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

Example 16 with BinaryDocValues

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

the class BytesRefFieldComparatorSource method newComparator.

@Override
public FieldComparator<?> newComparator(String fieldname, int numHits, int sortPos, boolean reversed) {
    assert indexFieldData == null || fieldname.equals(indexFieldData.getFieldName());
    final boolean sortMissingLast = sortMissingLast(missingValue) ^ reversed;
    final BytesRef missingBytes = (BytesRef) missingObject(missingValue, reversed);
    if (indexFieldData instanceof IndexOrdinalsFieldData) {
        return new FieldComparator.TermOrdValComparator(numHits, null, sortMissingLast) {

            @Override
            protected SortedDocValues getSortedDocValues(LeafReaderContext context, String field) throws IOException {
                final RandomAccessOrds values = ((IndexOrdinalsFieldData) indexFieldData).load(context).getOrdinalsValues();
                final SortedDocValues selectedValues;
                if (nested == null) {
                    selectedValues = sortMode.select(values);
                } else {
                    final BitSet rootDocs = nested.rootDocs(context);
                    final DocIdSetIterator innerDocs = nested.innerDocs(context);
                    selectedValues = sortMode.select(values, rootDocs, innerDocs);
                }
                if (sortMissingFirst(missingValue) || sortMissingLast(missingValue)) {
                    return selectedValues;
                } else {
                    return new ReplaceMissing(selectedValues, missingBytes);
                }
            }

            @Override
            public void setScorer(Scorer scorer) {
                BytesRefFieldComparatorSource.this.setScorer(scorer);
            }
        };
    }
    final BytesRef nullPlaceHolder = new BytesRef();
    final BytesRef nonNullMissingBytes = missingBytes == null ? nullPlaceHolder : missingBytes;
    return new FieldComparator.TermValComparator(numHits, null, sortMissingLast) {

        @Override
        protected BinaryDocValues getBinaryDocValues(LeafReaderContext context, String field) throws IOException {
            final SortedBinaryDocValues values = getValues(context);
            final BinaryDocValues selectedValues;
            if (nested == null) {
                selectedValues = sortMode.select(values, nonNullMissingBytes);
            } else {
                final BitSet rootDocs = nested.rootDocs(context);
                final DocIdSetIterator innerDocs = nested.innerDocs(context);
                selectedValues = sortMode.select(values, nonNullMissingBytes, rootDocs, innerDocs, context.reader().maxDoc());
            }
            return selectedValues;
        }

        @Override
        protected Bits getDocsWithField(LeafReaderContext context, String field) throws IOException {
            return new Bits.MatchAllBits(context.reader().maxDoc());
        }

        @Override
        protected boolean isNull(int doc, BytesRef term) {
            return term == nullPlaceHolder;
        }

        @Override
        public void setScorer(Scorer scorer) {
            BytesRefFieldComparatorSource.this.setScorer(scorer);
        }
    };
}

Example 17 with BinaryDocValues

use of org.apache.lucene.index.BinaryDocValues in project lucene-solr by apache.

the class TestFieldCache method test.

public void test() throws IOException {
    FieldCache cache = FieldCache.DEFAULT;
    NumericDocValues doubles = cache.getNumerics(reader, "theDouble", FieldCache.DOUBLE_POINT_PARSER);
    for (int i = 0; i < NUM_DOCS; i++) {
        assertEquals(i, doubles.nextDoc());
        assertEquals(Double.doubleToLongBits(Double.MAX_VALUE - i), doubles.longValue());
    }
    NumericDocValues longs = cache.getNumerics(reader, "theLong", FieldCache.LONG_POINT_PARSER);
    for (int i = 0; i < NUM_DOCS; i++) {
        assertEquals(i, longs.nextDoc());
        assertEquals(Long.MAX_VALUE - i, longs.longValue());
    }
    NumericDocValues ints = cache.getNumerics(reader, "theInt", FieldCache.INT_POINT_PARSER);
    for (int i = 0; i < NUM_DOCS; i++) {
        assertEquals(i, ints.nextDoc());
        assertEquals(Integer.MAX_VALUE - i, ints.longValue());
    }
    NumericDocValues floats = cache.getNumerics(reader, "theFloat", FieldCache.FLOAT_POINT_PARSER);
    for (int i = 0; i < NUM_DOCS; i++) {
        assertEquals(i, floats.nextDoc());
        assertEquals(Float.floatToIntBits(Float.MAX_VALUE - i), floats.longValue());
    }
    Bits docsWithField = cache.getDocsWithField(reader, "theLong", FieldCache.LONG_POINT_PARSER);
    assertSame("Second request to cache return same array", docsWithField, cache.getDocsWithField(reader, "theLong", FieldCache.LONG_POINT_PARSER));
    assertTrue("docsWithField(theLong) must be class Bits.MatchAllBits", docsWithField instanceof Bits.MatchAllBits);
    assertTrue("docsWithField(theLong) Size: " + docsWithField.length() + " is not: " + NUM_DOCS, docsWithField.length() == NUM_DOCS);
    for (int i = 0; i < docsWithField.length(); i++) {
        assertTrue(docsWithField.get(i));
    }
    docsWithField = cache.getDocsWithField(reader, "sparse", FieldCache.INT_POINT_PARSER);
    assertSame("Second request to cache return same array", docsWithField, cache.getDocsWithField(reader, "sparse", FieldCache.INT_POINT_PARSER));
    assertFalse("docsWithField(sparse) must not be class Bits.MatchAllBits", docsWithField instanceof Bits.MatchAllBits);
    assertTrue("docsWithField(sparse) Size: " + docsWithField.length() + " is not: " + NUM_DOCS, docsWithField.length() == NUM_DOCS);
    for (int i = 0; i < docsWithField.length(); i++) {
        assertEquals(i % 2 == 0, docsWithField.get(i));
    }
    // getTermsIndex
    SortedDocValues termsIndex = cache.getTermsIndex(reader, "theRandomUnicodeString");
    for (int i = 0; i < NUM_DOCS; i++) {
        final String s;
        if (i > termsIndex.docID()) {
            termsIndex.advance(i);
        }
        if (i == termsIndex.docID()) {
            s = termsIndex.binaryValue().utf8ToString();
        } else {
            s = null;
        }
        assertTrue("for doc " + i + ": " + s + " does not equal: " + unicodeStrings[i], unicodeStrings[i] == null || unicodeStrings[i].equals(s));
    }
    int nTerms = termsIndex.getValueCount();
    TermsEnum tenum = termsIndex.termsEnum();
    for (int i = 0; i < nTerms; i++) {
        BytesRef val1 = BytesRef.deepCopyOf(tenum.next());
        final BytesRef val = termsIndex.lookupOrd(i);
        // System.out.println("i="+i);
        assertEquals(val, val1);
    }
    // seek the enum around (note this isn't a great test here)
    int num = atLeast(100);
    for (int i = 0; i < num; i++) {
        int k = random().nextInt(nTerms);
        final BytesRef val = BytesRef.deepCopyOf(termsIndex.lookupOrd(k));
        assertEquals(TermsEnum.SeekStatus.FOUND, tenum.seekCeil(val));
        assertEquals(val, tenum.term());
    }
    for (int i = 0; i < nTerms; i++) {
        final BytesRef val = BytesRef.deepCopyOf(termsIndex.lookupOrd(i));
        assertEquals(TermsEnum.SeekStatus.FOUND, tenum.seekCeil(val));
        assertEquals(val, tenum.term());
    }
    // test bad field
    termsIndex = cache.getTermsIndex(reader, "bogusfield");
    // getTerms
    BinaryDocValues terms = cache.getTerms(reader, "theRandomUnicodeString");
    for (int i = 0; i < NUM_DOCS; i++) {
        if (terms.docID() < i) {
            terms.nextDoc();
        }
        if (terms.docID() == i) {
            assertEquals(unicodeStrings[i], terms.binaryValue().utf8ToString());
        } else {
            assertNull(unicodeStrings[i]);
        }
    }
    // test bad field
    terms = cache.getTerms(reader, "bogusfield");
    // getDocTermOrds
    SortedSetDocValues termOrds = cache.getDocTermOrds(reader, "theRandomUnicodeMultiValuedField", null);
    int numEntries = cache.getCacheEntries().length;
    // ask for it again, and check that we didnt create any additional entries:
    termOrds = cache.getDocTermOrds(reader, "theRandomUnicodeMultiValuedField", null);
    assertEquals(numEntries, cache.getCacheEntries().length);
    for (int i = 0; i < NUM_DOCS; i++) {
        // This will remove identical terms. A DocTermOrds doesn't return duplicate ords for a docId
        List<BytesRef> values = new ArrayList<>(new LinkedHashSet<>(Arrays.asList(multiValued[i])));
        for (BytesRef v : values) {
            if (v == null) {
                // why does this test use null values... instead of an empty list: confusing
                break;
            }
            if (i > termOrds.docID()) {
                assertEquals(i, termOrds.nextDoc());
            }
            long ord = termOrds.nextOrd();
            assert ord != SortedSetDocValues.NO_MORE_ORDS;
            BytesRef scratch = termOrds.lookupOrd(ord);
            assertEquals(v, scratch);
        }
        if (i == termOrds.docID()) {
            assertEquals(SortedSetDocValues.NO_MORE_ORDS, termOrds.nextOrd());
        }
    }
    // test bad field
    termOrds = cache.getDocTermOrds(reader, "bogusfield", null);
    assertTrue(termOrds.getValueCount() == 0);
    FieldCache.DEFAULT.purgeByCacheKey(reader.getCoreCacheHelper().getKey());
}

Example 18 with BinaryDocValues

use of org.apache.lucene.index.BinaryDocValues in project lucene-solr by apache.

the class TestFieldCache method testNonexistantFields.

public void testNonexistantFields() throws Exception {
    Directory dir = newDirectory();
    RandomIndexWriter iw = new RandomIndexWriter(random(), dir);
    Document doc = new Document();
    iw.addDocument(doc);
    DirectoryReader ir = iw.getReader();
    iw.close();
    LeafReader ar = getOnlyLeafReader(ir);
    final FieldCache cache = FieldCache.DEFAULT;
    cache.purgeAllCaches();
    assertEquals(0, cache.getCacheEntries().length);
    NumericDocValues ints = cache.getNumerics(ar, "bogusints", FieldCache.INT_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, ints.nextDoc());
    NumericDocValues longs = cache.getNumerics(ar, "boguslongs", FieldCache.LONG_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, longs.nextDoc());
    NumericDocValues floats = cache.getNumerics(ar, "bogusfloats", FieldCache.FLOAT_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, floats.nextDoc());
    NumericDocValues doubles = cache.getNumerics(ar, "bogusdoubles", FieldCache.DOUBLE_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, doubles.nextDoc());
    BinaryDocValues binaries = cache.getTerms(ar, "bogusterms");
    assertEquals(NO_MORE_DOCS, binaries.nextDoc());
    SortedDocValues sorted = cache.getTermsIndex(ar, "bogustermsindex");
    assertEquals(NO_MORE_DOCS, sorted.nextDoc());
    SortedSetDocValues sortedSet = cache.getDocTermOrds(ar, "bogusmultivalued", null);
    assertEquals(NO_MORE_DOCS, sortedSet.nextDoc());
    Bits bits = cache.getDocsWithField(ar, "bogusbits", null);
    assertFalse(bits.get(0));
    // check that we cached nothing
    assertEquals(0, cache.getCacheEntries().length);
    ir.close();
    dir.close();
}

Example 19 with BinaryDocValues

use of org.apache.lucene.index.BinaryDocValues in project lucene-solr by apache.

the class TestFieldCache method testNonIndexedFields.

public void testNonIndexedFields() throws Exception {
    Directory dir = newDirectory();
    RandomIndexWriter iw = new RandomIndexWriter(random(), dir);
    Document doc = new Document();
    doc.add(new StoredField("bogusbytes", "bogus"));
    doc.add(new StoredField("bogusshorts", "bogus"));
    doc.add(new StoredField("bogusints", "bogus"));
    doc.add(new StoredField("boguslongs", "bogus"));
    doc.add(new StoredField("bogusfloats", "bogus"));
    doc.add(new StoredField("bogusdoubles", "bogus"));
    doc.add(new StoredField("bogusterms", "bogus"));
    doc.add(new StoredField("bogustermsindex", "bogus"));
    doc.add(new StoredField("bogusmultivalued", "bogus"));
    doc.add(new StoredField("bogusbits", "bogus"));
    iw.addDocument(doc);
    DirectoryReader ir = iw.getReader();
    iw.close();
    LeafReader ar = getOnlyLeafReader(ir);
    final FieldCache cache = FieldCache.DEFAULT;
    cache.purgeAllCaches();
    assertEquals(0, cache.getCacheEntries().length);
    NumericDocValues ints = cache.getNumerics(ar, "bogusints", FieldCache.INT_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, ints.nextDoc());
    NumericDocValues longs = cache.getNumerics(ar, "boguslongs", FieldCache.LONG_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, longs.nextDoc());
    NumericDocValues floats = cache.getNumerics(ar, "bogusfloats", FieldCache.FLOAT_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, floats.nextDoc());
    NumericDocValues doubles = cache.getNumerics(ar, "bogusdoubles", FieldCache.DOUBLE_POINT_PARSER);
    assertEquals(NO_MORE_DOCS, doubles.nextDoc());
    BinaryDocValues binaries = cache.getTerms(ar, "bogusterms");
    assertEquals(NO_MORE_DOCS, binaries.nextDoc());
    SortedDocValues sorted = cache.getTermsIndex(ar, "bogustermsindex");
    assertEquals(NO_MORE_DOCS, sorted.nextDoc());
    SortedSetDocValues sortedSet = cache.getDocTermOrds(ar, "bogusmultivalued", null);
    assertEquals(NO_MORE_DOCS, sortedSet.nextDoc());
    Bits bits = cache.getDocsWithField(ar, "bogusbits", null);
    assertFalse(bits.get(0));
    // check that we cached nothing
    assertEquals(0, cache.getCacheEntries().length);
    ir.close();
    dir.close();
}

Example 20 with BinaryDocValues

use of org.apache.lucene.index.BinaryDocValues in project lucene-solr by apache.

the class SimpleTextDocValuesWriter method doAddBinaryField.

private void doAddBinaryField(FieldInfo field, DocValuesProducer valuesProducer) throws IOException {
    int maxLength = 0;
    BinaryDocValues values = valuesProducer.getBinary(field);
    for (int doc = values.nextDoc(); doc != DocIdSetIterator.NO_MORE_DOCS; doc = values.nextDoc()) {
        maxLength = Math.max(maxLength, values.binaryValue().length);
    }
    writeFieldEntry(field, DocValuesType.BINARY);
    // write maxLength
    SimpleTextUtil.write(data, MAXLENGTH);
    SimpleTextUtil.write(data, Integer.toString(maxLength), scratch);
    SimpleTextUtil.writeNewline(data);
    int maxBytesLength = Long.toString(maxLength).length();
    StringBuilder sb = new StringBuilder();
    for (int i = 0; i < maxBytesLength; i++) {
        sb.append('0');
    }
    // write our pattern for encoding lengths
    SimpleTextUtil.write(data, PATTERN);
    SimpleTextUtil.write(data, sb.toString(), scratch);
    SimpleTextUtil.writeNewline(data);
    final DecimalFormat encoder = new DecimalFormat(sb.toString(), new DecimalFormatSymbols(Locale.ROOT));
    values = valuesProducer.getBinary(field);
    int numDocsWritten = 0;
    for (int i = 0; i < numDocs; ++i) {
        if (values.docID() < i) {
            values.nextDoc();
            assert values.docID() >= i;
        }
        // write length
        final int length = values.docID() != i ? 0 : values.binaryValue().length;
        SimpleTextUtil.write(data, LENGTH);
        SimpleTextUtil.write(data, encoder.format(length), scratch);
        SimpleTextUtil.writeNewline(data);
        // because it escapes:
        if (values.docID() == i) {
            BytesRef value = values.binaryValue();
            data.writeBytes(value.bytes, value.offset, value.length);
        }
        // pad to fit
        for (int j = length; j < maxLength; j++) {
            data.writeByte((byte) ' ');
        }
        SimpleTextUtil.writeNewline(data);
        if (values.docID() != i) {
            SimpleTextUtil.write(data, "F", scratch);
        } else {
            SimpleTextUtil.write(data, "T", scratch);
        }
        SimpleTextUtil.writeNewline(data);
        numDocsWritten++;
    }
    assert numDocs == numDocsWritten;
}