Examples with RandomIndexWriter org.apache.lucene.index.RandomIndexWriter used on opensource projects

Example 6 with RandomIndexWriter

use of org.apache.lucene.index.RandomIndexWriter in project elasticsearch by elastic.

the class NestedAggregatorTests method testResetRootDocId.

public void testResetRootDocId() throws Exception {
    Directory directory = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(null);
    iwc.setMergePolicy(NoMergePolicy.INSTANCE);
    RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory, iwc);
    List<Document> documents = new ArrayList<>();
    // 1 segment with, 1 root document, with 3 nested sub docs
    Document document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#1", UidFieldMapper.Defaults.FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    indexWriter.addDocuments(documents);
    indexWriter.commit();
    documents.clear();
    // 1 segment with:
    // 1 document, with 1 nested subdoc
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#2", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#2", UidFieldMapper.Defaults.FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    indexWriter.addDocuments(documents);
    documents.clear();
    // and 1 document, with 1 nested subdoc
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#3", UidFieldMapper.Defaults.NESTED_FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "__nested_field", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    document = new Document();
    document.add(new Field(UidFieldMapper.NAME, "type#3", UidFieldMapper.Defaults.FIELD_TYPE));
    document.add(new Field(TypeFieldMapper.NAME, "test", TypeFieldMapper.Defaults.FIELD_TYPE));
    documents.add(document);
    indexWriter.addDocuments(documents);
    indexWriter.commit();
    indexWriter.close();
    IndexService indexService = createIndex("test");
    DirectoryReader directoryReader = DirectoryReader.open(directory);
    directoryReader = ElasticsearchDirectoryReader.wrap(directoryReader, new ShardId(indexService.index(), 0));
    IndexSearcher searcher = new IndexSearcher(directoryReader);
    indexService.mapperService().merge("test", new CompressedXContent(PutMappingRequest.buildFromSimplifiedDef("test", "nested_field", "type=nested").string()), MapperService.MergeReason.MAPPING_UPDATE, false);
    SearchContext context = createSearchContext(indexService);
    AggregatorFactories.Builder builder = AggregatorFactories.builder();
    NestedAggregationBuilder factory = new NestedAggregationBuilder("test", "nested_field");
    builder.addAggregator(factory);
    AggregatorFactories factories = builder.build(context, null);
    context.aggregations(new SearchContextAggregations(factories));
    Aggregator[] aggs = factories.createTopLevelAggregators();
    BucketCollector collector = BucketCollector.wrap(Arrays.asList(aggs));
    collector.preCollection();
    // A regular search always exclude nested docs, so we use NonNestedDocsFilter.INSTANCE here (otherwise MatchAllDocsQuery would be sufficient)
    // We exclude root doc with uid type#2, this will trigger the bug if we don't reset the root doc when we process a new segment, because
    // root doc type#3 and root doc type#1 have the same segment docid
    BooleanQuery.Builder bq = new BooleanQuery.Builder();
    bq.add(Queries.newNonNestedFilter(), Occur.MUST);
    bq.add(new TermQuery(new Term(UidFieldMapper.NAME, "type#2")), Occur.MUST_NOT);
    searcher.search(new ConstantScoreQuery(bq.build()), collector);
    collector.postCollection();
    Nested nested = (Nested) aggs[0].buildAggregation(0);
    // The bug manifests if 6 docs are returned, because currentRootDoc isn't reset the previous child docs from the first segment are emitted as hits.
    assertThat(nested.getDocCount(), equalTo(4L));
    directoryReader.close();
    directory.close();
}

Example 7 with RandomIndexWriter

use of org.apache.lucene.index.RandomIndexWriter in project elasticsearch by elastic.

the class ParentToChildrenAggregatorTests method testParentChild.

public void testParentChild() throws IOException {
    Directory directory = newDirectory();
    RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory);
    final Map<String, Tuple<Integer, Integer>> expectedParentChildRelations = setupIndex(indexWriter);
    indexWriter.close();
    IndexReader indexReader = ElasticsearchDirectoryReader.wrap(DirectoryReader.open(directory), new ShardId(new Index("foo", "_na_"), 1));
    // TODO set "maybeWrap" to true for IndexSearcher once #23338 is resolved
    IndexSearcher indexSearcher = newSearcher(indexReader, false, true);
    testCase(new MatchAllDocsQuery(), indexSearcher, child -> {
        int expectedTotalChildren = 0;
        int expectedMinValue = Integer.MAX_VALUE;
        for (Tuple<Integer, Integer> expectedValues : expectedParentChildRelations.values()) {
            expectedTotalChildren += expectedValues.v1();
            expectedMinValue = Math.min(expectedMinValue, expectedValues.v2());
        }
        assertEquals(expectedTotalChildren, child.getDocCount());
        assertEquals(expectedMinValue, ((InternalMin) child.getAggregations().get("in_child")).getValue(), Double.MIN_VALUE);
    });
    for (String parent : expectedParentChildRelations.keySet()) {
        testCase(new TermInSetQuery(UidFieldMapper.NAME, new BytesRef(Uid.createUid(PARENT_TYPE, parent))), indexSearcher, child -> {
            assertEquals((long) expectedParentChildRelations.get(parent).v1(), child.getDocCount());
            assertEquals(expectedParentChildRelations.get(parent).v2(), ((InternalMin) child.getAggregations().get("in_child")).getValue(), Double.MIN_VALUE);
        });
    }
    indexReader.close();
    directory.close();
}

Example 8 with RandomIndexWriter

use of org.apache.lucene.index.RandomIndexWriter in project elasticsearch by elastic.

the class TermsAggregatorTests method createIndexWithLongs.

private IndexReader createIndexWithLongs() throws IOException {
    Directory directory = newDirectory();
    RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory);
    Document document = new Document();
    document.add(new SortedNumericDocValuesField("number", 10));
    document.add(new SortedNumericDocValuesField("number", 100));
    indexWriter.addDocument(document);
    document = new Document();
    document.add(new SortedNumericDocValuesField("number", 1));
    document.add(new SortedNumericDocValuesField("number", 100));
    indexWriter.addDocument(document);
    document = new Document();
    document.add(new SortedNumericDocValuesField("number", 10));
    document.add(new SortedNumericDocValuesField("number", 1000));
    indexWriter.addDocument(document);
    indexWriter.close();
    return DirectoryReader.open(directory);
}

Example 9 with RandomIndexWriter

use of org.apache.lucene.index.RandomIndexWriter in project elasticsearch by elastic.

the class TermsAggregatorTests method createIndexWithDoubles.

private IndexReader createIndexWithDoubles() throws IOException {
    Directory directory = newDirectory();
    RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory);
    Document document = new Document();
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(10.0d)));
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(100.0d)));
    indexWriter.addDocument(document);
    document = new Document();
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(1.0d)));
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(100.0d)));
    indexWriter.addDocument(document);
    document = new Document();
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(10.0d)));
    document.add(new SortedNumericDocValuesField("number", NumericUtils.doubleToSortableLong(1000.0d)));
    indexWriter.addDocument(document);
    indexWriter.close();
    return DirectoryReader.open(directory);
}

Example 10 with RandomIndexWriter

use of org.apache.lucene.index.RandomIndexWriter in project elasticsearch by elastic.

the class MaxAggregatorTests method testCase.

private void testCase(Query query, CheckedConsumer<RandomIndexWriter, IOException> buildIndex, Consumer<InternalMax> verify) throws IOException {
    Directory directory = newDirectory();
    RandomIndexWriter indexWriter = new RandomIndexWriter(random(), directory);
    buildIndex.accept(indexWriter);
    indexWriter.close();
    IndexReader indexReader = DirectoryReader.open(directory);
    IndexSearcher indexSearcher = newSearcher(indexReader, true, true);
    MaxAggregationBuilder aggregationBuilder = new MaxAggregationBuilder("_name").field("number");
    MappedFieldType fieldType = new NumberFieldMapper.NumberFieldType(NumberFieldMapper.NumberType.LONG);
    fieldType.setName("number");
    try (MaxAggregator aggregator = createAggregator(aggregationBuilder, indexSearcher, fieldType)) {
        aggregator.preCollection();
        indexSearcher.search(query, aggregator);
        aggregator.postCollection();
        verify.accept((InternalMax) aggregator.buildAggregation(0L));
    }
    indexReader.close();
    directory.close();
}