Skip docs with Docvalues in NumericLeafComparator (#12405)

* Skip document by docValues

*When the queue is full with only one Comparator, we could better tune the maxValueAsBytes/minValueAsBytes. For instance, if the sort is ascending and bottom value is 5, we will use a range on [MIN_VALUE, 4].
---------

Co-authored-by: Adrien Grand <[email protected]>

Author: LuXugang

lucene/CHANGES.txt

@@ -262,6 +262,8 @@ Optimizations
 
 * GITHUB#11903: Faster sort on high-cardinality string fields. (Adrien Grand)
 
+* GITHUB#12381: Skip docs with DocValues in NumericLeafComparator. (Lu Xugang, Adrien Grand)
+
 Changes in runtime behavior
 ---------------------
 

lucene/core/src/java/org/apache/lucene/document/FeatureSortField.java

@@ -23,6 +23,7 @@
 import org.apache.lucene.index.Terms;
 import org.apache.lucene.index.TermsEnum;
 import org.apache.lucene.search.FieldComparator;
+import org.apache.lucene.search.Pruning;
 import org.apache.lucene.search.SimpleFieldComparator;
 import org.apache.lucene.search.SortField;
 import org.apache.lucene.util.BytesRef;
@@ -44,7 +45,7 @@ public FeatureSortField(String field, String featureName) {
   }
 
   @Override
-  public FieldComparator<?> getComparator(int numHits, boolean enableSkipping) {
+  public FieldComparator<?> getComparator(int numHits, Pruning pruning) {
     return new FeatureComparator(numHits, getField(), featureName);
   }
 

lucene/core/src/java/org/apache/lucene/document/LatLonPointSortField.java

@@ -18,6 +18,7 @@
 
 import org.apache.lucene.geo.GeoUtils;
 import org.apache.lucene.search.FieldComparator;
+import org.apache.lucene.search.Pruning;
 import org.apache.lucene.search.SortField;
 
 /** Sorts by distance from an origin location. */
@@ -38,7 +39,7 @@ final class LatLonPointSortField extends SortField {
   }
 
   @Override
-  public FieldComparator<?> getComparator(int numHits, boolean enableSkipping) {
+  public FieldComparator<?> getComparator(int numHits, Pruning pruning) {
     return new LatLonPointDistanceComparator(getField(), latitude, longitude, numHits);
   }
 

lucene/core/src/java/org/apache/lucene/document/XYPointSortField.java

@@ -17,6 +17,7 @@
 package org.apache.lucene.document;
 
 import org.apache.lucene.search.FieldComparator;
+import org.apache.lucene.search.Pruning;
 import org.apache.lucene.search.SortField;
 
 /** Sorts by distance from an origin location. */
@@ -35,7 +36,7 @@ final class XYPointSortField extends SortField {
   }
 
   @Override
-  public FieldComparator<?> getComparator(int numHits, boolean enableSkipping) {
+  public FieldComparator<?> getComparator(int numHits, Pruning pruning) {
     return new XYPointDistanceComparator(getField(), x, y, numHits);
   }
 

lucene/core/src/java/org/apache/lucene/index/CheckIndex.java

@@ -60,6 +60,7 @@
 import org.apache.lucene.search.FieldExistsQuery;
 import org.apache.lucene.search.KnnCollector;
 import org.apache.lucene.search.LeafFieldComparator;
+import org.apache.lucene.search.Pruning;
 import org.apache.lucene.search.Sort;
 import org.apache.lucene.search.SortField;
 import org.apache.lucene.search.TopDocs;
@@ -1119,7 +1120,7 @@ public static Status.IndexSortStatus testSort(
 
       for (int i = 0; i < fields.length; i++) {
         reverseMul[i] = fields[i].getReverse() ? -1 : 1;
-        comparators[i] = fields[i].getComparator(1, false).getLeafComparator(readerContext);
+        comparators[i] = fields[i].getComparator(1, Pruning.NONE).getLeafComparator(readerContext);
       }
 
       int maxDoc = reader.maxDoc();

lucene/core/src/java/org/apache/lucene/search/DoubleValuesSource.java

@@ -528,8 +528,8 @@ void setMissingValue(double missingValue) {
 
     @Override
     public FieldComparator<Double> newComparator(
-        String fieldname, int numHits, boolean enableSkipping, boolean reversed) {
-      return new DoubleComparator(numHits, fieldname, missingValue, reversed, false) {
+        String fieldname, int numHits, Pruning pruning, boolean reversed) {
+      return new DoubleComparator(numHits, fieldname, missingValue, reversed, Pruning.NONE) {
         @Override
         public LeafFieldComparator getLeafComparator(LeafReaderContext context) throws IOException {
           DoubleValuesHolder holder = new DoubleValuesHolder();

lucene/core/src/java/org/apache/lucene/search/FieldComparatorSource.java

@@ -30,5 +30,5 @@ public abstract class FieldComparatorSource {
    * @return FieldComparator.
    */
   public abstract FieldComparator<?> newComparator(
-      String fieldname, int numHits, boolean enableSkipping, boolean reversed);
+      String fieldname, int numHits, Pruning pruning, boolean reversed);
 }

lucene/core/src/java/org/apache/lucene/search/FieldValueHitQueue.java

@@ -125,7 +125,7 @@ private FieldValueHitQueue(SortField[] fields, int size) {
     // need to check it again.
 
     // All these are required by this class's API - need to return arrays.
-    // Therefore even in the case of a single comparator, create an array
+    // Therefore, even in the case of a single comparator, create an array
     // anyway.
     this.fields = fields;
     int numComparators = fields.length;
@@ -134,7 +134,12 @@ private FieldValueHitQueue(SortField[] fields, int size) {
     for (int i = 0; i < numComparators; ++i) {
       SortField field = fields[i];
       reverseMul[i] = field.reverse ? -1 : 1;
-      comparators[i] = field.getComparator(size, i == 0);
+      comparators[i] =
+          field.getComparator(
+              size,
+              i == 0
+                  ? (numComparators > 1 ? Pruning.GREATER_THAN : Pruning.GREATER_THAN_OR_EQUAL_TO)
+                  : Pruning.NONE);
     }
   }
 

lucene/core/src/java/org/apache/lucene/search/IndexSortSortedNumericDocValuesRangeQuery.java

@@ -608,7 +608,7 @@ private static ValueComparator loadComparator(
       throws IOException {
     @SuppressWarnings("unchecked")
     FieldComparator<Number> fieldComparator =
-        (FieldComparator<Number>) sortField.getComparator(1, false);
+        (FieldComparator<Number>) sortField.getComparator(1, Pruning.NONE);
     if (type == Type.INT) {
       fieldComparator.setTopValue((int) topValue);
     } else {

lucene/core/src/java/org/apache/lucene/search/LongValuesSource.java

@@ -337,8 +337,8 @@ void setMissingValue(long missingValue) {
 
     @Override
     public FieldComparator<Long> newComparator(
-        String fieldname, int numHits, boolean enableSkipping, boolean reversed) {
-      return new LongComparator(numHits, fieldname, missingValue, reversed, false) {
+        String fieldname, int numHits, Pruning pruning, boolean reversed) {
+      return new LongComparator(numHits, fieldname, missingValue, reversed, Pruning.NONE) {
         @Override
         public LeafFieldComparator getLeafComparator(LeafReaderContext context) throws IOException {
           LongValuesHolder holder = new LongValuesHolder();