OpenBitSet和OpenBitSetIterator在TermRangeQuery中的运用

OpenBitSet和OpenBitSetIterator在TermRangeQuery中的运用
在MultiTermQuery 的rewrite方法中，如果    if (pendingTerms.size() >= termCountLimit || docVisitCount >= docCountCutoff)  的就会使用MultiTermQueryWrapperFilter，如果查询出来的term的总数目大于termCountLimit或者docVisitCount是 df ，如果df 大于docCountCutoff 则使用MultiTermQueryWrapperFilter，否则使用BooleanQuery，他们之间的关系是or的关系, MultiTermQueryWrapperFilter 使用OpenBitSet收集docId,使用OpenBitSetIterator还原docId

  @Override
    public Query rewrite(IndexReader reader, MultiTermQuery query) throws IOException {
      // Get the enum and start visiting terms.  If we
      // exhaust the enum before hitting either of the
      // cutoffs, we use ConstantBooleanQueryRewrite; else,
      // ConstantFilterRewrite:
      final Collection<Term> pendingTerms = new ArrayList<Term>();
      final int docCountCutoff = (int) ((docCountPercent / 100.) * reader.maxDoc());
      final int termCountLimit = Math.min(BooleanQuery.getMaxClauseCount(), termCountCutoff);
      int docVisitCount = 0;

      FilteredTermEnum enumerator = query.getEnum(reader);
      try {
        while(true) {
          Term t = enumerator.term();
          if (t != null) {
            pendingTerms.add(t);
            // Loading the TermInfo from the terms dict here
            // should not be costly, because 1) the
            // query/filter will load the TermInfo when it
            // runs, and 2) the terms dict has a cache:
            docVisitCount += reader.docFreq(t);
          }

          if (pendingTerms.size() >= termCountLimit || docVisitCount >= docCountCutoff) {
            // Too many terms -- make a filter.
            Query result = new ConstantScoreQuery(new MultiTermQueryWrapperFilter<MultiTermQuery>(query));
            result.setBoost(query.getBoost());
            return result;
          } else  if (!enumerator.next()) {
            // Enumeration is done, and we hit a small
            // enough number of terms & docs -- just make a
            // BooleanQuery, now
            BooleanQuery bq = new BooleanQuery(true);
            for (final Term term: pendingTerms) {
              TermQuery tq = new TermQuery(term);
              bq.add(tq, BooleanClause.Occur.SHOULD);
            }
            // Strip scores
            Query result = new ConstantScoreQuery(new QueryWrapperFilter(bq));
            result.setBoost(query.getBoost());
            query.incTotalNumberOfTerms(pendingTerms.size());
            return result;
          }
        }
      } finally {
        enumerator.close();
      }
    }
   
收集的docId的代码 调用如图所示


先new 一个OpenBitSet，大小是查询出来的当前segemt中最大文档的数目，然后通过
  SegmentTermDocs 的public int read(final int[] docs, final int[] freqs)
这个方法读取docId和frg，然后 通过for循环
for(int i=0;i<count;i++) {
                bitSet.set(docs[i]);
              }
把docId放到OpenBitSet里面




代码如下和注释如下
public DocIdSet getDocIdSet(IndexReader reader) throws IOException {
//返回TermRangeTermEnum对象，这个对象先用用小的那个string new一个term，然后定位到tis文件,while循环读取term信息，然后去frg文件里面读取docId，在while循环里面，通过SegmentTermDocs读取frg文件的docId和frg。
    final TermEnum enumerator = query.getEnum(reader);
    try {
      // if current term in enum is null, the enum is empty -> shortcut
      if (enumerator.term() == null)
        return DocIdSet.EMPTY_DOCIDSET;
      // else fill into a OpenBitSet
      final OpenBitSet bitSet = new OpenBitSet(reader.maxDoc());
      final int[] docs = new int[32];
      final int[] freqs = new int[32];

// new 一个SegmentTermDocs的实例，会调用它的read方法读取docId
      TermDocs termDocs = reader.termDocs();
      try {
        int termCount = 0;
        do {
          Term term = enumerator.term();
          if (term == null)
            break;
          termCount++;
SegmentTermDocs 在frg文件里面seek到term的对应的docid的开始位置
          termDocs.seek(term);
          while (true) {
       //  读取docId，一次读取到32 个docId到 docs数组里面，如果没有32个则读取实际的数目
            final int count = termDocs.read(docs, freqs);
            if (count != 0) {
              for(int i=0;i<count;i++) {
                bitSet.set(docs[i]);
              }
            } else {
              break;
            }
          }

        } while (enumerator.next());

        query.incTotalNumberOfTerms(termCount);

      } finally {
        termDocs.close();
      }
      return bitSet;
    } finally {
      enumerator.close();
    }
  }



enumerator.next()方法截图如下，enumerator是TermRangeTermEnum，会调用父类的FilteredTermEnum next方法。



FilteredTermEnum的next方法如下，他会调用actualEnum读取tis文件里面的下一个term，然后调用termCompare 方法，termCompare 这个方法是抽象方法，留给子类实现，
TermRangeTermEnum方法的实现逻辑是和右边的区间的term做一个比较，看查询的term是否超出区间
public boolean next() throws IOException {
        if (actualEnum == null) return false; // the actual enumerator is not initialized!
        currentTerm = null;
        while (currentTerm == null) {
            if (endEnum()) return false;
            if (actualEnum.next()) {
                Term term = actualEnum.term();
                if (termCompare(term)) {
                    currentTerm = term;
                    return true;
                }
            }
            else return false;
        }
        currentTerm = null;
        return false;
}

还原是在ConstantScorer的nextDoc方法调用的如下图


public int nextDoc() throws IOException {
      return docIdSetIterator.nextDoc();
    }