Examples with Sort org.apache.lucene.search.Sort used on opensource projects

Example 76 with Sort

use of org.apache.lucene.search.Sort in project lucene-solr by apache.

the class TestIndexSorting method testBasicMultiValuedString.

public void testBasicMultiValuedString() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(new MockAnalyzer(random()));
    Sort indexSort = new Sort(new SortedSetSortField("foo", false));
    iwc.setIndexSort(indexSort);
    IndexWriter w = new IndexWriter(dir, iwc);
    Document doc = new Document();
    doc.add(new NumericDocValuesField("id", 3));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("zzz")));
    w.addDocument(doc);
    // so we get more than one segment, so that forceMerge actually does merge, since we only get a sorted segment by merging:
    w.commit();
    doc = new Document();
    doc.add(new NumericDocValuesField("id", 1));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("aaa")));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("zzz")));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("bcg")));
    w.addDocument(doc);
    w.commit();
    doc = new Document();
    doc.add(new NumericDocValuesField("id", 2));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("mmm")));
    doc.add(new SortedSetDocValuesField("foo", new BytesRef("pppp")));
    w.addDocument(doc);
    w.forceMerge(1);
    DirectoryReader r = DirectoryReader.open(w);
    LeafReader leaf = getOnlyLeafReader(r);
    assertEquals(3, leaf.maxDoc());
    NumericDocValues values = leaf.getNumericDocValues("id");
    assertEquals(0, values.nextDoc());
    assertEquals(1l, values.longValue());
    assertEquals(1, values.nextDoc());
    assertEquals(2l, values.longValue());
    assertEquals(2, values.nextDoc());
    assertEquals(3l, values.longValue());
    r.close();
    w.close();
    dir.close();
}

Example 77 with Sort

use of org.apache.lucene.search.Sort in project lucene-solr by apache.

the class TestIndexSorting method testMissingFloatLast.

public void testMissingFloatLast() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(new MockAnalyzer(random()));
    SortField sortField = new SortField("foo", SortField.Type.FLOAT);
    sortField.setMissingValue(Float.POSITIVE_INFINITY);
    Sort indexSort = new Sort(sortField);
    iwc.setIndexSort(indexSort);
    IndexWriter w = new IndexWriter(dir, iwc);
    Document doc = new Document();
    doc.add(new FloatDocValuesField("foo", 18.0f));
    w.addDocument(doc);
    // so we get more than one segment, so that forceMerge actually does merge, since we only get a sorted segment by merging:
    w.commit();
    // missing
    w.addDocument(new Document());
    w.commit();
    doc = new Document();
    doc.add(new FloatDocValuesField("foo", 7.0f));
    w.addDocument(doc);
    w.forceMerge(1);
    DirectoryReader r = DirectoryReader.open(w);
    LeafReader leaf = getOnlyLeafReader(r);
    assertEquals(3, leaf.maxDoc());
    NumericDocValues values = leaf.getNumericDocValues("foo");
    assertEquals(0, values.nextDoc());
    assertEquals(7.0f, Float.intBitsToFloat((int) values.longValue()), 0.0f);
    assertEquals(1, values.nextDoc());
    assertEquals(18.0f, Float.intBitsToFloat((int) values.longValue()), 0.0f);
    assertEquals(NO_MORE_DOCS, values.nextDoc());
    r.close();
    w.close();
    dir.close();
}

Example 78 with Sort

use of org.apache.lucene.search.Sort in project lucene-solr by apache.

the class TestIndexSorting method testMissingLongLast.

public void testMissingLongLast() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(new MockAnalyzer(random()));
    SortField sortField = new SortField("foo", SortField.Type.LONG);
    sortField.setMissingValue(Long.valueOf(Long.MAX_VALUE));
    Sort indexSort = new Sort(sortField);
    iwc.setIndexSort(indexSort);
    IndexWriter w = new IndexWriter(dir, iwc);
    Document doc = new Document();
    doc.add(new NumericDocValuesField("foo", 18));
    w.addDocument(doc);
    // so we get more than one segment, so that forceMerge actually does merge, since we only get a sorted segment by merging:
    w.commit();
    // missing
    w.addDocument(new Document());
    w.commit();
    doc = new Document();
    doc.add(new NumericDocValuesField("foo", 7));
    w.addDocument(doc);
    w.forceMerge(1);
    DirectoryReader r = DirectoryReader.open(w);
    LeafReader leaf = getOnlyLeafReader(r);
    assertEquals(3, leaf.maxDoc());
    NumericDocValues values = leaf.getNumericDocValues("foo");
    assertEquals(0, values.nextDoc());
    assertEquals(7, values.longValue());
    assertEquals(1, values.nextDoc());
    assertEquals(18, values.longValue());
    assertEquals(NO_MORE_DOCS, values.nextDoc());
    r.close();
    w.close();
    dir.close();
}

Example 79 with Sort

use of org.apache.lucene.search.Sort in project lucene-solr by apache.

the class TestIndexSorting method testMultiValuedRandom1.

public void testMultiValuedRandom1() throws IOException {
    boolean withDeletes = random().nextBoolean();
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(new MockAnalyzer(random()));
    Sort indexSort = new Sort(new SortedNumericSortField("foo", SortField.Type.LONG));
    iwc.setIndexSort(indexSort);
    IndexWriter w = new IndexWriter(dir, iwc);
    final int numDocs = atLeast(1000);
    final FixedBitSet deleted = new FixedBitSet(numDocs);
    for (int i = 0; i < numDocs; ++i) {
        Document doc = new Document();
        int num = random().nextInt(10);
        for (int j = 0; j < num; j++) {
            doc.add(new SortedNumericDocValuesField("foo", random().nextInt(2000)));
        }
        doc.add(new StringField("id", Integer.toString(i), Store.YES));
        doc.add(new NumericDocValuesField("id", i));
        w.addDocument(doc);
        if (random().nextInt(5) == 0) {
            w.getReader().close();
        } else if (random().nextInt(30) == 0) {
            w.forceMerge(2);
        } else if (random().nextInt(4) == 0) {
            final int id = TestUtil.nextInt(random(), 0, i);
            deleted.set(id);
            w.deleteDocuments(new Term("id", Integer.toString(id)));
        }
    }
    DirectoryReader reader = w.getReader();
    // Now check that the index is consistent
    IndexSearcher searcher = newSearcher(reader);
    for (int i = 0; i < numDocs; ++i) {
        TermQuery termQuery = new TermQuery(new Term("id", Integer.toString(i)));
        final TopDocs topDocs = searcher.search(termQuery, 1);
        if (deleted.get(i)) {
            assertEquals(0, topDocs.totalHits);
        } else {
            assertEquals(1, topDocs.totalHits);
            NumericDocValues values = MultiDocValues.getNumericValues(reader, "id");
            assertEquals(topDocs.scoreDocs[0].doc, values.advance(topDocs.scoreDocs[0].doc));
            assertEquals(i, values.longValue());
            Document document = reader.document(topDocs.scoreDocs[0].doc);
            assertEquals(Integer.toString(i), document.get("id"));
        }
    }
    reader.close();
    w.close();
    dir.close();
}

Example 80 with Sort

use of org.apache.lucene.search.Sort in project lucene-solr by apache.

the class TestIndexSorting method testIndexSortWithSparseField.

public void testIndexSortWithSparseField() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = new IndexWriterConfig(new MockAnalyzer(random()));
    SortField sortField = new SortField("dense_int", SortField.Type.INT, true);
    Sort indexSort = new Sort(sortField);
    iwc.setIndexSort(indexSort);
    IndexWriter w = new IndexWriter(dir, iwc);
    Field textField = newTextField("sparse_text", "", Field.Store.NO);
    for (int i = 0; i < 128; i++) {
        Document doc = new Document();
        doc.add(new NumericDocValuesField("dense_int", i));
        if (i < 64) {
            doc.add(new NumericDocValuesField("sparse_int", i));
            doc.add(new BinaryDocValuesField("sparse_binary", new BytesRef(Integer.toString(i))));
            textField.setStringValue("foo");
            doc.add(textField);
        }
        w.addDocument(doc);
    }
    w.commit();
    w.forceMerge(1);
    DirectoryReader r = DirectoryReader.open(w);
    assertEquals(1, r.leaves().size());
    LeafReader leafReader = r.leaves().get(0).reader();
    NumericDocValues denseValues = leafReader.getNumericDocValues("dense_int");
    NumericDocValues sparseValues = leafReader.getNumericDocValues("sparse_int");
    BinaryDocValues sparseBinaryValues = leafReader.getBinaryDocValues("sparse_binary");
    NumericDocValues normsValues = leafReader.getNormValues("sparse_text");
    for (int docID = 0; docID < 128; docID++) {
        assertTrue(denseValues.advanceExact(docID));
        assertEquals(127 - docID, (int) denseValues.longValue());
        if (docID >= 64) {
            assertTrue(denseValues.advanceExact(docID));
            assertTrue(sparseValues.advanceExact(docID));
            assertTrue(sparseBinaryValues.advanceExact(docID));
            assertTrue(normsValues.advanceExact(docID));
            assertEquals(1, normsValues.longValue());
            assertEquals(127 - docID, (int) sparseValues.longValue());
            assertEquals(new BytesRef(Integer.toString(127 - docID)), sparseBinaryValues.binaryValue());
        } else {
            assertFalse(sparseBinaryValues.advanceExact(docID));
            assertFalse(sparseValues.advanceExact(docID));
            assertFalse(normsValues.advanceExact(docID));
        }
    }
    IOUtils.close(r, w, dir);
}