有关Solr中SolrIndexSearcher的search和getDocSet的差别

  最近项目中碰到问题，发现在调用SolrIndexSearcher的
public TopFieldDocs search(Query query, Filter filter, int n,
                             Sort sort) throws IOException;
和public DocSet getDocSet(Query query) throws IOException；
效果差别比较大；
   查阅了代码发现在SolrIndexSearcher.getDocSet(Query query)中第一次搜索query时，内部实现是调用 getDocSetNC(Query query, DocSet filter)；非第一次的话会直接从cache中获取，即
   if (filterCache != null) {
      DocSet absAnswer = filterCache.get(absQ);
      if (absAnswer!=null) {
        if (positive) return absAnswer;
        else return getPositiveDocSet(matchAllDocsQuery).andNot(absAnswer);
      }
    }
   DocSet absAnswer = getDocSetNC(absQ, null);
    DocSet answer = positive ? absAnswer :                                getPositiveDocSet(matchAllDocsQuery).andNot(absAnswer);

    if (filterCache != null) {
      // cache negative queries as positive
      filterCache.put(absQ, absAnswer);
    }
    继续查阅方法getDocSetNC(Query query, DocSet filter)可以发现当filter不存在。且query为TermQuery时，实现如下：
    if (query instanceof TermQuery) {
        Term t = ((TermQuery)query).getTerm();
        SolrIndexReader[] readers = reader.getLeafReaders();
        int[] offsets = reader.getLeafOffsets();
        int[] arr = new int[256];
        int[] freq = new int[256];
        for (int i=0; i<readers.length; i++) {
          SolrIndexReader sir = readers[i];
          int offset = offsets[i];
          collector.setNextReader(sir, offset);
          TermDocs tdocs = sir.termDocs(t);
          for(;;) {
            int num = tdocs.read(arr, freq);
            if (num==0) break;
            for (int j=0; j<num; j++) {
              collector.collect(arr[j]);
            }
          }
          tdocs.close();
        }
    其实情况则直接调用lucene的super.search(query, luceneFilter, collector);

而SolrIndexSearcher.search(query,filter,n,sort)则是直接调用lucene的同名方法；