Examples with SortedSetDocValuesField org.apache.lucene.document.SortedSetDocValuesField used on opensource projects

Example 6 with SortedSetDocValuesField

use of org.apache.lucene.document.SortedSetDocValuesField in project elasticsearch by elastic.

the class CollapsingTopDocsCollectorTests method testCollapseString.

public void testCollapseString() throws Exception {
    CollapsingDocValuesProducer producer = new CollapsingDocValuesProducer<BytesRef>() {

        @Override
        public BytesRef randomGroup(int maxGroup) {
            return new BytesRef(Integer.toString(randomIntBetween(0, maxGroup - 1)));
        }

        @Override
        public void add(Document doc, BytesRef value, boolean multivalued) {
            if (multivalued) {
                doc.add(new SortedSetDocValuesField("field", value));
            } else {
                doc.add(new SortedDocValuesField("field", value));
            }
        }

        @Override
        public SortField sortField(boolean multivalued) {
            if (multivalued) {
                return new SortedSetSortField("field", false);
            } else {
                return new SortField("field", SortField.Type.STRING_VAL);
            }
        }
    };
    assertSearchCollapse(producer, false);
}

Example 7 with SortedSetDocValuesField

use of org.apache.lucene.document.SortedSetDocValuesField in project lucene-solr by apache.

the class TestLucene54DocValuesFormat method doTestTermsEnumRandom.

// TODO: try to refactor this and some termsenum tests into the base class.
// to do this we need to fix the test class to get a DVF not a Codec so we can setup
// the postings format correctly.
private void doTestTermsEnumRandom(int numDocs, int minLength, int maxLength) throws Exception {
    Directory dir = newFSDirectory(createTempDir());
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random()));
    conf.setMergeScheduler(new SerialMergeScheduler());
    // set to duel against a codec which has ordinals:
    final PostingsFormat pf = TestUtil.getPostingsFormatWithOrds(random());
    final DocValuesFormat dv = new Lucene54DocValuesFormat();
    conf.setCodec(new AssertingCodec() {

        @Override
        public PostingsFormat getPostingsFormatForField(String field) {
            return pf;
        }

        @Override
        public DocValuesFormat getDocValuesFormatForField(String field) {
            return dv;
        }
    });
    RandomIndexWriter writer = new RandomIndexWriter(random(), dir, conf);
    // index some docs
    for (int i = 0; i < numDocs; i++) {
        Document doc = new Document();
        Field idField = new StringField("id", Integer.toString(i), Field.Store.NO);
        doc.add(idField);
        final int length = TestUtil.nextInt(random(), minLength, maxLength);
        int numValues = random().nextInt(17);
        // create a random list of strings
        List<String> values = new ArrayList<>();
        for (int v = 0; v < numValues; v++) {
            values.add(TestUtil.randomSimpleString(random(), minLength, length));
        }
        // add in any order to the indexed field
        ArrayList<String> unordered = new ArrayList<>(values);
        Collections.shuffle(unordered, random());
        for (String v : values) {
            doc.add(newStringField("indexed", v, Field.Store.NO));
        }
        // add in any order to the dv field
        ArrayList<String> unordered2 = new ArrayList<>(values);
        Collections.shuffle(unordered2, random());
        for (String v : unordered2) {
            doc.add(new SortedSetDocValuesField("dv", new BytesRef(v)));
        }
        writer.addDocument(doc);
        if (random().nextInt(31) == 0) {
            writer.commit();
        }
    }
    // delete some docs
    int numDeletions = random().nextInt(numDocs / 10);
    for (int i = 0; i < numDeletions; i++) {
        int id = random().nextInt(numDocs);
        writer.deleteDocuments(new Term("id", Integer.toString(id)));
    }
    // compare per-segment
    DirectoryReader ir = writer.getReader();
    for (LeafReaderContext context : ir.leaves()) {
        LeafReader r = context.reader();
        Terms terms = r.terms("indexed");
        if (terms != null) {
            SortedSetDocValues ssdv = r.getSortedSetDocValues("dv");
            assertEquals(terms.size(), ssdv.getValueCount());
            TermsEnum expected = terms.iterator();
            TermsEnum actual = r.getSortedSetDocValues("dv").termsEnum();
            assertEquals(terms.size(), expected, actual);
            doTestSortedSetEnumAdvanceIndependently(ssdv);
        }
    }
    ir.close();
    writer.forceMerge(1);
    // now compare again after the merge
    ir = writer.getReader();
    LeafReader ar = getOnlyLeafReader(ir);
    Terms terms = ar.terms("indexed");
    if (terms != null) {
        assertEquals(terms.size(), ar.getSortedSetDocValues("dv").getValueCount());
        TermsEnum expected = terms.iterator();
        TermsEnum actual = ar.getSortedSetDocValues("dv").termsEnum();
        assertEquals(terms.size(), expected, actual);
    }
    ir.close();
    writer.close();
    dir.close();
}

Example 8 with SortedSetDocValuesField

use of org.apache.lucene.document.SortedSetDocValuesField in project lucene-solr by apache.

the class TestLucene54DocValuesFormat method doTestSparseDocValuesVsStoredFields.

private void doTestSparseDocValuesVsStoredFields() throws Exception {
    final long[] values = new long[TestUtil.nextInt(random(), 1, 500)];
    for (int i = 0; i < values.length; ++i) {
        values[i] = random().nextLong();
    }
    Directory dir = newFSDirectory(createTempDir());
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random()));
    conf.setMergeScheduler(new SerialMergeScheduler());
    RandomIndexWriter writer = new RandomIndexWriter(random(), dir, conf);
    // sparse compression is only enabled if less than 1% of docs have a value
    final int avgGap = 100;
    final int numDocs = atLeast(200);
    for (int i = random().nextInt(avgGap * 2); i >= 0; --i) {
        writer.addDocument(new Document());
    }
    final int maxNumValuesPerDoc = random().nextBoolean() ? 1 : TestUtil.nextInt(random(), 2, 5);
    for (int i = 0; i < numDocs; ++i) {
        Document doc = new Document();
        // single-valued
        long docValue = values[random().nextInt(values.length)];
        doc.add(new NumericDocValuesField("numeric", docValue));
        doc.add(new SortedDocValuesField("sorted", new BytesRef(Long.toString(docValue))));
        doc.add(new BinaryDocValuesField("binary", new BytesRef(Long.toString(docValue))));
        doc.add(new StoredField("value", docValue));
        // multi-valued
        final int numValues = TestUtil.nextInt(random(), 1, maxNumValuesPerDoc);
        for (int j = 0; j < numValues; ++j) {
            docValue = values[random().nextInt(values.length)];
            doc.add(new SortedNumericDocValuesField("sorted_numeric", docValue));
            doc.add(new SortedSetDocValuesField("sorted_set", new BytesRef(Long.toString(docValue))));
            doc.add(new StoredField("values", docValue));
        }
        writer.addDocument(doc);
        // add a gap
        for (int j = TestUtil.nextInt(random(), 0, avgGap * 2); j >= 0; --j) {
            writer.addDocument(new Document());
        }
    }
    if (random().nextBoolean()) {
        writer.forceMerge(1);
    }
    final IndexReader indexReader = writer.getReader();
    TestUtil.checkReader(indexReader);
    writer.close();
    for (LeafReaderContext context : indexReader.leaves()) {
        final LeafReader reader = context.reader();
        final NumericDocValues numeric = DocValues.getNumeric(reader, "numeric");
        final SortedDocValues sorted = DocValues.getSorted(reader, "sorted");
        final BinaryDocValues binary = DocValues.getBinary(reader, "binary");
        final SortedNumericDocValues sortedNumeric = DocValues.getSortedNumeric(reader, "sorted_numeric");
        final SortedSetDocValues sortedSet = DocValues.getSortedSet(reader, "sorted_set");
        for (int i = 0; i < reader.maxDoc(); ++i) {
            final Document doc = reader.document(i);
            final IndexableField valueField = doc.getField("value");
            final Long value = valueField == null ? null : valueField.numericValue().longValue();
            if (value == null) {
                assertTrue(numeric.docID() + " vs " + i, numeric.docID() < i);
            } else {
                assertEquals(i, numeric.nextDoc());
                assertEquals(i, binary.nextDoc());
                assertEquals(i, sorted.nextDoc());
                assertEquals(value.longValue(), numeric.longValue());
                assertTrue(sorted.ordValue() >= 0);
                assertEquals(new BytesRef(Long.toString(value)), sorted.lookupOrd(sorted.ordValue()));
                assertEquals(new BytesRef(Long.toString(value)), binary.binaryValue());
            }
            final IndexableField[] valuesFields = doc.getFields("values");
            if (valuesFields.length == 0) {
                assertTrue(sortedNumeric.docID() + " vs " + i, sortedNumeric.docID() < i);
            } else {
                final Set<Long> valueSet = new HashSet<>();
                for (IndexableField sf : valuesFields) {
                    valueSet.add(sf.numericValue().longValue());
                }
                assertEquals(i, sortedNumeric.nextDoc());
                assertEquals(valuesFields.length, sortedNumeric.docValueCount());
                for (int j = 0; j < sortedNumeric.docValueCount(); ++j) {
                    assertTrue(valueSet.contains(sortedNumeric.nextValue()));
                }
                assertEquals(i, sortedSet.nextDoc());
                int sortedSetCount = 0;
                while (true) {
                    long ord = sortedSet.nextOrd();
                    if (ord == SortedSetDocValues.NO_MORE_ORDS) {
                        break;
                    }
                    assertTrue(valueSet.contains(Long.parseLong(sortedSet.lookupOrd(ord).utf8ToString())));
                    sortedSetCount++;
                }
                assertEquals(valueSet.size(), sortedSetCount);
            }
        }
    }
    indexReader.close();
    dir.close();
}

Example 9 with SortedSetDocValuesField

use of org.apache.lucene.document.SortedSetDocValuesField in project lucene-solr by apache.

the class TestLucene54DocValuesFormat method testSortedSetAroundBlockSize.

@Slow
public void testSortedSetAroundBlockSize() throws IOException {
    final int frontier = 1 << Lucene54DocValuesFormat.DIRECT_MONOTONIC_BLOCK_SHIFT;
    for (int maxDoc = frontier - 1; maxDoc <= frontier + 1; ++maxDoc) {
        final Directory dir = newDirectory();
        IndexWriter w = new IndexWriter(dir, newIndexWriterConfig().setMergePolicy(newLogMergePolicy()));
        RAMFile buffer = new RAMFile();
        RAMOutputStream out = new RAMOutputStream(buffer, false);
        Document doc = new Document();
        SortedSetDocValuesField field1 = new SortedSetDocValuesField("sset", new BytesRef());
        doc.add(field1);
        SortedSetDocValuesField field2 = new SortedSetDocValuesField("sset", new BytesRef());
        doc.add(field2);
        for (int i = 0; i < maxDoc; ++i) {
            BytesRef s1 = new BytesRef(TestUtil.randomSimpleString(random(), 2));
            BytesRef s2 = new BytesRef(TestUtil.randomSimpleString(random(), 2));
            field1.setBytesValue(s1);
            field2.setBytesValue(s2);
            w.addDocument(doc);
            Set<BytesRef> set = new TreeSet<>(Arrays.asList(s1, s2));
            out.writeVInt(set.size());
            for (BytesRef ref : set) {
                out.writeVInt(ref.length);
                out.writeBytes(ref.bytes, ref.offset, ref.length);
            }
        }
        out.close();
        w.forceMerge(1);
        DirectoryReader r = DirectoryReader.open(w);
        w.close();
        LeafReader sr = getOnlyLeafReader(r);
        assertEquals(maxDoc, sr.maxDoc());
        SortedSetDocValues values = sr.getSortedSetDocValues("sset");
        assertNotNull(values);
        RAMInputStream in = new RAMInputStream("", buffer);
        BytesRefBuilder b = new BytesRefBuilder();
        for (int i = 0; i < maxDoc; ++i) {
            assertEquals(i, values.nextDoc());
            final int numValues = in.readVInt();
            for (int j = 0; j < numValues; ++j) {
                b.setLength(in.readVInt());
                b.grow(b.length());
                in.readBytes(b.bytes(), 0, b.length());
                assertEquals(b.get(), values.lookupOrd(values.nextOrd()));
            }
            assertEquals(SortedSetDocValues.NO_MORE_ORDS, values.nextOrd());
        }
        r.close();
        dir.close();
    }
}

Example 10 with SortedSetDocValuesField

use of org.apache.lucene.document.SortedSetDocValuesField in project lucene-solr by apache.

the class TestBackwardsCompatibility method addDoc.

private void addDoc(IndexWriter writer, int id) throws IOException {
    Document doc = new Document();
    doc.add(new TextField("content", "aaa", Field.Store.NO));
    doc.add(new StringField("id", Integer.toString(id), Field.Store.YES));
    FieldType customType2 = new FieldType(TextField.TYPE_STORED);
    customType2.setStoreTermVectors(true);
    customType2.setStoreTermVectorPositions(true);
    customType2.setStoreTermVectorOffsets(true);
    doc.add(new Field("autf8", "Lu𝄞ce𝅘𝅥𝅮ne  ☠ ab񕰗cd", customType2));
    doc.add(new Field("utf8", "Lu𝄞ce𝅘𝅥𝅮ne  ☠ ab񕰗cd", customType2));
    doc.add(new Field("content2", "here is more content with aaa aaa aaa", customType2));
    doc.add(new Field("fieⱷld", "field with non-ascii name", customType2));
    // add docvalues fields
    doc.add(new NumericDocValuesField("dvByte", (byte) id));
    byte[] bytes = new byte[] { (byte) (id >>> 24), (byte) (id >>> 16), (byte) (id >>> 8), (byte) id };
    BytesRef ref = new BytesRef(bytes);
    doc.add(new BinaryDocValuesField("dvBytesDerefFixed", ref));
    doc.add(new BinaryDocValuesField("dvBytesDerefVar", ref));
    doc.add(new SortedDocValuesField("dvBytesSortedFixed", ref));
    doc.add(new SortedDocValuesField("dvBytesSortedVar", ref));
    doc.add(new BinaryDocValuesField("dvBytesStraightFixed", ref));
    doc.add(new BinaryDocValuesField("dvBytesStraightVar", ref));
    doc.add(new DoubleDocValuesField("dvDouble", (double) id));
    doc.add(new FloatDocValuesField("dvFloat", (float) id));
    doc.add(new NumericDocValuesField("dvInt", id));
    doc.add(new NumericDocValuesField("dvLong", id));
    doc.add(new NumericDocValuesField("dvPacked", id));
    doc.add(new NumericDocValuesField("dvShort", (short) id));
    doc.add(new SortedSetDocValuesField("dvSortedSet", ref));
    doc.add(new SortedNumericDocValuesField("dvSortedNumeric", id));
    doc.add(new IntPoint("intPoint1d", id));
    doc.add(new IntPoint("intPoint2d", id, 2 * id));
    doc.add(new FloatPoint("floatPoint1d", (float) id));
    doc.add(new FloatPoint("floatPoint2d", (float) id, (float) 2 * id));
    doc.add(new LongPoint("longPoint1d", id));
    doc.add(new LongPoint("longPoint2d", id, 2 * id));
    doc.add(new DoublePoint("doublePoint1d", (double) id));
    doc.add(new DoublePoint("doublePoint2d", (double) id, (double) 2 * id));
    doc.add(new BinaryPoint("binaryPoint1d", bytes));
    doc.add(new BinaryPoint("binaryPoint2d", bytes, bytes));
    // a field with both offsets and term vectors for a cross-check
    FieldType customType3 = new FieldType(TextField.TYPE_STORED);
    customType3.setStoreTermVectors(true);
    customType3.setStoreTermVectorPositions(true);
    customType3.setStoreTermVectorOffsets(true);
    customType3.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
    doc.add(new Field("content5", "here is more content with aaa aaa aaa", customType3));
    // a field that omits only positions
    FieldType customType4 = new FieldType(TextField.TYPE_STORED);
    customType4.setStoreTermVectors(true);
    customType4.setStoreTermVectorPositions(false);
    customType4.setStoreTermVectorOffsets(true);
    customType4.setIndexOptions(IndexOptions.DOCS_AND_FREQS);
    doc.add(new Field("content6", "here is more content with aaa aaa aaa", customType4));
    // TODO: 
    //   index different norms types via similarity (we use a random one currently?!)
    //   remove any analyzer randomness, explicitly add payloads for certain fields.
    writer.addDocument(doc);
}