org.apache.lucene.index.ReadersAndUpdates Maven / Gradle / Ivy

Show more of this group Show more artifacts with this name
Show all versions of infinispan-embedded-query
Infinispan Embedded Query All-in-One module
There is a newer version: 9.1.7.Final
package org.apache.lucene.index;

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

import java.io.IOException;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;
import java.util.Map.Entry;
import java.util.NoSuchElementException;
import java.util.Set;
import java.util.concurrent.atomic.AtomicInteger;

import org.apache.lucene.codecs.Codec;
import org.apache.lucene.codecs.DocValuesConsumer;
import org.apache.lucene.codecs.DocValuesFormat;
import org.apache.lucene.codecs.LiveDocsFormat;
import org.apache.lucene.document.BinaryDocValuesField;
import org.apache.lucene.document.NumericDocValuesField;
import org.apache.lucene.store.Directory;
import org.apache.lucene.store.IOContext;
import org.apache.lucene.store.TrackingDirectoryWrapper;
import org.apache.lucene.util.Bits;
import org.apache.lucene.util.BytesRef;
import org.apache.lucene.util.IOUtils;
import org.apache.lucene.util.MutableBits;

// Used by IndexWriter to hold open SegmentReaders (for
// searching or merging), plus pending deletes and updates,
// for a given segment
class ReadersAndUpdates {
  // Not final because we replace (clone) when we need to
  // change it and it's been shared:
  public final SegmentCommitInfo info;

  // Tracks how many consumers are using this instance:
  private final AtomicInteger refCount = new AtomicInteger(1);

  private final IndexWriter writer;

  // Set once (null, and then maybe set, and never set again):
  private SegmentReader reader;

  // TODO: it's sometimes wasteful that we hold open two
  // separate SRs (one for merging one for
  // reading)... maybe just use a single SR?  The gains of
  // not loading the terms index (for merging in the
  // non-NRT case) are far less now... and if the app has
  // any deletes it'll open real readers anyway.

  // Set once (null, and then maybe set, and never set again):
  private SegmentReader mergeReader;

  // Holds the current shared (readable and writable)
  // liveDocs.  This is null when there are no deleted
  // docs, and it's copy-on-write (cloned whenever we need
  // to change it but it's been shared to an external NRT
  // reader).
  private Bits liveDocs;

  // How many further deletions we've done against
  // liveDocs vs when we loaded it or last wrote it:
  private int pendingDeleteCount;

  // True if the current liveDocs is referenced by an
  // external NRT reader:
  private boolean liveDocsShared;

  // Indicates whether this segment is currently being merged. While a segment
  // is merging, all field updates are also registered in the
  // mergingNumericUpdates map. Also, calls to writeFieldUpdates merge the 
  // updates with mergingNumericUpdates.
  // That way, when the segment is done merging, IndexWriter can apply the
  // updates on the merged segment too.
  private boolean isMerging = false;
  
  private final Map mergingDVUpdates = new HashMap<>();
  
  public ReadersAndUpdates(IndexWriter writer, SegmentCommitInfo info) {
    this.info = info;
    this.writer = writer;
    liveDocsShared = true;
  }

  public void incRef() {
    final int rc = refCount.incrementAndGet();
    assert rc > 1;
  }

  public void decRef() {
    final int rc = refCount.decrementAndGet();
    assert rc >= 0;
  }

  public int refCount() {
    final int rc = refCount.get();
    assert rc >= 0;
    return rc;
  }

  public synchronized int getPendingDeleteCount() {
    return pendingDeleteCount;
  }
  
  // Call only from assert!
  public synchronized boolean verifyDocCounts() {
    int count;
    if (liveDocs != null) {
      count = 0;
      for(int docID=0;docID e : dvUpdates.numericDVUpdates.entrySet()) {
        DocValuesFieldUpdates updates = mergingDVUpdates.get(e.getKey());
        if (updates == null) {
          mergingDVUpdates.put(e.getKey(), e.getValue());
        } else {
          updates.merge(e.getValue());
        }
      }
      for (Entry e : dvUpdates.binaryDVUpdates.entrySet()) {
        DocValuesFieldUpdates updates = mergingDVUpdates.get(e.getKey());
        if (updates == null) {
          mergingDVUpdates.put(e.getKey(), e.getValue());
        } else {
          updates.merge(e.getValue());
        }
      }
    }
    
    // create a new map, keeping only the gens that are in use
    Map> genUpdatesFiles = info.getUpdatesFiles();
    Map> newGenUpdatesFiles = new HashMap<>();
    final long fieldInfosGen = info.getFieldInfosGen();
    for (FieldInfo fi : fieldInfos) {
      long dvGen = fi.getDocValuesGen();
      if (dvGen != -1 && !newGenUpdatesFiles.containsKey(dvGen)) {
        if (dvGen == fieldInfosGen) {
          newGenUpdatesFiles.put(fieldInfosGen, trackingDir.getCreatedFiles());
        } else {
          newGenUpdatesFiles.put(dvGen, genUpdatesFiles.get(dvGen));
        }
      }
    }
    
    info.setGenUpdatesFiles(newGenUpdatesFiles);
    
    // wrote new files, should checkpoint()
    writer.checkpoint();

    // if there is a reader open, reopen it to reflect the updates
    if (reader != null) {
      SegmentReader newReader = new SegmentReader(info, reader, liveDocs, info.info.getDocCount() - info.getDelCount() - pendingDeleteCount);
      boolean reopened = false;
      try {
        reader.decRef();
        reader = newReader;
        reopened = true;
      } finally {
        if (!reopened) {
          newReader.decRef();
        }
      }
    }
  }

  /**
   * Returns a reader for merge. This method applies field updates if there are
   * any and marks that this segment is currently merging.
   */
  synchronized SegmentReader getReaderForMerge(IOContext context) throws IOException {
    assert Thread.holdsLock(writer);
    // must execute these two statements as atomic operation, otherwise we
    // could lose updates if e.g. another thread calls writeFieldUpdates in
    // between, or the updates are applied to the obtained reader, but then
    // re-applied in IW.commitMergedDeletes (unnecessary work and potential
    // bugs).
    isMerging = true;
    return getReader(context);
  }
  
  /**
   * Drops all merging updates. Called from IndexWriter after this segment
   * finished merging (whether successfully or not).
   */
  public synchronized void dropMergingUpdates() {
    mergingDVUpdates.clear();
    isMerging = false;
  }
  
  /** Returns updates that came in while this segment was merging. */
  public synchronized Map getMergingFieldUpdates() {
    return mergingDVUpdates;
  }
  
  @Override
  public String toString() {
    StringBuilder sb = new StringBuilder();
    sb.append("ReadersAndLiveDocs(seg=").append(info);
    sb.append(" pendingDeleteCount=").append(pendingDeleteCount);
    sb.append(" liveDocsShared=").append(liveDocsShared);
    return sb.toString();
  }
  
}