LUCENE-10598: SortedSetDocValues#docValueCount() should be always greater than zero (#934)

* LUCENE-10598: SortedSetDocValues#docValueCount() should be always greater than zero

lucene/CHANGES.txt

@@ -90,6 +90,8 @@ Bug Fixes
 
 * LUCENE-10582: Fix merging of overridden CollectionStatistics in CombinedFieldQuery (Yannick Welsch)
 
+* LUCENE-10598: SortedSetDocValues#docValueCount() should be always greater than zero. (Lu Xugang)
+
 Other
 ---------------------
 

lucene/core/src/java/org/apache/lucene/codecs/lucene90/Lucene90DocValuesProducer.java

@@ -1454,7 +1454,7 @@ final class Lucene90DocValuesProducer extends DocValuesProducer {
 
       @Override
       public long docValueCount() {
-        return count;
+        return ords.docValueCount();
       }
 
       @Override

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

@@ -3337,9 +3337,22 @@ public final class CheckIndex implements Closeable {
     LongBitSet seenOrds = new LongBitSet(dv.getValueCount());
     long maxOrd2 = -1;
     for (int docID = dv.nextDoc(); docID != NO_MORE_DOCS; docID = dv.nextDoc()) {
+      long count = dv.docValueCount();
+      if (count == 0) {
+        throw new CheckIndexException(
+            "sortedset dv for field: "
+                + fieldName
+                + " returned docValueCount=0 for docID="
+                + docID);
+      }
       if (dv2.advanceExact(docID) == false) {
         throw new CheckIndexException("advanceExact did not find matching doc ID: " + docID);
       }
+      long count2 = dv2.docValueCount();
+      if (count != count2) {
+        throw new CheckIndexException(
+            "advanceExact reports different value count: " + count + " != " + count2);
+      }
       long lastOrd = -1;
       long ord;
       int ordCount = 0;

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

@@ -45,8 +45,9 @@ public abstract class SortedSetDocValues extends DocValuesIterator {
   public abstract long nextOrd() throws IOException;
 
   /**
-   * Retrieves the number of values for the current document. This must always be greater than zero.
-   * It is illegal to call this method after {@link #advanceExact(int)} returned {@code false}.
+   * Retrieves the number of unique ords for the current document. This must always be greater than
+   * zero. It is illegal to call this method after {@link #advanceExact(int)} returned {@code
+   * false}.
    */
   public abstract long docValueCount();
 

lucene/core/src/test/org/apache/lucene/codecs/lucene90/TestLucene90DocValuesFormat.java

@@ -265,6 +265,7 @@ public class TestLucene90DocValuesFormat extends BaseCompressingDocValuesFormatT
             assertTrue(valueSet.contains(sortedNumeric.nextValue()));
           }
           assertEquals(i, sortedSet.nextDoc());
+          assertEquals(valueSet.size(), sortedSet.docValueCount());
           int sortedSetCount = 0;
           while (true) {
             long ord = sortedSet.nextOrd();
@@ -488,6 +489,7 @@ public class TestLucene90DocValuesFormat extends BaseCompressingDocValuesFormatT
       for (int i = 0; i < maxDoc; ++i) {
         assertEquals(i, values.nextDoc());
         final int numValues = in.readVInt();
+        assertEquals(numValues, values.docValueCount());
 
         for (int j = 0; j < numValues; ++j) {
           b.setLength(in.readVInt());

lucene/core/src/test/org/apache/lucene/index/TestMultiDocValues.java

@@ -279,6 +279,7 @@ public class TestMultiDocValues extends LuceneTestCase {
         if (docID == NO_MORE_DOCS) {
           break;
         }
+        assertEquals(single.docValueCount(), multi.docValueCount());
 
         ArrayList<Long> expectedList = new ArrayList<>();
         long ord;

lucene/core/src/test/org/apache/lucene/index/TestSortingCodecReader.java

@@ -230,6 +230,7 @@ public class TestSortingCodecReader extends LuceneTestCase {
               assertEquals(
                   new BytesRef(ids.longValue() + ""),
                   sorted_set_dv.lookupOrd(sorted_set_dv.nextOrd()));
+              assertEquals(1, sorted_set_dv.docValueCount());
               assertEquals(1, sorted_numeric_dv.docValueCount());
               assertEquals(ids.longValue(), sorted_numeric_dv.nextValue());
 

lucene/memory/src/test/org/apache/lucene/index/memory/TestMemoryIndex.java

@@ -297,6 +297,7 @@ public class TestMemoryIndex extends LuceneTestCase {
     SortedSetDocValues sortedSetDocValues = leafReader.getSortedSetDocValues("sorted_set");
     assertEquals(3, sortedSetDocValues.getValueCount());
     assertEquals(0, sortedSetDocValues.nextDoc());
+    assertEquals(3, sortedSetDocValues.docValueCount());
     assertEquals(0L, sortedSetDocValues.nextOrd());
     assertEquals(1L, sortedSetDocValues.nextOrd());
     assertEquals(2L, sortedSetDocValues.nextOrd());