Beispiel #1
0
  public void runParseTest(
      String fieldTerminator,
      String lineTerminator,
      String encloser,
      String escape,
      boolean encloseRequired)
      throws IOException {

    ClassLoader prevClassLoader = null;

    String[] argv =
        getArgv(true, fieldTerminator, lineTerminator, encloser, escape, encloseRequired);
    runImport(argv);
    try {
      String tableClassName = getTableName();

      argv = getArgv(false, fieldTerminator, lineTerminator, encloser, escape, encloseRequired);
      SqoopOptions opts = new ImportTool().parseArguments(argv, null, null, true);

      CompilationManager compileMgr = new CompilationManager(opts);
      String jarFileName = compileMgr.getJarFilename();

      // Make sure the user's class is loaded into our address space.
      prevClassLoader = ClassLoaderStack.addJarFile(jarFileName, tableClassName);

      JobConf job = new JobConf();
      job.setJar(jarFileName);

      // Tell the job what class we're testing.
      job.set(ReparseMapper.USER_TYPE_NAME_KEY, tableClassName);

      // use local mode in the same JVM.
      ConfigurationHelper.setJobtrackerAddr(job, "local");
      if (!BaseSqoopTestCase.isOnPhysicalCluster()) {
        job.set(CommonArgs.FS_DEFAULT_NAME, CommonArgs.LOCAL_FS);
      }
      String warehouseDir = getWarehouseDir();
      Path warehousePath = new Path(warehouseDir);
      Path inputPath = new Path(warehousePath, getTableName());
      Path outputPath = new Path(warehousePath, getTableName() + "-out");

      job.setMapperClass(ReparseMapper.class);
      job.setNumReduceTasks(0);
      FileInputFormat.addInputPath(job, inputPath);
      FileOutputFormat.setOutputPath(job, outputPath);

      job.setOutputKeyClass(Text.class);
      job.setOutputValueClass(NullWritable.class);

      JobClient.runJob(job);
    } catch (InvalidOptionsException ioe) {
      fail(ioe.toString());
    } catch (ParseException pe) {
      fail(pe.toString());
    } finally {
      if (null != prevClassLoader) {
        ClassLoaderStack.setCurrentClassLoader(prevClassLoader);
      }
    }
  }
Beispiel #2
0
  public List<InputSplit> split(Configuration conf, ResultSet results, String colName)
      throws SQLException {

    LOG.warn("Generating splits for a floating-point index column. Due to the");
    LOG.warn("imprecise representation of floating-point values in Java, this");
    LOG.warn("may result in an incomplete import.");
    LOG.warn("You are strongly encouraged to choose an integral split column.");

    List<InputSplit> splits = new ArrayList<InputSplit>();

    if (results.getString(1) == null && results.getString(2) == null) {
      // Range is null to null. Return a null split accordingly.
      splits.add(
          new DataDrivenDBInputFormat.DataDrivenDBInputSplit(
              colName + " IS NULL", colName + " IS NULL"));
      return splits;
    }

    double minVal = results.getDouble(1);
    double maxVal = results.getDouble(2);

    // Use this as a hint. May need an extra task if the size doesn't
    // divide cleanly.
    int numSplits = ConfigurationHelper.getConfNumMaps(conf);
    double splitSize = (maxVal - minVal) / (double) numSplits;

    if (splitSize < MIN_INCREMENT) {
      splitSize = MIN_INCREMENT;
    }

    String lowClausePrefix = colName + " >= ";
    String highClausePrefix = colName + " < ";

    double curLower = minVal;
    double curUpper = curLower + splitSize;

    while (curUpper < maxVal) {
      splits.add(
          new DataDrivenDBInputFormat.DataDrivenDBInputSplit(
              lowClausePrefix + Double.toString(curLower),
              highClausePrefix + Double.toString(curUpper)));

      curLower = curUpper;
      curUpper += splitSize;
    }

    // Catch any overage and create the closed interval for the last split.
    if (curLower <= maxVal || splits.size() == 1) {
      splits.add(
          new DataDrivenDBInputFormat.DataDrivenDBInputSplit(
              lowClausePrefix + Double.toString(curUpper),
              colName + " <= " + Double.toString(maxVal)));
    }

    if (results.getString(1) == null || results.getString(2) == null) {
      // At least one extrema is null; add a null split.
      splits.add(
          new DataDrivenDBInputFormat.DataDrivenDBInputSplit(
              colName + " IS NULL", colName + " IS NULL"));
    }

    return splits;
  }