lucene2.4源码学习8 查询 coord

lucene计算得分的时候要*coord，那这个coord是什么意思呢。就是这个document满足了多少个查询条件。如：

Document doc = new Document();  
doc.add(new Field("title", "search engine", Field.Store.YES, Field.Index.ANALYZED));  
doc.add(new Field("content", "good lucene luke lucene search server", Field.Store.YES, Field.Index.ANALYZED));  
indexWriter.addDocument(doc); 

当查询条件为title:search content:lucene的时候，
在title这个Field中（"search engine"）匹配上title:search，
在content这个Field中（"good lucene luke lucene search server"）匹配上content:lucene，也就是这个document满足所有的查询条件，coord为1.

公式是这样的：

 public float coord(int overlap, int maxOverlap) {
    return overlap / (float)maxOverlap;
  }

源码的来龙去脉是这样的。
BooleanScorer2中

  public float score() throws IOException {
    coordinator.initDoc();
    float sum = countingSumScorer.score();
    return sum * coordinator.coordFactor();
  }

float coordFactor() {
      return coordFactors[nrMatchers];
    }

 private float[] coordFactors = null;
    
    void init() { // use after all scorers have been added.
      coordFactors = new float[maxCoord + 1];
      Similarity sim = getSimilarity();
      for (int i = 0; i <= maxCoord; i++) {
        coordFactors[i] = sim.coord(i, maxCoord);
      }
    }

maxCoord就是查询语句中一共有多少个查询条件。

  public void add(final Scorer scorer, boolean required, boolean prohibited) {
    if (!prohibited) {
      coordinator.maxCoord++;
    }

可以看到每添加一个查询语句，只要不是条件非，maxCoord都会增加1.这个是总的查询条件。

那文档匹配的查询条件呢：

return new ConjunctionScorer(defaultSimilarity, requiredScorers) {
      private int lastScoredDoc = -1;

      public float score() throws IOException {
        if (this.doc() >= lastScoredDoc) {
          lastScoredDoc = this.doc();
          coordinator.nrMatchers += requiredNrMatchers;
        }

是由requiredNrMatchers决定的。

countingConjunctionSumScorer(optionalScorers);

 countingConjunctionSumScorer(requiredScorers);

optionalScorers表示的是should的条件，requiredScorers是must条件。所以满足的条件就是满足should和must的条件总和。