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

Example 11 with BinaryDocValuesField

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

the class TestBinaryDocValuesUpdates method testUpdateAllDeletedSegment.

public void testUpdateAllDeletedSegment() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random()));
    IndexWriter writer = new IndexWriter(dir, conf);
    Document doc = new Document();
    doc.add(new StringField("id", "doc", Store.NO));
    doc.add(new BinaryDocValuesField("f1", toBytes(1L)));
    writer.addDocument(doc);
    writer.addDocument(doc);
    writer.commit();
    // delete all docs in the first segment
    writer.deleteDocuments(new Term("id", "doc"));
    writer.addDocument(doc);
    writer.updateBinaryDocValue(new Term("id", "doc"), "f1", toBytes(2L));
    writer.close();
    DirectoryReader reader = DirectoryReader.open(dir);
    assertEquals(1, reader.leaves().size());
    BinaryDocValues bdv = reader.leaves().get(0).reader().getBinaryDocValues("f1");
    assertEquals(0, bdv.nextDoc());
    assertEquals(2L, getValue(bdv));
    reader.close();
    dir.close();
}

Example 12 with BinaryDocValuesField

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

the class TestBinaryDocValuesUpdates method testManyReopensAndFields.

public void testManyReopensAndFields() throws Exception {
    Directory dir = newDirectory();
    final Random random = random();
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random));
    LogMergePolicy lmp = newLogMergePolicy();
    // merge often
    lmp.setMergeFactor(3);
    conf.setMergePolicy(lmp);
    IndexWriter writer = new IndexWriter(dir, conf);
    final boolean isNRT = random.nextBoolean();
    DirectoryReader reader;
    if (isNRT) {
        reader = DirectoryReader.open(writer);
    } else {
        writer.commit();
        reader = DirectoryReader.open(dir);
    }
    //System.out.println("TEST: isNRT=" + isNRT);
    // 3-7
    final int numFields = random.nextInt(4) + 3;
    final long[] fieldValues = new long[numFields];
    for (int i = 0; i < fieldValues.length; i++) {
        fieldValues[i] = 1;
    }
    int numRounds = atLeast(15);
    int docID = 0;
    for (int i = 0; i < numRounds; i++) {
        int numDocs = atLeast(5);
        //System.out.println("[" + Thread.currentThread().getName() + "]: round=" + i + ", numDocs=" + numDocs);
        for (int j = 0; j < numDocs; j++) {
            Document doc = new Document();
            doc.add(new StringField("id", "doc-" + docID, Store.NO));
            // update key
            doc.add(new StringField("key", "all", Store.NO));
            // add all fields with their current value
            for (int f = 0; f < fieldValues.length; f++) {
                doc.add(new BinaryDocValuesField("f" + f, toBytes(fieldValues[f])));
            }
            writer.addDocument(doc);
            ++docID;
        }
        int fieldIdx = random.nextInt(fieldValues.length);
        String updateField = "f" + fieldIdx;
        //      System.out.println("[" + Thread.currentThread().getName() + "]: updated field '" + updateField + "' to value " + fieldValues[fieldIdx]);
        writer.updateBinaryDocValue(new Term("key", "all"), updateField, toBytes(++fieldValues[fieldIdx]));
        if (random.nextDouble() < 0.2) {
            // might also delete an already deleted document, ok!
            int deleteDoc = random.nextInt(docID);
            writer.deleteDocuments(new Term("id", "doc-" + deleteDoc));
        //        System.out.println("[" + Thread.currentThread().getName() + "]: deleted document: doc-" + deleteDoc);
        }
        // verify reader
        if (!isNRT) {
            writer.commit();
        }
        //      System.out.println("[" + Thread.currentThread().getName() + "]: reopen reader: " + reader);
        DirectoryReader newReader = DirectoryReader.openIfChanged(reader);
        assertNotNull(newReader);
        reader.close();
        reader = newReader;
        //      System.out.println("[" + Thread.currentThread().getName() + "]: reopened reader: " + reader);
        // we delete at most one document per round
        assertTrue(reader.numDocs() > 0);
        for (LeafReaderContext context : reader.leaves()) {
            LeafReader r = context.reader();
            //        System.out.println(((SegmentReader) r).getSegmentName());
            Bits liveDocs = r.getLiveDocs();
            for (int field = 0; field < fieldValues.length; field++) {
                String f = "f" + field;
                BinaryDocValues bdv = r.getBinaryDocValues(f);
                assertNotNull(bdv);
                int maxDoc = r.maxDoc();
                for (int doc = 0; doc < maxDoc; doc++) {
                    if (liveDocs == null || liveDocs.get(doc)) {
                        assertEquals(doc, bdv.advance(doc));
                        assertEquals("invalid value for doc=" + doc + ", field=" + f + ", reader=" + r, fieldValues[field], getValue(bdv));
                    }
                }
            }
        }
    //      System.out.println();
    }
    writer.close();
    IOUtils.close(reader, dir);
}

Example 13 with BinaryDocValuesField

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

the class TestBinaryDocValuesUpdates method testStressMultiThreading.

public void testStressMultiThreading() throws Exception {
    final Directory dir = newDirectory();
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random()));
    final IndexWriter writer = new IndexWriter(dir, conf);
    // create index
    final int numFields = TestUtil.nextInt(random(), 1, 4);
    final int numDocs = atLeast(2000);
    for (int i = 0; i < numDocs; i++) {
        Document doc = new Document();
        doc.add(new StringField("id", "doc" + i, Store.NO));
        double group = random().nextDouble();
        String g;
        if (group < 0.1)
            g = "g0";
        else if (group < 0.5)
            g = "g1";
        else if (group < 0.8)
            g = "g2";
        else
            g = "g3";
        doc.add(new StringField("updKey", g, Store.NO));
        for (int j = 0; j < numFields; j++) {
            long value = random().nextInt();
            doc.add(new BinaryDocValuesField("f" + j, toBytes(value)));
            // control, always updated to f * 2
            doc.add(new BinaryDocValuesField("cf" + j, toBytes(value * 2)));
        }
        writer.addDocument(doc);
    }
    final int numThreads = TestUtil.nextInt(random(), 3, 6);
    final CountDownLatch done = new CountDownLatch(numThreads);
    final AtomicInteger numUpdates = new AtomicInteger(atLeast(100));
    // same thread updates a field as well as reopens
    Thread[] threads = new Thread[numThreads];
    for (int i = 0; i < threads.length; i++) {
        threads[i] = new Thread("UpdateThread-" + i) {

            @Override
            public void run() {
                DirectoryReader reader = null;
                boolean success = false;
                try {
                    Random random = random();
                    while (numUpdates.getAndDecrement() > 0) {
                        double group = random.nextDouble();
                        Term t;
                        if (group < 0.1)
                            t = new Term("updKey", "g0");
                        else if (group < 0.5)
                            t = new Term("updKey", "g1");
                        else if (group < 0.8)
                            t = new Term("updKey", "g2");
                        else
                            t = new Term("updKey", "g3");
                        final int field = random().nextInt(numFields);
                        final String f = "f" + field;
                        final String cf = "cf" + field;
                        //              System.out.println("[" + Thread.currentThread().getName() + "] numUpdates=" + numUpdates + " updateTerm=" + t + " field=" + field);
                        long updValue = random.nextInt();
                        writer.updateDocValues(t, new BinaryDocValuesField(f, toBytes(updValue)), new BinaryDocValuesField(cf, toBytes(updValue * 2)));
                        if (random.nextDouble() < 0.2) {
                            // delete a random document
                            int doc = random.nextInt(numDocs);
                            //                System.out.println("[" + Thread.currentThread().getName() + "] deleteDoc=doc" + doc);
                            writer.deleteDocuments(new Term("id", "doc" + doc));
                        }
                        if (random.nextDouble() < 0.05) {
                            // commit every 20 updates on average
                            //                  System.out.println("[" + Thread.currentThread().getName() + "] commit");
                            writer.commit();
                        }
                        if (random.nextDouble() < 0.1) {
                            // reopen NRT reader (apply updates), on average once every 10 updates
                            if (reader == null) {
                                //                  System.out.println("[" + Thread.currentThread().getName() + "] open NRT");
                                reader = DirectoryReader.open(writer);
                            } else {
                                //                  System.out.println("[" + Thread.currentThread().getName() + "] reopen NRT");
                                DirectoryReader r2 = DirectoryReader.openIfChanged(reader, writer);
                                if (r2 != null) {
                                    reader.close();
                                    reader = r2;
                                }
                            }
                        }
                    }
                    //            System.out.println("[" + Thread.currentThread().getName() + "] DONE");
                    success = true;
                } catch (IOException e) {
                    throw new RuntimeException(e);
                } finally {
                    if (reader != null) {
                        try {
                            reader.close();
                        } catch (IOException e) {
                            if (success) {
                                // suppress this exception only if there was another exception
                                throw new RuntimeException(e);
                            }
                        }
                    }
                    done.countDown();
                }
            }
        };
    }
    for (Thread t : threads) t.start();
    done.await();
    writer.close();
    DirectoryReader reader = DirectoryReader.open(dir);
    for (LeafReaderContext context : reader.leaves()) {
        LeafReader r = context.reader();
        for (int i = 0; i < numFields; i++) {
            BinaryDocValues bdv = r.getBinaryDocValues("f" + i);
            BinaryDocValues control = r.getBinaryDocValues("cf" + i);
            Bits liveDocs = r.getLiveDocs();
            for (int j = 0; j < r.maxDoc(); j++) {
                if (liveDocs == null || liveDocs.get(j)) {
                    assertEquals(j, bdv.advance(j));
                    assertEquals(j, control.advance(j));
                    assertEquals(getValue(control), getValue(bdv) * 2);
                }
            }
        }
    }
    reader.close();
    dir.close();
}

Example 14 with BinaryDocValuesField

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

the class TestNumericDocValuesUpdates method testMultipleDocValuesTypes.

@Test
public void testMultipleDocValuesTypes() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig conf = newIndexWriterConfig(new MockAnalyzer(random()));
    // prevent merges
    conf.setMaxBufferedDocs(10);
    IndexWriter writer = new IndexWriter(dir, conf);
    for (int i = 0; i < 4; i++) {
        Document doc = new Document();
        doc.add(new StringField("dvUpdateKey", "dv", Store.NO));
        doc.add(new NumericDocValuesField("ndv", i));
        doc.add(new BinaryDocValuesField("bdv", new BytesRef(Integer.toString(i))));
        doc.add(new SortedDocValuesField("sdv", new BytesRef(Integer.toString(i))));
        doc.add(new SortedSetDocValuesField("ssdv", new BytesRef(Integer.toString(i))));
        doc.add(new SortedSetDocValuesField("ssdv", new BytesRef(Integer.toString(i * 2))));
        writer.addDocument(doc);
    }
    writer.commit();
    // update all docs' ndv field
    writer.updateNumericDocValue(new Term("dvUpdateKey", "dv"), "ndv", 17L);
    writer.close();
    final DirectoryReader reader = DirectoryReader.open(dir);
    LeafReader r = reader.leaves().get(0).reader();
    NumericDocValues ndv = r.getNumericDocValues("ndv");
    BinaryDocValues bdv = r.getBinaryDocValues("bdv");
    SortedDocValues sdv = r.getSortedDocValues("sdv");
    SortedSetDocValues ssdv = r.getSortedSetDocValues("ssdv");
    for (int i = 0; i < r.maxDoc(); i++) {
        assertEquals(i, ndv.nextDoc());
        assertEquals(17, ndv.longValue());
        assertEquals(i, bdv.nextDoc());
        BytesRef term = bdv.binaryValue();
        assertEquals(new BytesRef(Integer.toString(i)), term);
        assertEquals(i, sdv.nextDoc());
        term = sdv.binaryValue();
        assertEquals(new BytesRef(Integer.toString(i)), term);
        assertEquals(i, ssdv.nextDoc());
        long ord = ssdv.nextOrd();
        term = ssdv.lookupOrd(ord);
        assertEquals(i, Integer.parseInt(term.utf8ToString()));
        if (i != 0) {
            ord = ssdv.nextOrd();
            term = ssdv.lookupOrd(ord);
            assertEquals(i * 2, Integer.parseInt(term.utf8ToString()));
        }
        assertEquals(SortedSetDocValues.NO_MORE_ORDS, ssdv.nextOrd());
    }
    reader.close();
    dir.close();
}

Example 15 with BinaryDocValuesField

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

the class TestSearchAfter method setUp.

@Override
public void setUp() throws Exception {
    super.setUp();
    allSortFields = new ArrayList<>(Arrays.asList(new SortField[] { new SortField("int", SortField.Type.INT, false), new SortField("long", SortField.Type.LONG, false), new SortField("float", SortField.Type.FLOAT, false), new SortField("double", SortField.Type.DOUBLE, false), new SortField("bytes", SortField.Type.STRING, false), new SortField("bytesval", SortField.Type.STRING_VAL, false), new SortField("int", SortField.Type.INT, true), new SortField("long", SortField.Type.LONG, true), new SortField("float", SortField.Type.FLOAT, true), new SortField("double", SortField.Type.DOUBLE, true), new SortField("bytes", SortField.Type.STRING, true), new SortField("bytesval", SortField.Type.STRING_VAL, true), SortField.FIELD_SCORE, SortField.FIELD_DOC }));
    // Also test missing first / last for the "string" sorts:
    for (String field : new String[] { "bytes", "sortedbytesdocvalues" }) {
        for (int rev = 0; rev < 2; rev++) {
            boolean reversed = rev == 0;
            SortField sf = new SortField(field, SortField.Type.STRING, reversed);
            sf.setMissingValue(SortField.STRING_FIRST);
            allSortFields.add(sf);
            sf = new SortField(field, SortField.Type.STRING, reversed);
            sf.setMissingValue(SortField.STRING_LAST);
            allSortFields.add(sf);
        }
    }
    // Also test missing first / last for the "string_val" sorts:
    for (String field : new String[] { "sortedbytesdocvaluesval", "straightbytesdocvalues" }) {
        for (int rev = 0; rev < 2; rev++) {
            boolean reversed = rev == 0;
            SortField sf = new SortField(field, SortField.Type.STRING_VAL, reversed);
            sf.setMissingValue(SortField.STRING_FIRST);
            allSortFields.add(sf);
            sf = new SortField(field, SortField.Type.STRING_VAL, reversed);
            sf.setMissingValue(SortField.STRING_LAST);
            allSortFields.add(sf);
        }
    }
    int limit = allSortFields.size();
    for (int i = 0; i < limit; i++) {
        SortField sf = allSortFields.get(i);
        if (sf.getType() == SortField.Type.INT) {
            SortField sf2 = new SortField(sf.getField(), SortField.Type.INT, sf.getReverse());
            sf2.setMissingValue(random().nextInt());
            allSortFields.add(sf2);
        } else if (sf.getType() == SortField.Type.LONG) {
            SortField sf2 = new SortField(sf.getField(), SortField.Type.LONG, sf.getReverse());
            sf2.setMissingValue(random().nextLong());
            allSortFields.add(sf2);
        } else if (sf.getType() == SortField.Type.FLOAT) {
            SortField sf2 = new SortField(sf.getField(), SortField.Type.FLOAT, sf.getReverse());
            sf2.setMissingValue(random().nextFloat());
            allSortFields.add(sf2);
        } else if (sf.getType() == SortField.Type.DOUBLE) {
            SortField sf2 = new SortField(sf.getField(), SortField.Type.DOUBLE, sf.getReverse());
            sf2.setMissingValue(random().nextDouble());
            allSortFields.add(sf2);
        }
    }
    dir = newDirectory();
    RandomIndexWriter iw = new RandomIndexWriter(random(), dir);
    int numDocs = atLeast(200);
    Random r = random();
    for (int i = 0; i < numDocs; i++) {
        List<Field> fields = new ArrayList<>();
        fields.add(newTextField("english", English.intToEnglish(i), Field.Store.NO));
        fields.add(newTextField("oddeven", (i % 2 == 0) ? "even" : "odd", Field.Store.NO));
        fields.add(new NumericDocValuesField("byte", (byte) r.nextInt()));
        fields.add(new NumericDocValuesField("short", (short) r.nextInt()));
        fields.add(new NumericDocValuesField("int", r.nextInt()));
        fields.add(new NumericDocValuesField("long", r.nextLong()));
        fields.add(new FloatDocValuesField("float", r.nextFloat()));
        fields.add(new DoubleDocValuesField("double", r.nextDouble()));
        fields.add(new SortedDocValuesField("bytes", new BytesRef(TestUtil.randomRealisticUnicodeString(random()))));
        fields.add(new BinaryDocValuesField("bytesval", new BytesRef(TestUtil.randomRealisticUnicodeString(random()))));
        Document document = new Document();
        document.add(new StoredField("id", "" + i));
        if (VERBOSE) {
            System.out.println("  add doc id=" + i);
        }
        for (Field field : fields) {
            // So we are sometimes missing that field:
            if (random().nextInt(5) != 4) {
                document.add(field);
                if (VERBOSE) {
                    System.out.println("    " + field);
                }
            }
        }
        iw.addDocument(document);
        if (random().nextInt(50) == 17) {
            iw.commit();
        }
    }
    reader = iw.getReader();
    iw.close();
    searcher = newSearcher(reader);
    if (VERBOSE) {
        System.out.println("  searcher=" + searcher);
    }
}