = fs.exists(new Path(STORAGE_DIRECTORY + "/" + "0"));
  Assert.assertEquals("file should not exist", false, exists);
  r.close();
}
 

/**
 * Returns a boolean to denote whether a cache file is visible to all(public)
 * or not
 * @param conf
 * @param uri
 * @return true if the path in the uri is visible to all, false otherwise
 * @throws IOException
 */
static boolean isPublic(Configuration conf, URI uri,
    Map<URI, FileStatus> statCache) throws IOException {
  FileSystem fs = FileSystem.get(uri, conf);
  Path current = new Path(uri.getPath());
  //the leaf level file should be readable by others
  if (!checkPermissionOfOther(fs, current, FsAction.READ, statCache)) {
    return false;
  }
  return ancestorsHaveExecutePermissions(fs, current.getParent(), statCache);
}
 

@Test
public final void testCtasWithoutTableDefinition() throws Exception {
  ResultSet res = executeQuery();

  res.close();
  CatalogService catalog = testBase.getTestingCluster().getMaster().getCatalog();
  TableDesc desc = catalog.getTableDesc("testCtasWithoutTableDefinition");
  assertTrue(catalog.existsTable("testCtasWithoutTableDefinition"));

  assertTrue(desc.getSchema().contains("testCtasWithoutTableDefinition.col1"));
  PartitionMethodDesc partitionDesc = desc.getPartitionMethod();
  assertEquals(partitionDesc.getPartitionType(), CatalogProtos.PartitionType.COLUMN);
  assertEquals("key", partitionDesc.getExpressionSchema().getColumns().get(0).getSimpleName());

  FileSystem fs = FileSystem.get(testBase.getTestingCluster().getConfiguration());
  Path path = desc.getPath();
  assertTrue(fs.isDirectory(path));
  assertTrue(fs.isDirectory(new Path(path.toUri() + "/key=17.0")));
  assertTrue(fs.isDirectory(new Path(path.toUri() + "/key=36.0")));
  assertTrue(fs.isDirectory(new Path(path.toUri() + "/key=38.0")));
  assertTrue(fs.isDirectory(new Path(path.toUri() + "/key=45.0")));
  assertTrue(fs.isDirectory(new Path(path.toUri() + "/key=49.0")));
  assertEquals(5, desc.getStats().getNumRows().intValue());

  ResultSet res2 = executeFile("check1.sql");

  Map<Double, int []> resultRows1 = Maps.newHashMap();
  resultRows1.put(45.0d, new int[]{3, 2});
  resultRows1.put(38.0d, new int[]{2, 2});

  int i = 0;
  while(res2.next()) {
    assertEquals(resultRows1.get(res2.getDouble(3))[0], res2.getInt(1));
    assertEquals(resultRows1.get(res2.getDouble(3))[1], res2.getInt(2));
    i++;
  }
  res2.close();
  assertEquals(2, i);
}
 

public static void CHECK_INDEX(String schemaName, String tableName, String indexName, int level,
                               String outputFile, boolean fix, final ResultSet[] resultSet) throws Exception {
    FSDataOutputStream out = null;
    FileSystem fs = null;
    Map<String, List<String>> errors = null;

    try {
        LanguageConnectionContext lcc = ConnectionUtil.getCurrentLCC();
        Activation activation = lcc.getLastActivation();

        Configuration conf = (Configuration) SIDriver.driver().getConfiguration().getConfigSource().unwrapDelegate();
        String schema = EngineUtils.validateSchema(schemaName);
        String table = EngineUtils.validateTable(tableName);
        String index = EngineUtils.validateTable(indexName);

        fs = FileSystem.get(URI.create(outputFile), conf);
        out = fs.create(new Path(outputFile));

        errors = checkTable(schema, table, index, level, fix);
        resultSet[0] = processResults(errors, out, activation, outputFile);

    } finally {
        if (out != null) {
            out.close();
            if (errors == null || errors.size() == 0) {
                fs.delete(new Path(outputFile), true);
            }
        }

    }

}
 

@Test
public void testWebHdfsCustomDefaultPorts() throws IOException {
  URI uri = URI.create("webhdfs://localhost");
  WebHdfsFileSystem fs = (WebHdfsFileSystem) FileSystem.get(uri, conf);

  assertEquals(123, fs.getDefaultPort());
  assertEquals(uri, fs.getUri());
  assertEquals("127.0.0.1:123", fs.getCanonicalServiceName());
}
 

public void configure(JobConf conf) {
  try {
    FileSystem fs = FileSystem.get(conf);
    OutputStream os =
      fs.create(FileOutputFormat.getPathForCustomFile(conf, "test"));
    os.write(1);
    os.close();
  }
  catch (IOException ex) {
    throw new RuntimeException(ex);
  }
}
 

public HiveOrcSerDeManager(State props)
    throws IOException {
  super(props);
  this.fs = FileSystem.get(HadoopUtils.getConfFromState(props));

  List<String> extensions = props.getPropAsList(FILE_EXTENSIONS_KEY, DEFAULT_FILE_EXTENSIONS);
  this.fileExtensions = extensions.isEmpty() ? ImmutableList.of("") : extensions;

  this.ignoredFilePrefixes = props.getPropAsList(IGNORED_FILE_PREFIXES_KEY, DEFAULT_IGNORED_FILE_PREFIXES);
  this.checkOrcFormat = props.getPropAsBoolean(ENABLED_ORC_TYPE_CHECK, DEFAULT_ENABLED_ORC_TYPE_CHECK);
  this.metricContext = Instrumented.getMetricContext(props, HiveOrcSerDeManager.class);
  this.serDeWrapper = HiveSerDeWrapper.get(props.getProp(SERDE_TYPE_KEY, DEFAULT_SERDE_TYPE),
      Optional.of(props.getProp(INPUT_FORMAT_CLASS_KEY, DEFAULT_INPUT_FORMAT_CLASS)),
      Optional.of(props.getProp(OUTPUT_FORMAT_CLASS_KEY, DEFAULT_OUTPUT_FORMAT_CLASS)));
}
 

private void testOpen() throws Exception {
  FileSystem fs = FileSystem.get(getProxiedFSConf());
  Path path = new Path(getProxiedFSTestDir(), "foo.txt");
  OutputStream os = fs.create(path);
  os.write(1);
  os.close();
  fs.close();
  fs = getHttpFSFileSystem();
  InputStream is = fs.open(new Path(path.toUri().getPath()));
  Assert.assertEquals(is.read(), 1);
  is.close();
  fs.close();
}
 

public SegmentReader(Configuration conf, boolean co, boolean fe, boolean ge, boolean pa,
        boolean pd, boolean pt) {
  super(conf);
  this.co = co;
  this.fe = fe;
  this.ge = ge;
  this.pa = pa;
  this.pd = pd;
  this.pt = pt;
  try {
    this.fs = FileSystem.get(getConf());
  } catch (IOException e) {
    LOG.error("IOException:", e);
  }
}
 

private void cleanupStagingDir(Path uploadingDir)
{
    try {
        FileSystem hdfs = FileSystem.get(yarnClient.getConfig());
        if (hdfs.delete(uploadingDir, true)) {
            logger.info("Deleted staging directory {}", uploadingDir);
        }
    }
    catch (IOException e) {
        logger.warn("Failed to cleanup staging dir {}", uploadingDir, e);
    }
}
 

private void writeFile(String path, String message) {
    Path pt = new Path(path);
    logger.debug("Writing file out to {}, message {}", path, message);
    try {
        FileSystem fs = FileSystem.get(new Configuration());
        BufferedWriter br = new BufferedWriter(new OutputStreamWriter(fs.create(pt, true)));
        br.write(message);
        br.close();
    } catch (IOException e) {
        e.printStackTrace();
        throw new IllegalStateException("Could not write to file: " + path);
    }
}
 

private  long getNumberOfMobFiles(Configuration conf, TableName tableName, String family)
    throws IOException {
  FileSystem fs = FileSystem.get(conf);
  Path dir = MobUtils.getMobFamilyPath(conf, tableName, family);
  FileStatus[] stat = fs.listStatus(dir);
  for (FileStatus st : stat) {
    LOG.debug("MOB Directory content: {}", st.getPath());
  }
  LOG.debug("MOB Directory content total files: {}", stat.length);

  return stat.length;
}
 

/**
 * Write the sequence file.
 *
 * @param args the command-line arguments
 * @return the process exit code
 * @throws Exception if something goes wrong
 */
public int run(final String[] args) throws Exception {

  Cli cli = Cli.builder().setArgs(args).addOptions(CliCommonOpts.IOFileOpts.values()).build();
  int result = cli.runCmd();

  if (result != 0) {
    return result;
  }

  File inputFile = new File(cli.getArgValueAsString(CliCommonOpts.IOFileOpts.INPUT));
  Path outputPath = new Path(cli.getArgValueAsString(CliCommonOpts.IOFileOpts.OUTPUT));

  Configuration conf = super.getConf();

  FileSystem hdfs = FileSystem.get(conf);

  OutputStream os = hdfs.create(outputPath);
  writeToAvro(inputFile, os);

  return 0;
}
 

private FileSystem getFileSystem(String fsUri) throws IOException {
  return FileSystem.get(URI.create(fsUri), HadoopUtils.getConfFromState(this.props));
}
 

@Override
public FileSystem getFileSystemByUri(String uri) throws IOException {
  return FileSystem.get(URI.create(uri), conf);
}
 

public static void main(String[] args) throws IllegalArgumentException, IOException, URISyntaxException {
    String fromDirectory = args[0];
    String toEventsDirectory = args[1];
    String toOutliersDirectory = args[2];
    String metadataFile = args[3];
    
    // Detecting datasets.
    
    HashSet<String> datasets = new HashSet<String>();
    
    FileReader fileReader = new FileReader(metadataFile);
    BufferedReader bufferedReader = new BufferedReader(fileReader);

    String line;
    while((line = bufferedReader.readLine()) != null) {
        String[] parts = line.split(",");
        datasets.add(parts[0]);
    }    
    bufferedReader.close();
    
    // Downloading relationships.
    
    String relationshipPatternStr = "([a-zA-Z0-9]{4}\\-[a-zA-Z0-9]{4})\\-([a-zA-Z0-9]{4}\\-[a-zA-Z0-9]{4})";
    Pattern relationshipPattern = Pattern.compile(relationshipPatternStr);
    
    Configuration conf = new Configuration();
    FileSystem fs = FileSystem.get(conf);
    FileSystem localFS = FileSystem.getLocal(conf);

    for (FileStatus status : fs.listStatus(new Path(fs.getHomeDirectory() + "/" + fromDirectory))) {
        if (!status.isDirectory()) {
            continue;
        }
        Path file = status.getPath();
        
        Matcher m = relationshipPattern.matcher(file.getName());
        if (!m.find()) continue;
        
        String ds1 = m.group(1);
        String ds2 = m.group(2);
        
        if (!datasets.contains(ds1)) continue;
        if (!datasets.contains(ds2)) continue;
        
        for (FileStatus statusDir : fs.listStatus(file)) {
            if (!statusDir.isDirectory()) {
                continue;
            }
            
            Path fromPath = statusDir.getPath();
            String toPathStr;
            if (fromPath.getName().contains("events")) {
                toPathStr = toEventsDirectory + "/" +
                        fromPath.getParent().getName() + "-" + fromPath.getName();
            } else {
                toPathStr = toOutliersDirectory + "/" +
                        fromPath.getParent().getName() + "-" + fromPath.getName();
            }
            Path toPath = new Path(toPathStr);
            
            System.out.println("Copying:");
            System.out.println("  From: " + fromPath.toString());
            System.out.println("  To: " + toPath.toString());
            
            FileUtil.copyMerge(
                    fs, // HDFS File System
                    fromPath, // HDFS path
                    localFS, // Local File System
                    toPath, // Local Path
                    false, // Do not delete HDFS path
                    conf, // Configuration
                    null);
        }
    }
}
 

/**
 * Create distributed cache directory where distributed cache files will be
 * created by the MapReduce job {@link GenerateDistCacheData#JOB_NAME}.
 * @throws IOException
 */
private void createDistCacheDirectory() throws IOException {
  FileSystem fs = FileSystem.get(conf);
  FileSystem.mkdirs(fs, distCachePath, new FsPermission((short) 0777));
}