Java Code Examples for org.apache.lucene.index.DirectoryReader#numDocs()

private void assertData(int totalShardCount) throws IOException {
  Partitioner<IntWritable, IntWritable> partitioner = new HashPartitioner<IntWritable, IntWritable>();
  for (int i = 0; i < totalShardCount; i++) {
    HdfsDirectory directory = new HdfsDirectory(configuration, new Path(path, ShardUtil.getShardName(i)));
    DirectoryReader reader = DirectoryReader.open(directory);
    int numDocs = reader.numDocs();
    for (int d = 0; d < numDocs; d++) {
      Document document = reader.document(d);
      IndexableField field = document.getField("id");
      Integer id = (Integer) field.numericValue();
      int partition = partitioner.getPartition(new IntWritable(id), null, totalShardCount);
      assertEquals(i, partition);
    }
    reader.close();
  }
}
 

public int getLiveRecords() {
  checkIfChanged();
  try(Searcher searcher = acquireSearcher()) {
    DirectoryReader reader = (DirectoryReader) searcher.searcher.getIndexReader();
    return reader.numDocs();
  }
}
 

/**
 * Takes the categories from the given taxonomy directory, and adds the
 * missing ones to this taxonomy. Additionally, it fills the given
 * {@link OrdinalMap} with a mapping from the original ordinal to the new
 * ordinal.
 */
public void addTaxonomy(Directory taxoDir, OrdinalMap map) throws IOException {
  ensureOpen();
  DirectoryReader r = DirectoryReader.open(taxoDir);
  try {
    final int size = r.numDocs();
    final OrdinalMap ordinalMap = map;
    ordinalMap.setSize(size);
    int base = 0;
    PostingsEnum docs = null;
    for (final LeafReaderContext ctx : r.leaves()) {
      final LeafReader ar = ctx.reader();
      final Terms terms = ar.terms(Consts.FULL);
      // TODO: share per-segment TermsEnum here!
      TermsEnum te = terms.iterator();
      while (te.next() != null) {
        FacetLabel cp = new FacetLabel(FacetsConfig.stringToPath(te.term().utf8ToString()));
        final int ordinal = addCategory(cp);
        docs = te.postings(docs, PostingsEnum.NONE);
        ordinalMap.addMapping(docs.nextDoc() + base, ordinal);
      }
      base += ar.maxDoc(); // no deletions, so we're ok
    }
    ordinalMap.addDone();
  } finally {
    r.close();
  }
}
 

private void doTestSplitAlternately(SolrIndexSplitter.SplitMethod splitMethod) throws Exception {
  LocalSolrQueryRequest request = null;
  Directory directory = null;
  try {
    // add an even number of docs
    int max = (1 + random().nextInt(10)) * 3;
    log.info("Adding {} number of documents", max);
    for (int i = 0; i < max; i++) {
      assertU(adoc("id", String.valueOf(i)));
    }
    assertU(commit());

    request = lrf.makeRequest("q", "dummy");
    SolrQueryResponse rsp = new SolrQueryResponse();
    SplitIndexCommand command = new SplitIndexCommand(request, rsp,
        Lists.newArrayList(indexDir1.getAbsolutePath(), indexDir2.getAbsolutePath(), indexDir3.getAbsolutePath()),
        null, null, new PlainIdRouter(), null, null, splitMethod);
    doSplit(command);

    directory = h.getCore().getDirectoryFactory().get(indexDir1.getAbsolutePath(),
        DirectoryFactory.DirContext.DEFAULT, h.getCore().getSolrConfig().indexConfig.lockType);
    DirectoryReader reader = DirectoryReader.open(directory);
    int numDocs1 = reader.numDocs();
    reader.close();
    h.getCore().getDirectoryFactory().release(directory);
    directory = h.getCore().getDirectoryFactory().get(indexDir2.getAbsolutePath(),
        DirectoryFactory.DirContext.DEFAULT, h.getCore().getSolrConfig().indexConfig.lockType);
    reader = DirectoryReader.open(directory);
    int numDocs2 = reader.numDocs();
    reader.close();
    h.getCore().getDirectoryFactory().release(directory);
    directory = h.getCore().getDirectoryFactory().get(indexDir3.getAbsolutePath(),
        DirectoryFactory.DirContext.DEFAULT, h.getCore().getSolrConfig().indexConfig.lockType);
    reader = DirectoryReader.open(directory);
    int numDocs3 = reader.numDocs();
    reader.close();
    h.getCore().getDirectoryFactory().release(directory);
    directory = null;
    assertEquals("split indexes lost some documents!", max, numDocs1 + numDocs2 + numDocs3);
    assertEquals("split index1 has wrong number of documents", max / 3, numDocs1);
    assertEquals("split index2 has wrong number of documents", max / 3, numDocs2);
    assertEquals("split index3 has wrong number of documents", max / 3, numDocs3);
  } finally {
    if (request != null) request.close(); // decrefs the searcher
    if (directory != null)  {
      // perhaps an assert failed, release the directory
      h.getCore().getDirectoryFactory().release(directory);
    }
  }
}
 

@Test
public void testLargeDocs() throws IOException {
  Random random = new Random();
  Iterable<? extends IndexableField> doc = getLargeDoc(random);
  RAMDirectory directory = new RAMDirectory();
  IndexWriterConfig conf1 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf1.setCodec(new Blur024Codec());
  IndexWriter writer1 = new IndexWriter(directory, conf1);
  writer1.addDocument(doc);
  writer1.close();

  DirectoryReader reader1 = DirectoryReader.open(directory);
  int numDocs1 = reader1.numDocs();
  assertEquals(1, numDocs1);

  // for (int i = 0; i < numDocs1; i++) {
  // System.out.println(reader1.document(i));
  // }

  IndexWriterConfig conf2 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf2.setCodec(new Blur024Codec(1 << 16, CompressionMode.HIGH_COMPRESSION));
  IndexWriter writer2 = new IndexWriter(directory, conf2);
  writer2.addDocument(doc);
  writer2.close();

  DirectoryReader reader2 = DirectoryReader.open(directory);
  int numDocs2 = reader2.numDocs();
  assertEquals(2, numDocs2);

  for (int i = 0; i < 2; i++) {

    long t1 = System.nanoTime();
    Document document1 = reader1.document(0);
    long t2 = System.nanoTime();
    Document document2 = reader2.document(1);
    long t3 = System.nanoTime();

    System.out.println((t3 - t2) / 1000000.0);
    System.out.println((t2 - t1) / 1000000.0);

    System.out.println("doc1 " + document1.hashCode());
    System.out.println("doc2 " + document2.hashCode());
  }

  // for (int i = 0; i < numDocs2; i++) {
  // System.out.println(reader2.document(i));
  // }

  // long fileLength = directory.fileLength("_0.fdt");

  for (String name : directory.listAll()) {
    if (name.endsWith(".fdt")) {
      System.out.println(name);
      System.out.println(directory.fileLength(name));
    }
  }

}
 

@Test
public void testSmallDocs() throws IOException {

  RAMDirectory directory = new RAMDirectory();
  IndexWriterConfig conf1 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf1.setCodec(new Blur024Codec());
  Random random1 = new Random(1);
  IndexWriter writer1 = new IndexWriter(directory, conf1);
  for (int i = 0; i < 1000; i++) {
    writer1.addDocument(getSmallDoc(random1));
  }
  writer1.close();

  DirectoryReader reader1 = DirectoryReader.open(directory);
  int numDocs1 = reader1.numDocs();
  assertEquals(1000, numDocs1);

  // for (int i = 0; i < numDocs1; i++) {
  // System.out.println(reader1.document(i));
  // }

  IndexWriterConfig conf2 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf2.setCodec(new Blur024Codec(1 << 16, CompressionMode.HIGH_COMPRESSION));
  Random random2 = new Random(1);
  IndexWriter writer2 = new IndexWriter(directory, conf2);
  for (int i = 0; i < 1000; i++) {
    writer2.addDocument(getSmallDoc(random2));
  }
  writer2.close();

  DirectoryReader reader2 = DirectoryReader.open(directory);
  int numDocs2 = reader2.numDocs();
  assertEquals(2000, numDocs2);

  for (int i = 0; i < 2; i++) {

    long t1 = System.nanoTime();
    long hash1 = 0;
    long hash2 = 0;
    for (int d = 0; d < 1000; d++) {
      Document document1 = reader1.document(d);
      hash1 += document1.hashCode();
    }
    long t2 = System.nanoTime();
    for (int d = 0; d < 1000; d++) {
      Document document2 = reader2.document(d + 1000);
      hash2 += document2.hashCode();
    }
    long t3 = System.nanoTime();

    System.out.println((t3 - t2) / 1000000.0);
    System.out.println((t2 - t1) / 1000000.0);

    System.out.println("doc1 " + hash1);
    System.out.println("doc2 " + hash2);
  }

  // for (int i = 0; i < numDocs2; i++) {
  // System.out.println(reader2.document(i));
  // }

  // long fileLength = directory.fileLength("_0.fdt");

  for (String name : directory.listAll()) {
    if (name.endsWith(".fdt")) {
      System.out.println(name);
      System.out.println(directory.fileLength(name));
    }
  }
}
 

@Test
public void testLargeDocs() throws IOException {
  Random random = new Random();
  Iterable<? extends IndexableField> doc = getLargeDoc(random);
  RAMDirectory directory = new RAMDirectory();
  IndexWriterConfig conf1 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf1.setCodec(new Blur022Codec());
  IndexWriter writer1 = new IndexWriter(directory, conf1);
  writer1.addDocument(doc);
  writer1.close();

  DirectoryReader reader1 = DirectoryReader.open(directory);
  int numDocs1 = reader1.numDocs();
  assertEquals(1, numDocs1);

  // for (int i = 0; i < numDocs1; i++) {
  // System.out.println(reader1.document(i));
  // }

  IndexWriterConfig conf2 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf2.setCodec(new Blur022Codec(1 << 16, CompressionMode.HIGH_COMPRESSION));
  IndexWriter writer2 = new IndexWriter(directory, conf2);
  writer2.addDocument(doc);
  writer2.close();

  DirectoryReader reader2 = DirectoryReader.open(directory);
  int numDocs2 = reader2.numDocs();
  assertEquals(2, numDocs2);

  for (int i = 0; i < 2; i++) {

    long t1 = System.nanoTime();
    Document document1 = reader1.document(0);
    long t2 = System.nanoTime();
    Document document2 = reader2.document(1);
    long t3 = System.nanoTime();

    System.out.println((t3 - t2) / 1000000.0);
    System.out.println((t2 - t1) / 1000000.0);

    System.out.println("doc1 " + document1.hashCode());
    System.out.println("doc2 " + document2.hashCode());
  }

  // for (int i = 0; i < numDocs2; i++) {
  // System.out.println(reader2.document(i));
  // }

  // long fileLength = directory.fileLength("_0.fdt");

  for (String name : directory.listAll()) {
    if (name.endsWith(".fdt")) {
      System.out.println(name);
      System.out.println(directory.fileLength(name));
    }
  }

}
 

@Test
public void testSmallDocs() throws IOException {

  RAMDirectory directory = new RAMDirectory();
  IndexWriterConfig conf1 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf1.setCodec(new Blur022Codec());
  Random random1 = new Random(1);
  IndexWriter writer1 = new IndexWriter(directory, conf1);
  for (int i = 0; i < 1000; i++) {
    writer1.addDocument(getSmallDoc(random1));
  }
  writer1.close();

  DirectoryReader reader1 = DirectoryReader.open(directory);
  int numDocs1 = reader1.numDocs();
  assertEquals(1000, numDocs1);

  // for (int i = 0; i < numDocs1; i++) {
  // System.out.println(reader1.document(i));
  // }

  IndexWriterConfig conf2 = new IndexWriterConfig(Version.LUCENE_43, new WhitespaceAnalyzer(Version.LUCENE_43));
  conf2.setCodec(new Blur022Codec(1 << 16, CompressionMode.HIGH_COMPRESSION));
  Random random2 = new Random(1);
  IndexWriter writer2 = new IndexWriter(directory, conf2);
  for (int i = 0; i < 1000; i++) {
    writer2.addDocument(getSmallDoc(random2));
  }
  writer2.close();

  DirectoryReader reader2 = DirectoryReader.open(directory);
  int numDocs2 = reader2.numDocs();
  assertEquals(2000, numDocs2);

  for (int i = 0; i < 2; i++) {

    long t1 = System.nanoTime();
    long hash1 = 0;
    long hash2 = 0;
    for (int d = 0; d < 1000; d++) {
      Document document1 = reader1.document(d);
      hash1 += document1.hashCode();
    }
    long t2 = System.nanoTime();
    for (int d = 0; d < 1000; d++) {
      Document document2 = reader2.document(d + 1000);
      hash2 += document2.hashCode();
    }
    long t3 = System.nanoTime();

    System.out.println((t3 - t2) / 1000000.0);
    System.out.println((t2 - t1) / 1000000.0);

    System.out.println("doc1 " + hash1);
    System.out.println("doc2 " + hash2);
  }

  // for (int i = 0; i < numDocs2; i++) {
  // System.out.println(reader2.document(i));
  // }

  // long fileLength = directory.fileLength("_0.fdt");

  for (String name : directory.listAll()) {
    if (name.endsWith(".fdt")) {
      System.out.println(name);
      System.out.println(directory.fileLength(name));
    }
  }
}
 

@Test
public void testBlurOutputFormatOverFlowMultipleReducersTest() throws IOException, InterruptedException,
    ClassNotFoundException {
  Path input = getInDir();
  Path output = getOutDir();
  _fileSystem.delete(input, true);
  _fileSystem.delete(output, true);
  // 1500 * 50 = 75,000
  writeRecordsFile(new Path(input, "part1"), 1, 50, 1, 1500, "cf1");
  // 100 * 50 = 5,000
  writeRecordsFile(new Path(input, "part2"), 1, 50, 2000, 100, "cf1");

  Job job = Job.getInstance(_conf, "blur index");
  job.setJarByClass(BlurOutputFormatTest.class);
  job.setMapperClass(CsvBlurMapper.class);
  job.setInputFormatClass(TextInputFormat.class);

  FileInputFormat.addInputPath(job, input);
  CsvBlurMapper.addColumns(job, "cf1", "col");

  Path tablePath = new Path(new Path(_root, "table"), "test");

  TableDescriptor tableDescriptor = new TableDescriptor();
  tableDescriptor.setShardCount(2);
  tableDescriptor.setTableUri(tablePath.toString());
  tableDescriptor.setName("test");

  createShardDirectories(output, 2);

  BlurOutputFormat.setupJob(job, tableDescriptor);
  BlurOutputFormat.setOutputPath(job, output);
  BlurOutputFormat.setIndexLocally(job, false);
  BlurOutputFormat.setDocumentBufferStrategy(job, DocumentBufferStrategyHeapSize.class);
  BlurOutputFormat.setMaxDocumentBufferHeapSize(job, 128 * 1024);

  assertTrue(job.waitForCompletion(true));
  Counters ctrs = job.getCounters();
  System.out.println("Counters: " + ctrs);

  long total = 0;
  for (int i = 0; i < tableDescriptor.getShardCount(); i++) {
    Path path = new Path(output, ShardUtil.getShardName(i));
    Collection<Path> commitedTasks = getCommitedTasks(path);
    assertEquals(1, commitedTasks.size());

    DirectoryReader reader = DirectoryReader.open(new HdfsDirectory(_conf, commitedTasks.iterator().next()));
    total += reader.numDocs();
    reader.close();
  }
  assertEquals(80000, total);

}
 

@Test
public void testBlurOutputFormatOverFlowMultipleReducersWithReduceMultiplierTest() throws IOException,
    InterruptedException, ClassNotFoundException {
  Path input = getInDir();
  Path output = getOutDir();
  _fileSystem.delete(input, true);
  _fileSystem.delete(output, true);

  // 1500 * 50 = 75,000
  writeRecordsFile(new Path(input, "part1"), 1, 50, 1, 1500, "cf1");
  // 100 * 50 = 5,000
  writeRecordsFile(new Path(input, "part2"), 1, 50, 2000, 100, "cf1");

  Job job = Job.getInstance(_conf, "blur index");
  job.setJarByClass(BlurOutputFormatTest.class);
  job.setMapperClass(CsvBlurMapper.class);
  job.setInputFormatClass(TextInputFormat.class);

  FileInputFormat.addInputPath(job, input);
  CsvBlurMapper.addColumns(job, "cf1", "col");

  Path tablePath = new Path(new Path(_root, "table"), "test");

  TableDescriptor tableDescriptor = new TableDescriptor();
  tableDescriptor.setShardCount(7);
  tableDescriptor.setTableUri(tablePath.toString());
  tableDescriptor.setName("test");

  createShardDirectories(output, 7);

  BlurOutputFormat.setupJob(job, tableDescriptor);
  BlurOutputFormat.setOutputPath(job, output);
  int multiple = 2;
  BlurOutputFormat.setReducerMultiplier(job, multiple);

  assertTrue(job.waitForCompletion(true));
  Counters ctrs = job.getCounters();
  System.out.println("Counters: " + ctrs);

  long total = 0;
  for (int i = 0; i < tableDescriptor.getShardCount(); i++) {
    Path path = new Path(output, ShardUtil.getShardName(i));
    Collection<Path> commitedTasks = getCommitedTasks(path);
    assertTrue(commitedTasks.size() >= multiple);
    for (Path p : commitedTasks) {
      DirectoryReader reader = DirectoryReader.open(new HdfsDirectory(_conf, p));
      total += reader.numDocs();
      reader.close();
    }
  }
  assertEquals(80000, total);

}