Substantially improved the performance of DateFilter by adding the

ability to reuse TermDocs objects.


git-svn-id: https://svn.apache.org/repos/asf/lucene/java/trunk@149642 13f79535-47bb-0310-9956-ffa450edef68

src/java/org/apache/lucene/index/IndexReader.java

@@ -193,7 +193,7 @@ abstract public class IndexReader {
   abstract public int docFreq(Term t) throws IOException;
 
   /** Returns an enumeration of all the documents which contain
-    <code>Term</code>. For each document, the document number, the frequency of
+    <code>term</code>. For each document, the document number, the frequency of
     the term in that document is also provided, for use in search scoring.
     Thus, this method implements the mapping:
     <p><ul>
@@ -201,10 +201,17 @@ abstract public class IndexReader {
     </ul>
     <p>The enumeration is ordered by document number.  Each document number
     is greater than all that precede it in the enumeration. */
-  abstract public TermDocs termDocs(Term t) throws IOException;
+  public TermDocs termDocs(Term term) throws IOException {
+    TermDocs termDocs = termDocs();
+    termDocs.seek(term);
+    return termDocs;
+  }
+
+  /** Returns an unpositioned {@link TermDocs} enumerator. */
+  abstract public TermDocs termDocs() throws IOException;
 
   /** Returns an enumeration of all the documents which contain
-    <code>Term</code>.  For each document, in addition to the document number
+    <code>term</code>.  For each document, in addition to the document number
     and frequency of the term in that document, a list of all of the ordinal
     positions of the term in the document is available.  Thus, this method
     implements the mapping:
@@ -218,7 +225,14 @@ abstract public class IndexReader {
     <p> This positional information faciliates phrase and proximity searching.
     <p>The enumeration is ordered by document number.  Each document number is
     greater than all that precede it in the enumeration. */
-  abstract public TermPositions termPositions(Term t) throws IOException;
+  public TermPositions termPositions(Term term) throws IOException {
+    TermPositions termPositions = termPositions();
+    termPositions.seek(term);
+    return termPositions;
+  }
+
+  /** Returns an unpositioned {@link TermPositions} enumerator. */
+  abstract public TermPositions termPositions() throws IOException;
 
   /** Deletes the document numbered <code>docNum</code>.  Once a document is
     deleted it will not appear in TermDocs or TermPostitions enumerations.

src/java/org/apache/lucene/index/SegmentReader.java

@@ -78,8 +78,8 @@ final class SegmentReader extends IndexReader {
   BitVector deletedDocs = null;
   private boolean deletedDocsDirty = false;
 
-  private InputStream freqStream;
-  private InputStream proxStream;
+  InputStream freqStream;
+  InputStream proxStream;
 
 
   private static class Norm {
@@ -194,28 +194,12 @@ final class SegmentReader extends IndexReader {
     return (deletedDocs != null && deletedDocs.get(n));
   }
 
-  public final TermDocs termDocs(Term t) throws IOException {
-    TermInfo ti = tis.get(t);
-    if (ti != null)
-      return new SegmentTermDocs(this, ti);
-    else
-      return null;
+  public final TermDocs termDocs() throws IOException {
+    return new SegmentTermDocs(this);
   }
 
-  final InputStream getFreqStream () {
-    return (InputStream)freqStream.clone();
-  }
-
-  public final TermPositions termPositions(Term t) throws IOException {
-    TermInfo ti = tis.get(t);
-    if (ti != null)
-      return new SegmentTermPositions(this, ti);
-    else
-      return null;
-  }
-
-  final InputStream getProxStream () {
-    return (InputStream)proxStream.clone();
+  public final TermPositions termPositions() throws IOException {
+    return new SegmentTermPositions(this);
   }
 
   public final int docFreq(Term t) throws IOException {

src/java/org/apache/lucene/index/SegmentTermDocs.java

@@ -66,22 +66,27 @@ class SegmentTermDocs implements TermDocs {
   int doc = 0;
   int freq;
 
-  SegmentTermDocs(SegmentReader p) throws IOException {
-    parent = p;
-    freqStream = parent.getFreqStream();
-    deletedDocs = parent.deletedDocs;
+  SegmentTermDocs(SegmentReader parent)
+    throws IOException {
+    this.parent = parent;
+    this.freqStream = (InputStream)parent.freqStream.clone();
+    this.deletedDocs = parent.deletedDocs;
   }
   
-  SegmentTermDocs(SegmentReader p, TermInfo ti) throws IOException {
-    this(p);
+  public void seek(Term term) throws IOException {
+    TermInfo ti = parent.tis.get(term);
     seek(ti);
   }
   
   void seek(TermInfo ti) throws IOException {
+    if (ti == null) {
+      freqCount = 0;
+    } else {
       freqCount = ti.docFreq;
       doc = 0;
       freqStream.seek(ti.freqPointer);
     }
+  }
   
   public void close() throws IOException {
     freqStream.close();

src/java/org/apache/lucene/index/SegmentTermPositions.java

@@ -66,13 +66,7 @@ extends SegmentTermDocs implements TermPositions {
   
   SegmentTermPositions(SegmentReader p) throws IOException {
     super(p);
-    proxStream = parent.getProxStream();
-  }
-
-  SegmentTermPositions(SegmentReader p, TermInfo ti)
-       throws IOException {
-    this(p);
-    seek(ti);
+    this.proxStream = (InputStream)parent.proxStream.clone();
   }
 
   final void seek(TermInfo ti) throws IOException {

src/java/org/apache/lucene/index/SegmentsReader.java

@@ -151,12 +151,12 @@ final class SegmentsReader extends IndexReader {
     return total;
   }
 
-  public final TermDocs termDocs(Term term) throws IOException {
-    return new SegmentsTermDocs(readers, starts, term);
+  public final TermDocs termDocs() throws IOException {
+    return new SegmentsTermDocs(readers, starts);
   }
 
-  public final TermPositions termPositions(Term term) throws IOException {
-    return new SegmentsTermPositions(readers, starts, term);
+  public final TermPositions termPositions() throws IOException {
+    return new SegmentsTermPositions(readers, starts);
   }
 
   public final void close() throws IOException {
@@ -240,13 +240,15 @@ class SegmentsTermDocs implements TermDocs {
   protected int base = 0;
   protected int pointer = 0;
 
-  SegmentsTermDocs(SegmentReader[] r, int[] s, Term t) {
+  private SegmentTermDocs[] segTermDocs;
+  protected SegmentTermDocs current;              // == segTermDocs[pointer]
+  
+  SegmentsTermDocs(SegmentReader[] r, int[] s) {
     readers = r;
     starts = s;
-    term = t;
-  }
 
-  protected SegmentTermDocs current;
+    segTermDocs = new SegmentTermDocs[r.length];
+  }
 
   public final int doc() {
     return base + current.doc;
@@ -255,14 +257,19 @@ class SegmentsTermDocs implements TermDocs {
     return current.freq;
   }
 
+  public final void seek(Term term) {
+    this.term = term;
+    this.base = 0;
+    this.pointer = 0;
+    this.current = null;
+  }
+
   public final boolean next() throws IOException {
     if (current != null && current.next()) {
       return true;
     } else if (pointer < readers.length) {
-      if (current != null)
-	current.close();
       base = starts[pointer];
-      current = termDocs(readers[pointer++]);
+      current = termDocs(pointer++);
       return next();
     } else
       return false;
@@ -275,14 +282,13 @@ class SegmentsTermDocs implements TermDocs {
       while (current == null) {
 	if (pointer < readers.length) {		  // try next segment
 	  base = starts[pointer];
-	  current = termDocs(readers[pointer++]);
+	  current = termDocs(pointer++);
 	} else {
 	  return 0;
 	}
       }
       int end = current.read(docs, freqs);
       if (end == 0) {				  // none left in segment
-	current.close();
 	current = null;
       } else {					  // got some
 	final int b = base;			  // adjust doc numbers
@@ -302,25 +308,37 @@ class SegmentsTermDocs implements TermDocs {
     return true;
   }
 
+  private SegmentTermDocs termDocs(int i) throws IOException {
+    if (term == null)
+      return null;
+    SegmentTermDocs result = segTermDocs[i];
+    if (result == null)
+      result = segTermDocs[i] = termDocs(readers[i]);
+    result.seek(term);
+    return result;
+  }
+
   protected SegmentTermDocs termDocs(SegmentReader reader)
     throws IOException {
-    return (SegmentTermDocs)reader.termDocs(term);
+    return (SegmentTermDocs)reader.termDocs();
   }
 
   public final void close() throws IOException {
-    if (current != null)
-      current.close();
+    for (int i = 0; i < segTermDocs.length; i++) {
+      if (segTermDocs[i] != null)
+        segTermDocs[i].close();
+    }
   }
 }
 
 class SegmentsTermPositions extends SegmentsTermDocs implements TermPositions {
-  SegmentsTermPositions(SegmentReader[] r, int[] s, Term t) {
-    super(r,s,t);
+  SegmentsTermPositions(SegmentReader[] r, int[] s) {
+    super(r,s);
   }
 
   protected final SegmentTermDocs termDocs(SegmentReader reader)
        throws IOException {
-    return (SegmentTermDocs)reader.termPositions(term);
+    return (SegmentTermDocs)reader.termPositions();
   }
 
   public final int nextPosition() throws IOException {

src/java/org/apache/lucene/index/TermDocs.java

@@ -67,6 +67,11 @@ import org.apache.lucene.document.Document;
   */
 
 public interface TermDocs {
+  /** Sets this to the data for a term.
+   * The enumeration is reset to the start of the data for this term.
+   */
+  void seek(Term term) throws IOException;
+
   /** Returns the current document number.  <p> This is invalid until {@link
       #next()} is called for the first time.*/
   int doc();
@@ -91,7 +96,7 @@ public interface TermDocs {
   /** Skips entries to the first beyond the current whose document number is
    * greater than or equal to <i>target</i>. <p>Returns true iff there is such
    * an entry.  <p>Behaves as if written: <pre>
-   *   public boolean skipTo(int target) {
+   *   boolean skipTo(int target) {
    *     do {
    *       if (!next())
    * 	     return false;

src/java/org/apache/lucene/index/TermEnum.java

@@ -65,12 +65,10 @@ public abstract class TermEnum {
   /** Increments the enumeration to the next element.  True if one exists.*/
   abstract public boolean next() throws IOException;
 
-  /** Returns the current Term in the enumeration.
-    Initially invalid, valid after next() called for the first time.*/
+  /** Returns the current Term in the enumeration.*/
   abstract public Term term();
 
-  /** Returns the docFreq of the current Term in the enumeration.
-    Initially invalid, valid after next() called for the first time.*/
+  /** Returns the docFreq of the current Term in the enumeration.*/
   abstract public int docFreq();
 
   /** Closes the enumeration to further activity, freeing resources. */

src/java/org/apache/lucene/search/DateFilter.java

@@ -129,10 +129,11 @@ public final class DateFilter extends Filter {
   final public BitSet bits(IndexReader reader) throws IOException {
     BitSet bits = new BitSet(reader.maxDoc());
     TermEnum enum = reader.terms(new Term(field, start));
+    TermDocs termDocs = reader.termDocs();
     try {
       Term stop = new Term(field, end);
       while (enum.term().compareTo(stop) <= 0) {
-	TermDocs termDocs = reader.termDocs(enum.term());
+	termDocs.seek(enum.term());
 	try {
 	  while (termDocs.next())
 	    bits.set(termDocs.doc());