Skip to content

Apache Lucene open-source search software

Implement #intoBitset for DocIdSetIterator#all and DocIdSetIterator#range (#14269)

gf212107daa4ba1442
diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index 8a47ad202bb..19a40be2a61 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -131,6 +131,8 @@ Optimizations
* GITHUB#14176: Reduce when visiting bpv24-encoded doc ids in BKD leaves. (Guo Feng)
+* GITHUB#14269: Implement #intoBitset for DocIdSetIterator#all and DocIdSetIterator#range. (Guo Feng)
+
Bug Fixes
---------------------
diff --git a/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java b/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
index 42132344086..07c4a4f95e3 100644
--- a/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
+++ b/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
@@ -28,7 +28,7 @@ import org.apache.lucene.util.FixedBitSet;
public abstract class DocIdSetIterator {
/** An empty {@code DocIdSetIterator} instance */
- public static final DocIdSetIterator empty() {
+ public static DocIdSetIterator empty() {
return new DocIdSetIterator() {
boolean exhausted = false;
@@ -60,7 +60,7 @@ public abstract class DocIdSetIterator {
}
/** A {@link DocIdSetIterator} that matches all documents up to {@code maxDoc - 1}. */
- public static final DocIdSetIterator all(int maxDoc) {
+ public static DocIdSetIterator all(int maxDoc) {
return new DocIdSetIterator() {
int doc = -1;
@@ -70,12 +70,12 @@ public abstract class DocIdSetIterator {
}
@Override
- public int nextDoc() throws IOException {
+ public int nextDoc() {
return advance(doc + 1);
}
@Override
- public int advance(int target) throws IOException {
+ public int advance(int target) {
doc = target;
if (doc >= maxDoc) {
doc = NO_MORE_DOCS;
@@ -87,6 +87,16 @@ public abstract class DocIdSetIterator {
public long cost() {
return maxDoc;
}
+
+ @Override
+ public void intoBitSet(int upTo, FixedBitSet bitSet, int offset) {
+ assert offset <= doc;
+ upTo = Math.min(upTo, maxDoc);
+ if (upTo > doc) {
+ bitSet.set(doc - offset, upTo - offset);
+ advance(upTo);
+ }
+ }
};
}
@@ -94,7 +104,7 @@ public abstract class DocIdSetIterator {
* A {@link DocIdSetIterator} that matches a range documents from minDocID (inclusive) to maxDocID
* (exclusive).
*/
- public static final DocIdSetIterator range(int minDoc, int maxDoc) {
+ public static DocIdSetIterator range(int minDoc, int maxDoc) {
if (minDoc >= maxDoc) {
throw new IllegalArgumentException(
"minDoc must be < maxDoc but got minDoc=" + minDoc + " maxDoc=" + maxDoc);
@@ -111,12 +121,12 @@ public abstract class DocIdSetIterator {
}
@Override
- public int nextDoc() throws IOException {
+ public int nextDoc() {
return advance(doc + 1);
}
@Override
- public int advance(int target) throws IOException {
+ public int advance(int target) {
if (target < minDoc) {
doc = minDoc;
} else if (target >= maxDoc) {
@@ -131,6 +141,16 @@ public abstract class DocIdSetIterator {
public long cost() {
return maxDoc - minDoc;
}
+
+ @Override
+ public void intoBitSet(int upTo, FixedBitSet bitSet, int offset) {
+ assert offset <= doc;
+ upTo = Math.min(upTo, maxDoc);
+ if (upTo > doc) {
+ bitSet.set(doc - offset, upTo - offset);
+ advance(upTo);
+ }
+ }
};
}
@@ -224,7 +244,7 @@ public abstract class DocIdSetIterator {
* </pre>
*
* <p><b>Note</b>: {@code offset} must be less than or equal to the {@link #docID() current doc
- * ID}.
+ * ID}. Behaviour is undefined if this iterator is unpositioned.
*
* <p><b>Note</b>: It is important not to clear bits from {@code bitSet} that may be already set.
*
diff --git a/lucene/core/src/test/org/apache/lucene/search/TestDocIdSetIterator.java b/lucene/core/src/test/org/apache/lucene/search/TestDocIdSetIterator.java
index bfa5e1a4501..97decb4249d 100644
--- a/lucene/core/src/test/org/apache/lucene/search/TestDocIdSetIterator.java
+++ b/lucene/core/src/test/org/apache/lucene/search/TestDocIdSetIterator.java
@@ -20,6 +20,7 @@ package org.apache.lucene.search;
import static org.apache.lucene.search.DocIdSetIterator.NO_MORE_DOCS;
import org.apache.lucene.tests.util.LuceneTestCase;
+import org.apache.lucene.util.FixedBitSet;
public class TestDocIdSetIterator extends LuceneTestCase {
public void testRangeBasic() throws Exception {
@@ -64,4 +65,54 @@ public class TestDocIdSetIterator extends LuceneTestCase {
assertEquals(19, disi.nextDoc());
assertEquals(NO_MORE_DOCS, disi.nextDoc());
}
+
+ public void testIntoBitset() throws Exception {
+ for (int i = 0; i < 10; i++) {
+ int max = 1 + random().nextInt(500);
+ DocIdSetIterator expectedDisi;
+ DocIdSetIterator actualDisi;
+ if ((i & 1) == 0) {
+ int min = random().nextInt(max);
+ expectedDisi = DocIdSetIterator.range(min, max);
+ actualDisi = DocIdSetIterator.range(min, max);
+ } else {
+ expectedDisi = DocIdSetIterator.all(max);
+ actualDisi = DocIdSetIterator.all(max);
+ }
+ FixedBitSet expected = new FixedBitSet(max * 2);
+ FixedBitSet actual = new FixedBitSet(max * 2);
+ int doc = -1;
+ expectedDisi.nextDoc();
+ actualDisi.nextDoc();
+ while (doc != NO_MORE_DOCS) {
+ int r = random().nextInt(3);
+ switch (r) {
+ case 0 -> {
+ expectedDisi.nextDoc();
+ actualDisi.nextDoc();
+ }
+ case 1 -> {
+ int jump = expectedDisi.docID() + random().nextInt(5);
+ expectedDisi.advance(jump);
+ actualDisi.advance(jump);
+ }
+ case 2 -> {
+ expected.clear();
+ actual.clear();
+ int upTo =
+ random().nextBoolean()
+ ? expectedDisi.docID() - 1
+ : expectedDisi.docID() + random().nextInt(5);
+ int offset = expectedDisi.docID() - random().nextInt(max);
+ // use the default impl of intoBitSet
+ new FilterDocIdSetIterator(expectedDisi).intoBitSet(upTo, expected, offset);
+ actualDisi.intoBitSet(upTo, actual, offset);
+ assertArrayEquals(expected.getBits(), actual.getBits());
+ }
+ }
+ assertEquals(expectedDisi.docID(), actualDisi.docID());
+ doc = expectedDisi.docID();
+ }
+ }
+ }
}