DocSet.andNot(),andNotSize()

git-svn-id: https://svn.apache.org/repos/asf/incubator/solr/trunk@413399 13f79535-47bb-0310-9956-ffa450edef68

CHANGES.txt

@@ -17,6 +17,7 @@ New Features
  9. Added KeywordTokenizerFactory (hossman)
 10. copyField accepts dynamicfield-like names as the source.
     (Darren Erik Vengroff via yonik, SOLR-21)
+11. new DocSet.andNot(), DocSet.andNotSize() (yonik)
      
 Changes in runtime behavior
  1. classes reorganized into different packages, package names changed to Apache

src/java/org/apache/solr/search/BitDocSet.java

@@ -143,6 +143,7 @@ public class BitDocSet extends DocSetBase {
     return bits.get(doc);
   }
 
+  @Override
   public int intersectionSize(DocSet other) {
     if (other instanceof BitDocSet) {
       return (int)OpenBitSet.intersectionCount(this.bits, ((BitDocSet)other).bits);
@@ -152,6 +153,7 @@ public class BitDocSet extends DocSetBase {
     }
   }
 
+  @Override
   public int unionSize(DocSet other) {
     if (other instanceof BitDocSet) {
       return (int)OpenBitSet.unionCount(this.bits, ((BitDocSet)other).bits);
@@ -161,6 +163,18 @@ public class BitDocSet extends DocSetBase {
     }
   }
 
+  @Override
+  public int andNotSize(DocSet other) {
+    if (other instanceof BitDocSet) {
+      // if we don't know our current size, this is faster than
+      // size - intersection_size
+      return (int)OpenBitSet.andNotCount(this.bits, ((BitDocSet)other).bits);
+    } else {
+      // use BaseDocSet's size-intersection_size
+      return super.andNotSize(other);
+    }
+  }
+
   public long memSize() {
     return (bits.getBits().length << 3) + 16;
   }

src/java/org/apache/solr/search/DocSet.java

@@ -124,6 +124,17 @@ public interface DocSet /* extends Collection<Integer> */ {
    */
   public int unionSize(DocSet other);
 
+  /**
+   * Returns the documents in this set that are not in the other set. Neither set is modified - a new DocSet is
+   * created and returned.
+   * @return a DocSet representing this AND NOT other
+   */
+  public DocSet andNot(DocSet other);
+
+  /**
+   * Returns the number of documents in this set that are not in the other set.
+   */
+  public int andNotSize(DocSet other);
 }
 
 /** A base class that may be usefull for implementing DocSets */
@@ -208,11 +219,20 @@ abstract class DocSetBase implements DocSet {
     return intersection(other).size();
   }
 
-  // TODO: do an efficient implementation
+  // subclasses have more efficient implementations
   public int unionSize(DocSet other) {
     return union(other).size();
   }
 
+  public DocSet andNot(DocSet other) {
+    OpenBitSet newbits = (OpenBitSet)(this.getBits().clone());
+    newbits.andNot(other.getBits());
+    return new BitDocSet(newbits);
+  }
+
+  public int andNotSize(DocSet other) {
+    return this.size() - this.intersectionSize(other);
+  }
 }
 
 

src/java/org/apache/solr/search/HashDocSet.java

@@ -278,7 +278,7 @@ public final class HashDocSet extends DocSetBase {
   public int unionSize(DocSet other) {
    if (other instanceof HashDocSet) {
      // set "a" to the smallest doc set for the most efficient
-     // intersection.
+     // union count.
      final HashDocSet a = size()<=other.size() ? this : (HashDocSet)other;
      final HashDocSet b = size()<=other.size() ? (HashDocSet)other : this;
 
@@ -302,5 +302,6 @@ public final class HashDocSet extends DocSetBase {
    }
   }
 
-
+  // don't implement andNotSize() on purpose... if one of the sets is a HashDocSet,
+  // its easier to get the intersection size and subtract (implementation in BaseDocSet)
 }

src/test/org/apache/solr/search/DocSetPerf.java

@@ -14,7 +14,7 @@
  * limitations under the License.
  */
 
-package org.apache.solr.search.test;
+package org.apache.solr.search;
 
 import org.apache.solr.search.BitDocSet;
 import org.apache.solr.search.HashDocSet;

src/test/org/apache/solr/search/TestDocSet.java

@@ -64,14 +64,15 @@ public class TestDocSet extends TestCase {
     OpenBitSet a_and = (OpenBitSet)a1.clone(); a_and.and(a2);
     OpenBitSet a_or = (OpenBitSet)a1.clone(); a_or.or(a2);
     // OpenBitSet a_xor = (OpenBitSet)a1.clone(); a_xor.xor(a2);
-    // OpenBitSet a_andn = (OpenBitSet)a1.clone(); a_andn.andNot(a2);
+    OpenBitSet a_andn = (OpenBitSet)a1.clone(); a_andn.andNot(a2);
 
     checkEqual(a_and, b1.intersection(b2));
     checkEqual(a_or, b1.union(b2));
+    checkEqual(a_andn, b1.andNot(b2));
 
     assertEquals(a_and.cardinality(), b1.intersectionSize(b2));
     assertEquals(a_or.cardinality(), b1.unionSize(b2));
-
+    assertEquals(a_andn.cardinality(), b1.andNotSize(b2));
   }