Search in sources :

Example 1 with PagedBytesIndexFieldData

use of org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData in project OpenSearch by opensearch-project.

the class FieldDataCacheTests method testLoadGlobal_neverCacheIfFieldIsMissing.

public void testLoadGlobal_neverCacheIfFieldIsMissing() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(null);
    iwc.setMergePolicy(NoMergePolicy.INSTANCE);
    IndexWriter iw = new IndexWriter(dir, iwc);
    long numDocs = scaledRandomIntBetween(32, 128);
    for (int i = 1; i <= numDocs; i++) {
        Document doc = new Document();
        doc.add(new SortedSetDocValuesField("field1", new BytesRef(String.valueOf(i))));
        doc.add(new StringField("field2", String.valueOf(i), Field.Store.NO));
        iw.addDocument(doc);
        if (i % 24 == 0) {
            iw.commit();
        }
    }
    iw.close();
    DirectoryReader ir = OpenSearchDirectoryReader.wrap(DirectoryReader.open(dir), new ShardId("_index", "_na_", 0));
    DummyAccountingFieldDataCache fieldDataCache = new DummyAccountingFieldDataCache();
    // Testing SortedSetOrdinalsIndexFieldData:
    SortedSetOrdinalsIndexFieldData sortedSetOrdinalsIndexFieldData = createSortedDV("field1", fieldDataCache);
    sortedSetOrdinalsIndexFieldData.loadGlobal(ir);
    assertThat(fieldDataCache.cachedGlobally, equalTo(1));
    sortedSetOrdinalsIndexFieldData.loadGlobal(new FieldMaskingReader("field1", ir));
    assertThat(fieldDataCache.cachedGlobally, equalTo(1));
    // Testing PagedBytesIndexFieldData
    PagedBytesIndexFieldData pagedBytesIndexFieldData = createPagedBytes("field2", fieldDataCache);
    pagedBytesIndexFieldData.loadGlobal(ir);
    assertThat(fieldDataCache.cachedGlobally, equalTo(2));
    pagedBytesIndexFieldData.loadGlobal(new FieldMaskingReader("field2", ir));
    assertThat(fieldDataCache.cachedGlobally, equalTo(2));
    ir.close();
    dir.close();
}
Also used : PagedBytesIndexFieldData(org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData) OpenSearchDirectoryReader(org.opensearch.common.lucene.index.OpenSearchDirectoryReader) DirectoryReader(org.apache.lucene.index.DirectoryReader) SortedSetOrdinalsIndexFieldData(org.opensearch.index.fielddata.plain.SortedSetOrdinalsIndexFieldData) FieldMaskingReader(org.opensearch.test.FieldMaskingReader) Document(org.apache.lucene.document.Document) ShardId(org.opensearch.index.shard.ShardId) IndexWriter(org.apache.lucene.index.IndexWriter) StringField(org.apache.lucene.document.StringField) SortedSetDocValuesField(org.apache.lucene.document.SortedSetDocValuesField) BytesRef(org.apache.lucene.util.BytesRef) Directory(org.apache.lucene.store.Directory) IndexWriterConfig(org.apache.lucene.index.IndexWriterConfig)

Example 2 with PagedBytesIndexFieldData

use of org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData in project OpenSearch by opensearch-project.

the class NestedSortingTests method testNestedSorting.

public void testNestedSorting() throws Exception {
    List<Document> docs = new ArrayList<>();
    Document document = new Document();
    document.add(new StringField("field2", "a", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "b", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "c", Field.Store.NO));
    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(new StringField("field1", "a", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "c", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "d", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "e", Field.Store.NO));
    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(new StringField("field1", "b", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "e", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "f", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "g", Field.Store.NO));
    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(new StringField("field1", "c", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "g", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "h", Field.Store.NO));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "i", Field.Store.NO));
    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(new StringField("field1", "d", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    writer.commit();
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "i", Field.Store.NO));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "j", Field.Store.NO));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "k", Field.Store.NO));
    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(new StringField("field1", "f", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "k", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "l", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "m", Field.Store.NO));
    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(new StringField("field1", "g", Field.Store.NO));
    docs.add(document);
    writer.addDocuments(docs);
    // 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(new StringField("field1", "h", Field.Store.NO));
    writer.addDocument(document);
    docs.clear();
    document = new Document();
    document.add(new StringField("field2", "m", Field.Store.NO));
    document.add(new StringField("filter_1", "T", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "n", Field.Store.NO));
    document.add(new StringField("filter_1", "F", Field.Store.NO));
    docs.add(document);
    document = new Document();
    document.add(new StringField("field2", "o", Field.Store.NO));
    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(new StringField("field1", "i", Field.Store.NO));
    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.MIN;
    DirectoryReader reader = DirectoryReader.open(writer);
    reader = OpenSearchDirectoryReader.wrap(reader, new ShardId(indexService.index(), 0));
    IndexSearcher searcher = new IndexSearcher(reader);
    PagedBytesIndexFieldData indexFieldData = getForField("field2");
    Query parentFilter = new TermQuery(new Term("__type", "parent"));
    Query childFilter = Queries.not(parentFilter);
    BytesRefFieldComparatorSource nestedComparatorSource = new BytesRefFieldComparatorSource(indexFieldData, null, sortMode, 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.value, equalTo(7L));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(3));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).utf8ToString(), equalTo("a"));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(7));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).utf8ToString(), equalTo("c"));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(11));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).utf8ToString(), equalTo("e"));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(15));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).utf8ToString(), equalTo("g"));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(19));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).utf8ToString(), equalTo("i"));
    sortMode = MultiValueMode.MAX;
    nestedComparatorSource = new BytesRefFieldComparatorSource(indexFieldData, null, sortMode, createNested(searcher, parentFilter, childFilter));
    sort = new Sort(new SortField("field2", nestedComparatorSource, true));
    topDocs = searcher.search(query, 5, sort);
    assertThat(topDocs.totalHits.value, equalTo(7L));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(28));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).utf8ToString(), equalTo("o"));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(23));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).utf8ToString(), equalTo("m"));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(19));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).utf8ToString(), equalTo("k"));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(15));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).utf8ToString(), equalTo("i"));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(11));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).utf8ToString(), equalTo("g"));
    BooleanQuery.Builder bq = new BooleanQuery.Builder();
    bq.add(parentFilter, Occur.MUST_NOT);
    bq.add(new TermQuery(new Term("filter_1", "T")), Occur.MUST);
    childFilter = bq.build();
    nestedComparatorSource = new BytesRefFieldComparatorSource(indexFieldData, null, sortMode, 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.value, equalTo(6L));
    assertThat(topDocs.scoreDocs.length, equalTo(5));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(23));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[0]).fields[0]).utf8ToString(), equalTo("m"));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(28));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[1]).fields[0]).utf8ToString(), equalTo("m"));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(11));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[2]).fields[0]).utf8ToString(), equalTo("g"));
    assertThat(topDocs.scoreDocs[3].doc, equalTo(15));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[3]).fields[0]).utf8ToString(), equalTo("g"));
    assertThat(topDocs.scoreDocs[4].doc, equalTo(7));
    assertThat(((BytesRef) ((FieldDoc) topDocs.scoreDocs[4]).fields[0]).utf8ToString(), equalTo("e"));
    searcher.getIndexReader().close();
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) PagedBytesIndexFieldData(org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData) TermQuery(org.apache.lucene.search.TermQuery) BooleanQuery(org.apache.lucene.search.BooleanQuery) Query(org.apache.lucene.search.Query) ConstantScoreQuery(org.apache.lucene.search.ConstantScoreQuery) TermQuery(org.apache.lucene.search.TermQuery) BooleanQuery(org.apache.lucene.search.BooleanQuery) ToParentBlockJoinQuery(org.apache.lucene.search.join.ToParentBlockJoinQuery) DirectoryReader(org.apache.lucene.index.DirectoryReader) OpenSearchDirectoryReader(org.opensearch.common.lucene.index.OpenSearchDirectoryReader) FieldSortBuilder(org.opensearch.search.sort.FieldSortBuilder) NestedQueryBuilder(org.opensearch.index.query.NestedQueryBuilder) QueryBuilder(org.opensearch.index.query.QueryBuilder) RangeQueryBuilder(org.opensearch.index.query.RangeQueryBuilder) NestedSortBuilder(org.opensearch.search.sort.NestedSortBuilder) TermQueryBuilder(org.opensearch.index.query.TermQueryBuilder) XContentBuilder(org.opensearch.common.xcontent.XContentBuilder) MatchAllQueryBuilder(org.opensearch.index.query.MatchAllQueryBuilder) ArrayList(java.util.ArrayList) BytesRefFieldComparatorSource(org.opensearch.index.fielddata.fieldcomparator.BytesRefFieldComparatorSource) TopFieldDocs(org.apache.lucene.search.TopFieldDocs) SortField(org.apache.lucene.search.SortField) Term(org.apache.lucene.index.Term) Document(org.apache.lucene.document.Document) MultiValueMode(org.opensearch.search.MultiValueMode) ShardId(org.opensearch.index.shard.ShardId) ToParentBlockJoinQuery(org.apache.lucene.search.join.ToParentBlockJoinQuery) StringField(org.apache.lucene.document.StringField) ConstantScoreQuery(org.apache.lucene.search.ConstantScoreQuery) QueryBitSetProducer(org.apache.lucene.search.join.QueryBitSetProducer) Sort(org.apache.lucene.search.Sort)

Example 3 with PagedBytesIndexFieldData

use of org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData in project OpenSearch by opensearch-project.

the class NestedSortingTests method testDuel.

public void testDuel() throws Exception {
    final int numDocs = scaledRandomIntBetween(100, 1000);
    for (int i = 0; i < numDocs; ++i) {
        final int numChildren = randomInt(2);
        List<Document> docs = new ArrayList<>(numChildren + 1);
        for (int j = 0; j < numChildren; ++j) {
            Document doc = new Document();
            doc.add(new StringField("f", TestUtil.randomSimpleString(random(), 2), Field.Store.NO));
            doc.add(new StringField("__type", "child", Field.Store.NO));
            docs.add(doc);
        }
        if (randomBoolean()) {
            docs.add(new Document());
        }
        Document parent = new Document();
        parent.add(new StringField("__type", "parent", Field.Store.NO));
        docs.add(parent);
        writer.addDocuments(docs);
        if (rarely()) {
            // we need to have a bit more segments than what RandomIndexWriter would do by default
            DirectoryReader.open(writer).close();
        }
    }
    writer.commit();
    MultiValueMode sortMode = randomFrom(Arrays.asList(MultiValueMode.MIN, MultiValueMode.MAX));
    DirectoryReader reader = DirectoryReader.open(writer);
    reader = OpenSearchDirectoryReader.wrap(reader, new ShardId(indexService.index(), 0));
    IndexSearcher searcher = new IndexSearcher(reader);
    PagedBytesIndexFieldData indexFieldData1 = getForField("f");
    IndexFieldData<?> indexFieldData2 = NoOrdinalsStringFieldDataTests.hideOrdinals(indexFieldData1);
    final String missingValue = randomBoolean() ? null : TestUtil.randomSimpleString(random(), 2);
    final int n = randomIntBetween(1, numDocs + 2);
    final boolean reverse = randomBoolean();
    final TopDocs topDocs1 = getTopDocs(searcher, indexFieldData1, missingValue, sortMode, n, reverse);
    final TopDocs topDocs2 = getTopDocs(searcher, indexFieldData2, missingValue, sortMode, n, reverse);
    for (int i = 0; i < topDocs1.scoreDocs.length; ++i) {
        final FieldDoc fieldDoc1 = (FieldDoc) topDocs1.scoreDocs[i];
        final FieldDoc fieldDoc2 = (FieldDoc) topDocs2.scoreDocs[i];
        assertEquals(fieldDoc1.doc, fieldDoc2.doc);
        assertArrayEquals(fieldDoc1.fields, fieldDoc2.fields);
    }
    searcher.getIndexReader().close();
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) PagedBytesIndexFieldData(org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData) FieldDoc(org.apache.lucene.search.FieldDoc) DirectoryReader(org.apache.lucene.index.DirectoryReader) OpenSearchDirectoryReader(org.opensearch.common.lucene.index.OpenSearchDirectoryReader) ArrayList(java.util.ArrayList) Document(org.apache.lucene.document.Document) IntPoint(org.apache.lucene.document.IntPoint) MultiValueMode(org.opensearch.search.MultiValueMode) ShardId(org.opensearch.index.shard.ShardId) TopDocs(org.apache.lucene.search.TopDocs) StringField(org.apache.lucene.document.StringField)

Aggregations

Document (org.apache.lucene.document.Document)3 StringField (org.apache.lucene.document.StringField)3 DirectoryReader (org.apache.lucene.index.DirectoryReader)3 OpenSearchDirectoryReader (org.opensearch.common.lucene.index.OpenSearchDirectoryReader)3 PagedBytesIndexFieldData (org.opensearch.index.fielddata.plain.PagedBytesIndexFieldData)3 ShardId (org.opensearch.index.shard.ShardId)3 ArrayList (java.util.ArrayList)2 IndexSearcher (org.apache.lucene.search.IndexSearcher)2 MultiValueMode (org.opensearch.search.MultiValueMode)2 IntPoint (org.apache.lucene.document.IntPoint)1 SortedSetDocValuesField (org.apache.lucene.document.SortedSetDocValuesField)1 IndexWriter (org.apache.lucene.index.IndexWriter)1 IndexWriterConfig (org.apache.lucene.index.IndexWriterConfig)1 Term (org.apache.lucene.index.Term)1 BooleanQuery (org.apache.lucene.search.BooleanQuery)1 ConstantScoreQuery (org.apache.lucene.search.ConstantScoreQuery)1 FieldDoc (org.apache.lucene.search.FieldDoc)1 Query (org.apache.lucene.search.Query)1 Sort (org.apache.lucene.search.Sort)1 SortField (org.apache.lucene.search.SortField)1