public void testNumerics() throws Exception {
Directory dir = newDirectory();
Document doc = new Document();
Field field = new NumericDocValuesField("numbers", 0);
doc.add(field);
IndexWriterConfig iwc = newIndexWriterConfig(random(), null);
iwc.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
int numDocs = TEST_NIGHTLY ? atLeast(500) : atLeast(50);
for (int i = 0; i < numDocs; i++) {
field.setLongValue(random().nextLong());
iw.addDocument(doc);
if (random().nextInt(17) == 0) {
iw.commit();
}
}
DirectoryReader ir = iw.getReader();
iw.forceMerge(1);
DirectoryReader ir2 = iw.getReader();
LeafReader merged = getOnlyLeafReader(ir2);
iw.close();
NumericDocValues multi = MultiDocValues.getNumericValues(ir, "numbers");
NumericDocValues single = merged.getNumericDocValues("numbers");
for (int i = 0; i < numDocs; i++) {
assertEquals(single.get(i), multi.get(i));
}
ir.close();
ir2.close();
dir.close();
}
@Override
public void setUp() throws Exception {
super.setUp();
directory = newDirectory();
RandomIndexWriter writer =
new RandomIndexWriter(
random(),
directory,
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMergePolicy(newLogMergePolicy()));
// Add series of docs with filterable fields : url, text and dates flags
addDoc(writer, "http://lucene.apache.org", "lucene 1.4.3 available", "20040101");
addDoc(writer, "http://lucene.apache.org", "New release pending", "20040102");
addDoc(writer, "http://lucene.apache.org", "Lucene 1.9 out now", "20050101");
addDoc(writer, "http://www.bar.com", "Local man bites dog", "20040101");
addDoc(writer, "http://www.bar.com", "Dog bites local man", "20040102");
addDoc(writer, "http://www.bar.com", "Dog uses Lucene", "20050101");
addDoc(writer, "http://lucene.apache.org", "Lucene 2.0 out", "20050101");
addDoc(writer, "http://lucene.apache.org", "Oops. Lucene 2.1 out", "20050102");
// Until we fix LUCENE-2348, the index must
// have only 1 segment:
writer.forceMerge(1);
reader = writer.getReader();
writer.close();
searcher = newSearcher(reader);
}
/** make sure we downgrade positions and payloads correctly */
public void testMixing() throws Exception {
// no positions
FieldType ft = new FieldType(TextField.TYPE_NOT_STORED);
ft.setIndexOptions(IndexOptions.DOCS_AND_FREQS);
Directory dir = newDirectory();
RandomIndexWriter iw = new RandomIndexWriter(random(), dir);
for (int i = 0; i < 20; i++) {
Document doc = new Document();
if (i < 19 && random().nextBoolean()) {
for (int j = 0; j < 50; j++) {
doc.add(new TextField("foo", "i have positions", Field.Store.NO));
}
} else {
for (int j = 0; j < 50; j++) {
doc.add(new Field("foo", "i have no positions", ft));
}
}
iw.addDocument(doc);
iw.commit();
}
if (random().nextBoolean()) {
iw.forceMerge(1);
}
DirectoryReader ir = iw.getReader();
FieldInfos fis = MultiFields.getMergedFieldInfos(ir);
assertEquals(IndexOptions.DOCS_AND_FREQS, fis.fieldInfo("foo").getIndexOptions());
assertFalse(fis.fieldInfo("foo").hasPayloads());
iw.close();
ir.close();
dir.close(); // checkindex
}
public void testSeekCeilNotFound() throws Exception {
Directory dir = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), dir);
Document doc = new Document();
// Get empty string in there!
doc.add(newStringField("field", "", Field.Store.NO));
w.addDocument(doc);
for (int i = 0; i < 36; i++) {
doc = new Document();
String term = "" + (char) (97 + i);
String term2 = "a" + (char) (97 + i);
doc.add(newTextField("field", term + " " + term2, Field.Store.NO));
w.addDocument(doc);
}
w.forceMerge(1);
IndexReader r = w.getReader();
TermsEnum te = MultiFields.getTerms(r, "field").iterator(null);
assertEquals(TermsEnum.SeekStatus.NOT_FOUND, te.seekCeil(new BytesRef(new byte[] {0x22})));
assertEquals("a", te.term().utf8ToString());
assertEquals(1L, te.ord());
r.close();
w.close();
dir.close();
}
public void testDifferentTypedDocValuesField2() throws Exception {
Directory d = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), d);
Document doc = new Document();
// Index doc values are single-valued so we should not
// be able to add same field more than once:
Field f = new NumericDocValuesField("field", 17);
doc.add(f);
doc.add(new SortedDocValuesField("field", new BytesRef("hello")));
try {
w.addDocument(doc);
fail("didn't hit expected exception");
} catch (IllegalArgumentException iae) {
// expected
}
doc = new Document();
doc.add(f);
w.addDocument(doc);
w.forceMerge(1);
DirectoryReader r = w.getReader();
assertEquals(17, getOnlySegmentReader(r).getNumericDocValues("field").get(0));
r.close();
w.close();
d.close();
}
public void testMultiValuedDocValuesField() throws Exception {
Directory d = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), d);
Document doc = new Document();
Field f = new NumericDocValuesField("field", 17);
// Index doc values are single-valued so we should not
// be able to add same field more than once:
doc.add(f);
doc.add(f);
try {
w.addDocument(doc);
fail("didn't hit expected exception");
} catch (IllegalArgumentException iae) {
// expected
}
doc = new Document();
doc.add(f);
w.addDocument(doc);
w.forceMerge(1);
DirectoryReader r = w.getReader();
w.close();
assertEquals(17, FieldCache.DEFAULT.getInts(getOnlySegmentReader(r), "field", false).get(0));
r.close();
d.close();
}
public void testThreeBlocks() throws Exception {
Directory dir = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), dir);
List<String> terms = new ArrayList<>();
for (int i = 0; i < 36; i++) {
Document doc = new Document();
String term = "" + (char) (97 + i);
terms.add(term);
if (VERBOSE) {
System.out.println("i=" + i + " term=" + term);
}
doc.add(newTextField("field", term, Field.Store.NO));
w.addDocument(doc);
}
for (int i = 0; i < 36; i++) {
Document doc = new Document();
String term = "m" + (char) (97 + i);
terms.add(term);
if (VERBOSE) {
System.out.println("i=" + i + " term=" + term);
}
doc.add(newTextField("field", term, Field.Store.NO));
w.addDocument(doc);
}
for (int i = 0; i < 36; i++) {
Document doc = new Document();
String term = "mo" + (char) (97 + i);
terms.add(term);
if (VERBOSE) {
System.out.println("i=" + i + " term=" + term);
}
doc.add(newTextField("field", term, Field.Store.NO));
w.addDocument(doc);
}
w.forceMerge(1);
IndexReader r = w.getReader();
TermsEnum te = MultiFields.getTerms(r, "field").iterator(null);
if (VERBOSE) {
while (te.next() != null) {
System.out.println("TERM: " + te.ord() + " " + te.term().utf8ToString());
}
}
assertTrue(te.seekExact(new BytesRef("mo")));
assertEquals(27, te.ord());
te.seekExact(90);
assertEquals(new BytesRef("s"), te.term());
testEnum(te, terms);
r.close();
w.close();
dir.close();
}
public void testSortedNumeric() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig(random(), null);
iwc.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
int numDocs = TEST_NIGHTLY ? atLeast(500) : atLeast(50);
for (int i = 0; i < numDocs; i++) {
Document doc = new Document();
int numValues = random().nextInt(5);
for (int j = 0; j < numValues; j++) {
doc.add(
new SortedNumericDocValuesField(
"nums", TestUtil.nextLong(random(), Long.MIN_VALUE, Long.MAX_VALUE)));
}
iw.addDocument(doc);
if (random().nextInt(17) == 0) {
iw.commit();
}
}
DirectoryReader ir = iw.getReader();
iw.forceMerge(1);
DirectoryReader ir2 = iw.getReader();
LeafReader merged = getOnlyLeafReader(ir2);
iw.close();
SortedNumericDocValues multi = MultiDocValues.getSortedNumericValues(ir, "nums");
SortedNumericDocValues single = merged.getSortedNumericDocValues("nums");
if (multi == null) {
assertNull(single);
} else {
// check values
for (int i = 0; i < numDocs; i++) {
single.setDocument(i);
ArrayList<Long> expectedList = new ArrayList<>();
for (int j = 0; j < single.count(); j++) {
expectedList.add(single.valueAt(j));
}
multi.setDocument(i);
assertEquals(expectedList.size(), multi.count());
for (int j = 0; j < single.count(); j++) {
assertEquals(expectedList.get(j).longValue(), multi.valueAt(j));
}
}
}
ir.close();
ir2.close();
dir.close();
}
/**
* Indexes at least 1 document with f1, and at least 1 document with f2. returns the norms for
* "field".
*/
NumericDocValues getNorms(String field, Field f1, Field f2) throws IOException {
Directory dir = newDirectory();
IndexWriterConfig iwc =
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMergePolicy(newLogMergePolicy());
RandomIndexWriter riw = new RandomIndexWriter(random(), dir, iwc);
// add f1
Document d = new Document();
d.add(f1);
riw.addDocument(d);
// add f2
d = new Document();
d.add(f2);
riw.addDocument(d);
// add a mix of f1's and f2's
int numExtraDocs = TestUtil.nextInt(random(), 1, 1000);
for (int i = 0; i < numExtraDocs; i++) {
d = new Document();
d.add(random().nextBoolean() ? f1 : f2);
riw.addDocument(d);
}
IndexReader ir1 = riw.getReader();
// todo: generalize
NumericDocValues norms1 = MultiDocValues.getNormValues(ir1, field);
// fully merge and validate MultiNorms against single segment.
riw.forceMerge(1);
DirectoryReader ir2 = riw.getReader();
NumericDocValues norms2 = getOnlySegmentReader(ir2).getNormValues(field);
if (norms1 == null) {
assertNull(norms2);
} else {
for (int docID = 0; docID < ir1.maxDoc(); docID++) {
assertEquals(norms1.get(docID), norms2.get(docID));
}
}
ir1.close();
ir2.close();
riw.close();
dir.close();
return norms1;
}
public void testDocsWithField() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig(random(), null);
iwc.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
int numDocs = TEST_NIGHTLY ? atLeast(500) : atLeast(50);
for (int i = 0; i < numDocs; i++) {
Document doc = new Document();
if (random().nextInt(4) >= 0) {
doc.add(new NumericDocValuesField("numbers", random().nextLong()));
}
doc.add(new NumericDocValuesField("numbersAlways", random().nextLong()));
iw.addDocument(doc);
if (random().nextInt(17) == 0) {
iw.commit();
}
}
DirectoryReader ir = iw.getReader();
iw.forceMerge(1);
DirectoryReader ir2 = iw.getReader();
LeafReader merged = getOnlyLeafReader(ir2);
iw.close();
Bits multi = MultiDocValues.getDocsWithField(ir, "numbers");
Bits single = merged.getDocsWithField("numbers");
if (multi == null) {
assertNull(single);
} else {
assertEquals(single.length(), multi.length());
for (int i = 0; i < numDocs; i++) {
assertEquals(single.get(i), multi.get(i));
}
}
multi = MultiDocValues.getDocsWithField(ir, "numbersAlways");
single = merged.getDocsWithField("numbersAlways");
assertEquals(single.length(), multi.length());
for (int i = 0; i < numDocs; i++) {
assertEquals(single.get(i), multi.get(i));
}
ir.close();
ir2.close();
dir.close();
}
public void testSimple() throws Exception {
Directory dir = newDirectory();
RandomIndexWriter iw = new RandomIndexWriter(random(), dir);
int numDocs = atLeast(100);
for (int i = 0; i < numDocs; i++) {
iw.addDocument(doc());
}
IndexReader ir = iw.getReader();
verifyCount(ir);
ir.close();
iw.forceMerge(1);
ir = iw.getReader();
verifyCount(ir);
ir.close();
iw.close();
dir.close();
}
public void testSorted() throws Exception {
Directory dir = newDirectory();
Document doc = new Document();
Field field = new SortedDocValuesField("bytes", new BytesRef());
doc.add(field);
IndexWriterConfig iwc = newIndexWriterConfig(random(), null);
iwc.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
int numDocs = TEST_NIGHTLY ? atLeast(500) : atLeast(50);
for (int i = 0; i < numDocs; i++) {
BytesRef ref = new BytesRef(TestUtil.randomUnicodeString(random()));
field.setBytesValue(ref);
if (random().nextInt(7) == 0) {
iw.addDocument(new Document());
}
iw.addDocument(doc);
if (random().nextInt(17) == 0) {
iw.commit();
}
}
DirectoryReader ir = iw.getReader();
iw.forceMerge(1);
DirectoryReader ir2 = iw.getReader();
LeafReader merged = getOnlyLeafReader(ir2);
iw.close();
SortedDocValues multi = MultiDocValues.getSortedValues(ir, "bytes");
SortedDocValues single = merged.getSortedDocValues("bytes");
assertEquals(single.getValueCount(), multi.getValueCount());
for (int i = 0; i < numDocs; i++) {
// check ord
assertEquals(single.getOrd(i), multi.getOrd(i));
// check value
final BytesRef expected = BytesRef.deepCopyOf(single.get(i));
final BytesRef actual = multi.get(i);
assertEquals(expected, actual);
}
ir.close();
ir2.close();
dir.close();
}
public void testAddIndexes() throws IOException {
Directory d1 = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), d1);
Document doc = new Document();
doc.add(newStringField("id", "1", Field.Store.YES));
doc.add(new NumericDocValuesField("dv", 1));
w.addDocument(doc);
IndexReader r1 = w.getReader();
w.close();
Directory d2 = newDirectory();
w = new RandomIndexWriter(random(), d2);
doc = new Document();
doc.add(newStringField("id", "2", Field.Store.YES));
doc.add(new NumericDocValuesField("dv", 2));
w.addDocument(doc);
IndexReader r2 = w.getReader();
w.close();
Directory d3 = newDirectory();
w = new RandomIndexWriter(random(), d3);
w.addIndexes(SlowCompositeReaderWrapper.wrap(r1), SlowCompositeReaderWrapper.wrap(r2));
r1.close();
d1.close();
r2.close();
d2.close();
w.forceMerge(1);
DirectoryReader r3 = w.getReader();
w.close();
AtomicReader sr = getOnlySegmentReader(r3);
assertEquals(2, sr.numDocs());
NumericDocValues docValues = sr.getNumericDocValues("dv");
assertNotNull(docValues);
r3.close();
d3.close();
}
public void testTwoBlocks() throws Exception {
Directory dir = newDirectory();
RandomIndexWriter w = new RandomIndexWriter(random(), dir);
List<String> terms = new ArrayList<>();
for (int i = 0; i < 36; i++) {
Document doc = new Document();
String term = "" + (char) (97 + i);
terms.add(term);
if (VERBOSE) {
System.out.println("i=" + i + " term=" + term);
}
doc.add(newTextField("field", term, Field.Store.NO));
w.addDocument(doc);
}
for (int i = 0; i < 36; i++) {
Document doc = new Document();
String term = "m" + (char) (97 + i);
terms.add(term);
if (VERBOSE) {
System.out.println("i=" + i + " term=" + term);
}
doc.add(newTextField("field", term, Field.Store.NO));
w.addDocument(doc);
}
if (VERBOSE) {
System.out.println("TEST: now forceMerge");
}
w.forceMerge(1);
IndexReader r = w.getReader();
TermsEnum te = MultiFields.getTerms(r, "field").iterator(null);
assertTrue(te.seekExact(new BytesRef("mo")));
assertEquals(27, te.ord());
te.seekExact(54);
assertEquals(new BytesRef("s"), te.term());
Collections.sort(terms);
for (int i = terms.size() - 1; i >= 0; i--) {
te.seekExact(i);
assertEquals(i, te.ord());
assertEquals(terms.get(i), te.term().utf8ToString());
}
int iters = atLeast(1000);
for (int iter = 0; iter < iters; iter++) {
int ord = random().nextInt(terms.size());
BytesRef term = new BytesRef(terms.get(ord));
if (random().nextBoolean()) {
if (VERBOSE) {
System.out.println("TEST: iter=" + iter + " seek to ord=" + ord + " of " + terms.size());
}
te.seekExact(ord);
} else {
if (VERBOSE) {
System.out.println(
"TEST: iter="
+ iter
+ " seek to term="
+ terms.get(ord)
+ " ord="
+ ord
+ " of "
+ terms.size());
}
te.seekExact(term);
}
assertEquals(ord, te.ord());
assertEquals(term, te.term());
}
r.close();
w.close();
dir.close();
}
public void testSortedTermsEnum() throws IOException {
Directory directory = newDirectory();
Analyzer analyzer = new MockAnalyzer(random());
IndexWriterConfig iwconfig = newIndexWriterConfig(TEST_VERSION_CURRENT, analyzer);
iwconfig.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iwriter = new RandomIndexWriter(random(), directory, iwconfig);
Document doc = new Document();
doc.add(new StringField("field", "hello", Field.Store.NO));
iwriter.addDocument(doc);
doc = new Document();
doc.add(new StringField("field", "world", Field.Store.NO));
iwriter.addDocument(doc);
doc = new Document();
doc.add(new StringField("field", "beer", Field.Store.NO));
iwriter.addDocument(doc);
iwriter.forceMerge(1);
DirectoryReader ireader = iwriter.getReader();
iwriter.close();
AtomicReader ar = getOnlySegmentReader(ireader);
SortedSetDocValues dv = FieldCache.DEFAULT.getDocTermOrds(ar, "field");
assertEquals(3, dv.getValueCount());
TermsEnum termsEnum = dv.termsEnum();
// next()
assertEquals("beer", termsEnum.next().utf8ToString());
assertEquals(0, termsEnum.ord());
assertEquals("hello", termsEnum.next().utf8ToString());
assertEquals(1, termsEnum.ord());
assertEquals("world", termsEnum.next().utf8ToString());
assertEquals(2, termsEnum.ord());
// seekCeil()
assertEquals(SeekStatus.NOT_FOUND, termsEnum.seekCeil(new BytesRef("ha!")));
assertEquals("hello", termsEnum.term().utf8ToString());
assertEquals(1, termsEnum.ord());
assertEquals(SeekStatus.FOUND, termsEnum.seekCeil(new BytesRef("beer")));
assertEquals("beer", termsEnum.term().utf8ToString());
assertEquals(0, termsEnum.ord());
assertEquals(SeekStatus.END, termsEnum.seekCeil(new BytesRef("zzz")));
// seekExact()
assertTrue(termsEnum.seekExact(new BytesRef("beer"), true));
assertEquals("beer", termsEnum.term().utf8ToString());
assertEquals(0, termsEnum.ord());
assertTrue(termsEnum.seekExact(new BytesRef("hello"), true));
assertEquals("hello", termsEnum.term().utf8ToString());
assertEquals(1, termsEnum.ord());
assertTrue(termsEnum.seekExact(new BytesRef("world"), true));
assertEquals("world", termsEnum.term().utf8ToString());
assertEquals(2, termsEnum.ord());
assertFalse(termsEnum.seekExact(new BytesRef("bogus"), true));
// seek(ord)
termsEnum.seekExact(0);
assertEquals("beer", termsEnum.term().utf8ToString());
assertEquals(0, termsEnum.ord());
termsEnum.seekExact(1);
assertEquals("hello", termsEnum.term().utf8ToString());
assertEquals(1, termsEnum.ord());
termsEnum.seekExact(2);
assertEquals("world", termsEnum.term().utf8ToString());
assertEquals(2, termsEnum.ord());
ireader.close();
directory.close();
}
public void testRandomStoredFields() throws IOException {
Directory dir = newDirectory();
Random rand = random();
RandomIndexWriter w =
new RandomIndexWriter(
rand,
dir,
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMaxBufferedDocs(_TestUtil.nextInt(rand, 5, 20)));
// w.w.setNoCFSRatio(0.0);
final int docCount = atLeast(200);
final int fieldCount = _TestUtil.nextInt(rand, 1, 5);
final List<Integer> fieldIDs = new ArrayList<Integer>();
FieldType customType = new FieldType(TextField.TYPE_STORED);
customType.setTokenized(false);
Field idField = newField("id", "", customType);
for (int i = 0; i < fieldCount; i++) {
fieldIDs.add(i);
}
final Map<String, Document> docs = new HashMap<String, Document>();
if (VERBOSE) {
System.out.println("TEST: build index docCount=" + docCount);
}
FieldType customType2 = new FieldType();
customType2.setStored(true);
for (int i = 0; i < docCount; i++) {
Document doc = new Document();
doc.add(idField);
final String id = "" + i;
idField.setStringValue(id);
docs.put(id, doc);
if (VERBOSE) {
System.out.println("TEST: add doc id=" + id);
}
for (int field : fieldIDs) {
final String s;
if (rand.nextInt(4) != 3) {
s = _TestUtil.randomUnicodeString(rand, 1000);
doc.add(newField("f" + field, s, customType2));
} else {
s = null;
}
}
w.addDocument(doc);
if (rand.nextInt(50) == 17) {
// mixup binding of field name -> Number every so often
Collections.shuffle(fieldIDs);
}
if (rand.nextInt(5) == 3 && i > 0) {
final String delID = "" + rand.nextInt(i);
if (VERBOSE) {
System.out.println("TEST: delete doc id=" + delID);
}
w.deleteDocuments(new Term("id", delID));
docs.remove(delID);
}
}
if (VERBOSE) {
System.out.println("TEST: " + docs.size() + " docs in index; now load fields");
}
if (docs.size() > 0) {
String[] idsList = docs.keySet().toArray(new String[docs.size()]);
for (int x = 0; x < 2; x++) {
IndexReader r = w.getReader();
IndexSearcher s = newSearcher(r);
if (VERBOSE) {
System.out.println("TEST: cycle x=" + x + " r=" + r);
}
int num = atLeast(1000);
for (int iter = 0; iter < num; iter++) {
String testID = idsList[rand.nextInt(idsList.length)];
if (VERBOSE) {
System.out.println("TEST: test id=" + testID);
}
TopDocs hits = s.search(new TermQuery(new Term("id", testID)), 1);
assertEquals(1, hits.totalHits);
Document doc = r.document(hits.scoreDocs[0].doc);
Document docExp = docs.get(testID);
for (int i = 0; i < fieldCount; i++) {
assertEquals(
"doc " + testID + ", field f" + fieldCount + " is wrong",
docExp.get("f" + i),
doc.get("f" + i));
}
}
r.close();
w.forceMerge(1);
}
}
w.close();
dir.close();
}
@Nightly
public void testBigDocuments() throws IOException {
// "big" as "much bigger than the chunk size"
// for this test we force a FS dir
// we can't just use newFSDirectory, because this test doesn't really index anything.
// so if we get NRTCachingDir+SimpleText, we make massive stored fields and OOM (LUCENE-4484)
Directory dir =
new MockDirectoryWrapper(
random(), new MMapDirectory(_TestUtil.getTempDir("testBigDocuments")));
IndexWriterConfig iwConf =
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()));
iwConf.setMaxBufferedDocs(RandomInts.randomIntBetween(random(), 2, 30));
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwConf);
if (dir instanceof MockDirectoryWrapper) {
((MockDirectoryWrapper) dir).setThrottling(Throttling.NEVER);
}
final Document emptyDoc = new Document(); // emptyDoc
final Document bigDoc1 = new Document(); // lot of small fields
final Document bigDoc2 = new Document(); // 1 very big field
final Field idField = new StringField("id", "", Store.NO);
emptyDoc.add(idField);
bigDoc1.add(idField);
bigDoc2.add(idField);
final FieldType onlyStored = new FieldType(StringField.TYPE_STORED);
onlyStored.setIndexed(false);
final Field smallField =
new Field("fld", randomByteArray(random().nextInt(10), 256), onlyStored);
final int numFields = RandomInts.randomIntBetween(random(), 500000, 1000000);
for (int i = 0; i < numFields; ++i) {
bigDoc1.add(smallField);
}
final Field bigField =
new Field(
"fld",
randomByteArray(RandomInts.randomIntBetween(random(), 1000000, 5000000), 2),
onlyStored);
bigDoc2.add(bigField);
final int numDocs = atLeast(5);
final Document[] docs = new Document[numDocs];
for (int i = 0; i < numDocs; ++i) {
docs[i] = RandomPicks.randomFrom(random(), Arrays.asList(emptyDoc, bigDoc1, bigDoc2));
}
for (int i = 0; i < numDocs; ++i) {
idField.setStringValue("" + i);
iw.addDocument(docs[i]);
if (random().nextInt(numDocs) == 0) {
iw.commit();
}
}
iw.commit();
iw.forceMerge(1); // look at what happens when big docs are merged
final DirectoryReader rd = DirectoryReader.open(dir);
final IndexSearcher searcher = new IndexSearcher(rd);
for (int i = 0; i < numDocs; ++i) {
final Query query = new TermQuery(new Term("id", "" + i));
final TopDocs topDocs = searcher.search(query, 1);
assertEquals("" + i, 1, topDocs.totalHits);
final Document doc = rd.document(topDocs.scoreDocs[0].doc);
assertNotNull(doc);
final IndexableField[] fieldValues = doc.getFields("fld");
assertEquals(docs[i].getFields("fld").length, fieldValues.length);
if (fieldValues.length > 0) {
assertEquals(docs[i].getFields("fld")[0].binaryValue(), fieldValues[0].binaryValue());
}
}
rd.close();
iw.close();
dir.close();
}
public void testWriteReadMerge() throws IOException {
// get another codec, other than the default: so we are merging segments across different codecs
final Codec otherCodec;
if ("SimpleText".equals(Codec.getDefault().getName())) {
otherCodec = new Lucene46Codec();
} else {
otherCodec = new SimpleTextCodec();
}
Directory dir = newDirectory();
IndexWriterConfig iwConf =
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()));
iwConf.setMaxBufferedDocs(RandomInts.randomIntBetween(random(), 2, 30));
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwConf.clone());
final int docCount = atLeast(200);
final byte[][][] data = new byte[docCount][][];
for (int i = 0; i < docCount; ++i) {
final int fieldCount =
rarely()
? RandomInts.randomIntBetween(random(), 1, 500)
: RandomInts.randomIntBetween(random(), 1, 5);
data[i] = new byte[fieldCount][];
for (int j = 0; j < fieldCount; ++j) {
final int length = rarely() ? random().nextInt(1000) : random().nextInt(10);
final int max = rarely() ? 256 : 2;
data[i][j] = randomByteArray(length, max);
}
}
final FieldType type = new FieldType(StringField.TYPE_STORED);
type.setIndexed(false);
type.freeze();
IntField id = new IntField("id", 0, Store.YES);
for (int i = 0; i < data.length; ++i) {
Document doc = new Document();
doc.add(id);
id.setIntValue(i);
for (int j = 0; j < data[i].length; ++j) {
Field f = new Field("bytes" + j, data[i][j], type);
doc.add(f);
}
iw.w.addDocument(doc);
if (random().nextBoolean() && (i % (data.length / 10) == 0)) {
iw.w.close();
// test merging against a non-compressing codec
if (iwConf.getCodec() == otherCodec) {
iwConf.setCodec(Codec.getDefault());
} else {
iwConf.setCodec(otherCodec);
}
iw = new RandomIndexWriter(random(), dir, iwConf.clone());
}
}
for (int i = 0; i < 10; ++i) {
final int min = random().nextInt(data.length);
final int max = min + random().nextInt(20);
iw.deleteDocuments(NumericRangeQuery.newIntRange("id", min, max, true, false));
}
iw.forceMerge(2); // force merges with deletions
iw.commit();
final DirectoryReader ir = DirectoryReader.open(dir);
assertTrue(ir.numDocs() > 0);
int numDocs = 0;
for (int i = 0; i < ir.maxDoc(); ++i) {
final Document doc = ir.document(i);
if (doc == null) {
continue;
}
++numDocs;
final int docId = doc.getField("id").numericValue().intValue();
assertEquals(data[docId].length + 1, doc.getFields().size());
for (int j = 0; j < data[docId].length; ++j) {
final byte[] arr = data[docId][j];
final BytesRef arr2Ref = doc.getBinaryValue("bytes" + j);
final byte[] arr2 =
Arrays.copyOfRange(arr2Ref.bytes, arr2Ref.offset, arr2Ref.offset + arr2Ref.length);
assertArrayEquals(arr, arr2);
}
}
assertTrue(ir.numDocs() <= numDocs);
ir.close();
iw.deleteAll();
iw.commit();
iw.forceMerge(1);
iw.close();
dir.close();
}
// tries to make more dups than testSortedSet
public void testSortedSetWithDups() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig(random(), null);
iwc.setMergePolicy(newLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
int numDocs = TEST_NIGHTLY ? atLeast(500) : atLeast(50);
for (int i = 0; i < numDocs; i++) {
Document doc = new Document();
int numValues = random().nextInt(5);
for (int j = 0; j < numValues; j++) {
doc.add(
new SortedSetDocValuesField(
"bytes", new BytesRef(TestUtil.randomSimpleString(random(), 2))));
}
iw.addDocument(doc);
if (random().nextInt(17) == 0) {
iw.commit();
}
}
DirectoryReader ir = iw.getReader();
iw.forceMerge(1);
DirectoryReader ir2 = iw.getReader();
LeafReader merged = getOnlyLeafReader(ir2);
iw.close();
SortedSetDocValues multi = MultiDocValues.getSortedSetValues(ir, "bytes");
SortedSetDocValues single = merged.getSortedSetDocValues("bytes");
if (multi == null) {
assertNull(single);
} else {
assertEquals(single.getValueCount(), multi.getValueCount());
// check values
for (long i = 0; i < single.getValueCount(); i++) {
final BytesRef expected = BytesRef.deepCopyOf(single.lookupOrd(i));
final BytesRef actual = multi.lookupOrd(i);
assertEquals(expected, actual);
}
// check ord list
for (int i = 0; i < numDocs; i++) {
single.setDocument(i);
ArrayList<Long> expectedList = new ArrayList<>();
long ord;
while ((ord = single.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
expectedList.add(ord);
}
multi.setDocument(i);
int upto = 0;
while ((ord = multi.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
assertEquals(expectedList.get(upto).longValue(), ord);
upto++;
}
assertEquals(expectedList.size(), upto);
}
}
ir.close();
ir2.close();
dir.close();
}
