Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

add LongPoint.newSetQuery; improve PointInSetQuery.toString

This commit is contained in:
Mike McCandless 2016-02-24 04:50:37 -05:00
parent 3f5ed6eb67
commit 79c8bce1af
4 changed files with 143 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
lucene/core/src/java/org/apache/lucene/document/IntPoint.java
View File

@ -219,14 +219,14 @@ public final class IntPoint extends Field {
}
/**
* Returns a query efficiently finding all documents indexed with any of the specified 1D values.
* Create a query matching any of the specified 1D values. This is the points equivalent of {@code TermsQuery}.
*
* @param field field name. must not be {@code null}.
* @param valuesIn all int values to search for
* @param valuesIn all int values to match
*/
public static PointInSetQuery newSetQuery(String field, int... valuesIn) throws IOException {
// Don't unexpectedly change the user's incoming array:
// Don't unexpectedly change the user's incoming values array:
int[] values = valuesIn.clone();
Arrays.sort(values);

45
lucene/core/src/java/org/apache/lucene/document/LongPoint.java
View File

@ -16,8 +16,13 @@
*/
package org.apache.lucene.document;
import java.io.IOException;
import java.util.Arrays;
import org.apache.lucene.search.PointInSetQuery;
import org.apache.lucene.search.PointRangeQuery;
import org.apache.lucene.util.BytesRef;
import org.apache.lucene.util.BytesRefIterator;
import org.apache.lucene.util.NumericUtils;
/**
@ -212,4 +217,44 @@ public final class LongPoint extends Field {
}
};
}
/**
* Create a query matching any of the specified 1D values. This is the points equivalent of {@code TermsQuery}.
*
* @param field field name. must not be {@code null}.
* @param valuesIn all int values to match
*/
public static PointInSetQuery newSetQuery(String field, long... valuesIn) throws IOException {
// Don't unexpectedly change the user's incoming values array:
long[] values = valuesIn.clone();
Arrays.sort(values);
final BytesRef value = new BytesRef(new byte[Long.BYTES]);
value.length = Long.BYTES;
return new PointInSetQuery(field, 1, Long.BYTES,
new BytesRefIterator() {
int upto;
@Override
public BytesRef next() {
if (upto == values.length) {
return null;
} else {
LongPoint.encodeDimension(values[upto], value.bytes, 0);
upto++;
return value;
}
}
}) {
@Override
protected String toString(byte[] value) {
assert value.length == Long.BYTES;
return Long.toString(decodeDimension(value, 0));
}
};
}
}

13
lucene/core/src/java/org/apache/lucene/search/PointInSetQuery.java
View File

@ -320,24 +320,25 @@ public class PointInSetQuery extends Query {
@Override
public String toString(String field) {
final StringBuilder sb = new StringBuilder();
sb.append(getClass().getSimpleName());
sb.append(':');
if (this.field.equals(field) == false) {
sb.append(" field=");
sb.append(this.field);
sb.append(':');
}
sb.append(" points:");
sb.append("{");
TermIterator iterator = sortedPackedPoints.iterator();
byte[] pointBytes = new byte[numDims * bytesPerDim];
boolean first = true;
for (BytesRef point = iterator.next(); point != null; point = iterator.next()) {
sb.append(' ');
if (first == false) {
sb.append(" ");
}
first = false;
System.arraycopy(point.bytes, point.offset, pointBytes, 0, pointBytes.length);
sb.append(toString(pointBytes));
}
sb.append("}");
return sb.toString();
}

94
lucene/core/src/test/org/apache/lucene/search/TestPointQueries.java
View File

@ -1127,11 +1127,31 @@ public class TestPointQueries extends LuceneTestCase {
return values;
}
private static int randomIntValue(Integer min, Integer max) {
if (min == null) {
return random().nextInt();
} else {
return TestUtil.nextInt(random(), min, max);
}
}
public void testRandomPointInSetQuery() throws Exception {
boolean useNarrowRange = random().nextBoolean();
final Integer valueMin;
final Integer valueMax;
if (useNarrowRange) {
int gap = random().nextInt(100);
valueMin = random().nextInt(Integer.MAX_VALUE-gap);
valueMax = valueMin + gap;
} else {
valueMin = null;
valueMax = null;
}
final Set<Integer> valuesSet = new HashSet<>();
int numValues = TestUtil.nextInt(random(), 1, 100);
while (valuesSet.size() < numValues) {
valuesSet.add(random().nextInt());
valuesSet.add(randomIntValue(valueMin, valueMax));
}
int[] values = toArray(valuesSet);
int numDocs = TestUtil.nextInt(random(), 1, 10000);
@ -1211,8 +1231,7 @@ public class TestPointQueries extends LuceneTestCase {
int numExtraValuesToQuery = random().nextInt(20);
while (valuesToQuery.size() < numValidValuesToQuery + numExtraValuesToQuery) {
// nocommit fix test to sometimes use "narrow" range of values
valuesToQuery.add(random().nextInt());
valuesToQuery.add(randomIntValue(valueMin, valueMax));
}
int expectedCount = 0;
@ -1241,8 +1260,6 @@ public class TestPointQueries extends LuceneTestCase {
IOUtils.close(r, dir);
}
// nocommit fix existing randomized tests to sometimes randomly use PointInSet instead
// nocommit need 2D test too
public void testBasicPointInSetQuery() throws Exception {
@ -1253,14 +1270,17 @@ public class TestPointQueries extends LuceneTestCase {
Document doc = new Document();
doc.add(new IntPoint("int", 17));
doc.add(new LongPoint("long", 17));
w.addDocument(doc);
doc = new Document();
doc.add(new IntPoint("int", 42));
doc.add(new LongPoint("long", 42));
w.addDocument(doc);
doc = new Document();
doc.add(new IntPoint("int", 97));
doc.add(new LongPoint("long", 97));
w.addDocument(doc);
IndexReader r = DirectoryReader.open(w);
@ -1271,6 +1291,46 @@ public class TestPointQueries extends LuceneTestCase {
assertEquals(3, s.count(IntPoint.newSetQuery("int", -7, 17, 42, 97)));
assertEquals(3, s.count(IntPoint.newSetQuery("int", 17, 20, 42, 97)));
assertEquals(3, s.count(IntPoint.newSetQuery("int", 17, 105, 42, 97)));
assertEquals(0, s.count(LongPoint.newSetQuery("long", 16)));
assertEquals(1, s.count(LongPoint.newSetQuery("long", 17)));
assertEquals(3, s.count(LongPoint.newSetQuery("long", 17, 97, 42)));
assertEquals(3, s.count(LongPoint.newSetQuery("long", -7, 17, 42, 97)));
assertEquals(3, s.count(LongPoint.newSetQuery("long", 17, 20, 42, 97)));
assertEquals(3, s.count(LongPoint.newSetQuery("long", 17, 105, 42, 97)));
w.close();
r.close();
dir.close();
}
public void testBasicMultiValuedPointInSetQuery() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig();
iwc.setCodec(getCodec());
IndexWriter w = new IndexWriter(dir, iwc);
Document doc = new Document();
doc.add(new IntPoint("int", 17));
doc.add(new IntPoint("int", 42));
doc.add(new LongPoint("long", 17));
doc.add(new LongPoint("long", 42));
w.addDocument(doc);
IndexReader r = DirectoryReader.open(w);
IndexSearcher s = newSearcher(r);
assertEquals(0, s.count(IntPoint.newSetQuery("int", 16)));
assertEquals(1, s.count(IntPoint.newSetQuery("int", 17)));
assertEquals(1, s.count(IntPoint.newSetQuery("int", 17, 97, 42)));
assertEquals(1, s.count(IntPoint.newSetQuery("int", -7, 17, 42, 97)));
assertEquals(0, s.count(IntPoint.newSetQuery("int", 16, 20, 41, 97)));
assertEquals(0, s.count(LongPoint.newSetQuery("long", 16)));
assertEquals(1, s.count(LongPoint.newSetQuery("long", 17)));
assertEquals(1, s.count(LongPoint.newSetQuery("long", 17, 97, 42)));
assertEquals(1, s.count(LongPoint.newSetQuery("long", -7, 17, 42, 97)));
assertEquals(0, s.count(LongPoint.newSetQuery("long", 16, 20, 41, 97)));
w.close();
r.close();
dir.close();
@ -1290,6 +1350,7 @@ public class TestPointQueries extends LuceneTestCase {
}
Document doc = new Document();
doc.add(new IntPoint("int", x));
doc.add(new LongPoint("long", x));
w.addDocument(doc);
}
@ -1300,12 +1361,18 @@ public class TestPointQueries extends LuceneTestCase {
assertEquals(zeroCount, s.count(IntPoint.newSetQuery("int", 7, 0)));
assertEquals(10000-zeroCount, s.count(IntPoint.newSetQuery("int", 1)));
assertEquals(0, s.count(IntPoint.newSetQuery("int", 2)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 0)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 0, -7)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 7, 0)));
assertEquals(10000-zeroCount, s.count(LongPoint.newSetQuery("long", 1)));
assertEquals(0, s.count(LongPoint.newSetQuery("long", 2)));
w.close();
r.close();
dir.close();
}
public void testPointInSetQueryManyEqualValuesBigGap() throws Exception {
public void testPointInSetQueryManyEqualValuesWithBigGap() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig();
iwc.setCodec(getCodec());
@ -1319,6 +1386,7 @@ public class TestPointQueries extends LuceneTestCase {
}
Document doc = new Document();
doc.add(new IntPoint("int", x));
doc.add(new LongPoint("long", x));
w.addDocument(doc);
}
@ -1329,6 +1397,12 @@ public class TestPointQueries extends LuceneTestCase {
assertEquals(zeroCount, s.count(IntPoint.newSetQuery("int", 7, 0)));
assertEquals(10000-zeroCount, s.count(IntPoint.newSetQuery("int", 200)));
assertEquals(0, s.count(IntPoint.newSetQuery("int", 2)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 0)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 0, -7)));
assertEquals(zeroCount, s.count(LongPoint.newSetQuery("long", 7, 0)));
assertEquals(10000-zeroCount, s.count(LongPoint.newSetQuery("long", 200)));
assertEquals(0, s.count(LongPoint.newSetQuery("long", 2)));
w.close();
r.close();
dir.close();
@ -1347,4 +1421,12 @@ public class TestPointQueries extends LuceneTestCase {
});
assertEquals("packed point length should be 12 but got 3; field=\"foo\" numDims=3 bytesPerDim=4", expected.getMessage());
}
public void testPointInSetQueryToString() throws Exception {
// int
assertEquals("int:{-42 18}", IntPoint.newSetQuery("int", -42, 18).toString());
// long
assertEquals("long:{-42 18}", LongPoint.newSetQuery("long", -42L, 18L).toString());
}
}