Examples with TopFieldDocs org.apache.lucene.search.TopFieldDocs used on opensource projects

Example 26 with TopFieldDocs

use of org.apache.lucene.search.TopFieldDocs in project elasticsearch by elastic.

the class AbstractNumberNestedSortingTestCase method testNestedSorting.

public void testNestedSorting() throws Exception {
    List<Document> docs = new ArrayList<>();
    Document document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 1));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 2));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 2));
    docs.add(document);
    writer.addDocuments(docs);
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 1));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 3));
    docs.add(document);
    writer.addDocuments(docs);
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 4));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 4));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 5));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 5));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 6));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 6));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    // This doc will not be included, because it doesn't have nested docs
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 7));
    writer.addDocument(document);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 3));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(createField("field2", 7));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("__type", "parent", Field.Store.NO));
    document.add(createField("field1", 8));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    // Some garbage docs, just to check if the NestedFieldComparator can deal with this.
    document = new Document();
    document.add(new StringField("fieldXXX", "x", Field.Store.NO));
    writer.addDocument(document);
    document = new Document();
    document.add(new StringField("fieldXXX", "x", Field.Store.NO));
    writer.addDocument(document);
    document = new Document();
    document.add(new StringField("fieldXXX", "x", Field.Store.NO));
    writer.addDocument(document);
    MultiValueMode sortMode = MultiValueMode.SUM;
    DirectoryReader directoryReader = DirectoryReader.open(writer);
    directoryReader = ElasticsearchDirectoryReader.wrap(directoryReader, new ShardId(indexService.index(), 0));
    IndexSearcher searcher = new IndexSearcher(directoryReader);
    Query parentFilter = new TermQuery(new Term("__type", "parent"));
    Query childFilter = Queries.not(parentFilter);
    XFieldComparatorSource nestedComparatorSource = createFieldComparator("field2", sortMode, null, createNested(searcher, parentFilter, childFilter));
    ToParentBlockJoinQuery query = new ToParentBlockJoinQuery(new ConstantScoreQuery(childFilter), new QueryBitSetProducer(parentFilter), ScoreMode.None);
    Sort sort = new Sort(new SortField("field2", nestedComparatorSource));
    TopFieldDocs topDocs = searcher.search(query, 5, sort);
    assertThat(topDocs.totalHits, equalTo(7));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(11));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(7));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(7));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(8));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(3));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(9));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(15));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(10));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(19));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(11));
    sort = new Sort(new SortField("field2", nestedComparatorSource, true));
    topDocs = searcher.search(query, 5, sort);
    assertThat(topDocs.totalHits, equalTo(7));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(28));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(13));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(23));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(12));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(19));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(11));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(15));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(10));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(3));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(9));
    childFilter = new TermQuery(new Term("filter_1", "T"));
    nestedComparatorSource = createFieldComparator("field2", sortMode, null, createNested(searcher, parentFilter, childFilter));
    query = new ToParentBlockJoinQuery(new ConstantScoreQuery(childFilter), new QueryBitSetProducer(parentFilter), ScoreMode.None);
    sort = new Sort(new SortField("field2", nestedComparatorSource, true));
    topDocs = searcher.search(query, 5, sort);
    assertThat(topDocs.totalHits, equalTo(6));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(23));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(12));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(3));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(9));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(7));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(8));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(11));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(7));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(15));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(3));
    sort = new Sort(new SortField("field2", nestedComparatorSource));
    topDocs = searcher.search(query, 5, sort);
    assertThat(topDocs.totalHits, equalTo(6));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(15));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(3));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(28));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(3));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(11));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(7));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(7));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(8));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(3));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(9));
    nestedComparatorSource = createFieldComparator("field2", sortMode, 127, createNested(searcher, parentFilter, childFilter));
    sort = new Sort(new SortField("field2", nestedComparatorSource, true));
    topDocs = searcher.search(new TermQuery(new Term("__type", "parent")), 5, sort);
    assertThat(topDocs.totalHits, equalTo(8));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(19));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(127));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(24));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(127));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(23));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(12));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(3));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(9));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(7));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(8));
    nestedComparatorSource = createFieldComparator("field2", sortMode, -127, createNested(searcher, parentFilter, childFilter));
    sort = new Sort(new SortField("field2", nestedComparatorSource));
    topDocs = searcher.search(new TermQuery(new Term("__type", "parent")), 5, sort);
    assertThat(topDocs.totalHits, equalTo(8));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(19));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).intValue(), equalTo(-127));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(24));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).intValue(), equalTo(-127));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(15));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).intValue(), equalTo(3));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(28));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).intValue(), equalTo(3));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(11));
    assertThat(((Number) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).intValue(), equalTo(7));
    // Moved to method, because floating point based XFieldComparatorSource have different outcome for SortMode avg,
    // than integral number based implementations...
    assertAvgScoreMode(parentFilter, searcher);
    searcher.getIndexReader().close();
}

Example 27 with TopFieldDocs

use of org.apache.lucene.search.TopFieldDocs in project elasticsearch by elastic.

the class InternalTopHitsTests method createTestInstance.

@Override
protected InternalTopHits createTestInstance(String name, List<PipelineAggregator> pipelineAggregators, Map<String, Object> metaData) {
    int from = 0;
    int requestedSize = between(1, 40);
    int actualSize = between(0, requestedSize);
    float maxScore = Float.MIN_VALUE;
    ScoreDoc[] scoreDocs = new ScoreDoc[actualSize];
    SearchHit[] hits = new SearchHit[actualSize];
    Set<Integer> usedDocIds = new HashSet<>();
    for (int i = 0; i < actualSize; i++) {
        float score = randomFloat();
        maxScore = max(maxScore, score);
        int docId = randomValueOtherThanMany(usedDocIds::contains, () -> between(0, IndexWriter.MAX_DOCS));
        usedDocIds.add(docId);
        Map<String, SearchHitField> searchHitFields = new HashMap<>();
        if (testInstancesLookSortedByField) {
            Object[] fields = new Object[testInstancesSortFields.length];
            for (int f = 0; f < testInstancesSortFields.length; f++) {
                fields[f] = randomOfType(testInstancesSortFields[f].getType());
            }
            scoreDocs[i] = new FieldDoc(docId, score, fields);
        } else {
            scoreDocs[i] = new ScoreDoc(docId, score);
        }
        hits[i] = new SearchHit(docId, Integer.toString(i), new Text("test"), searchHitFields);
        hits[i].score(score);
    }
    int totalHits = between(actualSize, 500000);
    SearchHits searchHits = new SearchHits(hits, totalHits, maxScore);
    TopDocs topDocs;
    Arrays.sort(scoreDocs, scoreDocComparator());
    if (testInstancesLookSortedByField) {
        topDocs = new TopFieldDocs(totalHits, scoreDocs, testInstancesSortFields, maxScore);
    } else {
        topDocs = new TopDocs(totalHits, scoreDocs, maxScore);
    }
    return new InternalTopHits(name, from, requestedSize, topDocs, searchHits, pipelineAggregators, metaData);
}

Example 28 with TopFieldDocs

use of org.apache.lucene.search.TopFieldDocs in project elasticsearch by elastic.

the class SimpleLuceneTests method testSortValues.

public void testSortValues() throws Exception {
    Directory dir = new RAMDirectory();
    IndexWriter indexWriter = new IndexWriter(dir, new IndexWriterConfig(Lucene.STANDARD_ANALYZER));
    for (int i = 0; i < 10; i++) {
        Document document = new Document();
        String text = new String(new char[] { (char) (97 + i), (char) (97 + i) });
        document.add(new TextField("str", text, Field.Store.YES));
        document.add(new SortedDocValuesField("str", new BytesRef(text)));
        indexWriter.addDocument(document);
    }
    IndexReader reader = SlowCompositeReaderWrapper.wrap(DirectoryReader.open(indexWriter));
    IndexSearcher searcher = new IndexSearcher(reader);
    TopFieldDocs docs = searcher.search(new MatchAllDocsQuery(), 10, new Sort(new SortField("str", SortField.Type.STRING)));
    for (int i = 0; i < 10; i++) {
        FieldDoc fieldDoc = (FieldDoc) docs.scoreDocs[i];
        assertThat((BytesRef) fieldDoc.fields[0], equalTo(new BytesRef(new String(new char[] { (char) (97 + i), (char) (97 + i) }))));
    }
}

Example 29 with TopFieldDocs

use of org.apache.lucene.search.TopFieldDocs in project elasticsearch by elastic.

the class AbstractFieldDataImplTestCase method testMultiValueAllSet.

public void testMultiValueAllSet() throws Exception {
    fillMultiValueAllSet();
    // the segments are force merged to a single segment so that the sorted binary doc values can be asserted within a single segment.
    // Previously we used the SlowCompositeReaderWrapper but this is an unideal solution so force merging is a better idea.
    writer.forceMerge(1);
    IndexFieldData indexFieldData = getForField("value");
    List<LeafReaderContext> readerContexts = refreshReader();
    for (LeafReaderContext readerContext : readerContexts) {
        AtomicFieldData fieldData = indexFieldData.load(readerContext);
        assertThat(fieldData.ramBytesUsed(), greaterThanOrEqualTo(minRamBytesUsed()));
        SortedBinaryDocValues bytesValues = fieldData.getBytesValues();
        assertValues(bytesValues, 0, two(), four());
        assertValues(bytesValues, 1, one());
        assertValues(bytesValues, 2, three());
        IndexSearcher searcher = new IndexSearcher(DirectoryReader.open(writer));
        TopFieldDocs topDocs = searcher.search(new MatchAllDocsQuery(), 10, new Sort(new SortField("value", indexFieldData.comparatorSource(null, MultiValueMode.MIN, null))));
        assertThat(topDocs.totalHits, equalTo(3));
        assertThat(topDocs.scoreDocs.length, equalTo(3));
        assertThat(topDocs.scoreDocs[0].doc, equalTo(1));
        assertThat(topDocs.scoreDocs[1].doc, equalTo(0));
        assertThat(topDocs.scoreDocs[2].doc, equalTo(2));
        topDocs = searcher.search(new MatchAllDocsQuery(), 10, new Sort(new SortField("value", indexFieldData.comparatorSource(null, MultiValueMode.MAX, null), true)));
        assertThat(topDocs.totalHits, equalTo(3));
        assertThat(topDocs.scoreDocs.length, equalTo(3));
        assertThat(topDocs.scoreDocs[0].doc, equalTo(0));
        assertThat(topDocs.scoreDocs[1].doc, equalTo(2));
        assertThat(topDocs.scoreDocs[2].doc, equalTo(1));
    }
}

Example 30 with TopFieldDocs

use of org.apache.lucene.search.TopFieldDocs in project elasticsearch by elastic.

the class AbstractFieldDataImplTestCase method testSingleValueAllSet.

public void testSingleValueAllSet() throws Exception {
    fillSingleValueAllSet();
    IndexFieldData indexFieldData = getForField("value");
    List<LeafReaderContext> readerContexts = refreshReader();
    for (LeafReaderContext readerContext : readerContexts) {
        AtomicFieldData fieldData = indexFieldData.load(readerContext);
        assertThat(fieldData.ramBytesUsed(), greaterThanOrEqualTo(minRamBytesUsed()));
        SortedBinaryDocValues bytesValues = fieldData.getBytesValues();
        bytesValues.setDocument(0);
        assertThat(bytesValues.count(), equalTo(1));
        assertThat(bytesValues.valueAt(0), equalTo(new BytesRef(two())));
        bytesValues.setDocument(1);
        assertThat(bytesValues.count(), equalTo(1));
        assertThat(bytesValues.valueAt(0), equalTo(new BytesRef(one())));
        bytesValues.setDocument(2);
        assertThat(bytesValues.count(), equalTo(1));
        assertThat(bytesValues.valueAt(0), equalTo(new BytesRef(three())));
        assertValues(bytesValues, 0, two());
        assertValues(bytesValues, 1, one());
        assertValues(bytesValues, 2, three());
        IndexSearcher searcher = new IndexSearcher(readerContext.reader());
        TopFieldDocs topDocs;
        topDocs = searcher.search(new MatchAllDocsQuery(), 10, new Sort(new SortField("value", indexFieldData.comparatorSource(null, MultiValueMode.MIN, null))));
        assertThat(topDocs.totalHits, equalTo(3));
        assertThat(topDocs.scoreDocs[0].doc, equalTo(1));
        assertThat(toString(((FieldDoc) topDocs.scoreDocs[0]).fields[0]), equalTo(one()));
        assertThat(topDocs.scoreDocs[1].doc, equalTo(0));
        assertThat(toString(((FieldDoc) topDocs.scoreDocs[1]).fields[0]), equalTo(two()));
        assertThat(topDocs.scoreDocs[2].doc, equalTo(2));
        assertThat(toString(((FieldDoc) topDocs.scoreDocs[2]).fields[0]), equalTo(three()));
        topDocs = searcher.search(new MatchAllDocsQuery(), 10, new Sort(new SortField("value", indexFieldData.comparatorSource(null, MultiValueMode.MAX, null), true)));
        assertThat(topDocs.totalHits, equalTo(3));
        assertThat(topDocs.scoreDocs[0].doc, equalTo(2));
        assertThat(topDocs.scoreDocs[1].doc, equalTo(0));
        assertThat(topDocs.scoreDocs[2].doc, equalTo(1));
    }
}