public override void BeforeClass()
{
base.BeforeClass();
Random random = Random;
m_directory = NewDirectory();
m_stopword = "" + GetRandomChar();
CharacterRunAutomaton stopset = new CharacterRunAutomaton(BasicAutomata.MakeString(m_stopword));
m_analyzer = new MockAnalyzer(random, MockTokenizer.WHITESPACE, false, stopset);
RandomIndexWriter iw = new RandomIndexWriter(random, m_directory, m_analyzer);
Document doc = new Document();
Field id = new StringField("id", "", Field.Store.NO);
Field field = new TextField("field", "", Field.Store.NO);
doc.Add(id);
doc.Add(field);
// index some docs
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; i++)
{
id.SetStringValue(Convert.ToString(i, CultureInfo.InvariantCulture));
field.SetStringValue(RandomFieldContents());
iw.AddDocument(doc);
}
// delete some docs
int numDeletes = numDocs / 20;
for (int i = 0; i < numDeletes; i++)
{
Term toDelete = new Term("id", Convert.ToString(random.Next(numDocs), CultureInfo.InvariantCulture));
if (random.NextBoolean())
{
iw.DeleteDocuments(toDelete);
}
else
{
iw.DeleteDocuments(new TermQuery(toDelete));
}
}
m_reader = iw.GetReader();
m_s1 = NewSearcher(m_reader);
m_s2 = NewSearcher(m_reader);
iw.Dispose();
}
public static void BeforeClass()
{
Random random = Random();
Directory = NewDirectory();
Stopword = "" + RandomChar();
CharacterRunAutomaton stopset = new CharacterRunAutomaton(BasicAutomata.MakeString(Stopword));
Analyzer = new MockAnalyzer(random, MockTokenizer.WHITESPACE, false, stopset);
RandomIndexWriter iw = new RandomIndexWriter(random, Directory, Analyzer);
Document doc = new Document();
Field id = new StringField("id", "", Field.Store.NO);
Field field = new TextField("field", "", Field.Store.NO);
doc.Add(id);
doc.Add(field);
// index some docs
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; i++)
{
id.StringValue = Convert.ToString(i);
field.StringValue = RandomFieldContents();
iw.AddDocument(doc);
}
// delete some docs
int numDeletes = numDocs / 20;
for (int i = 0; i < numDeletes; i++)
{
Term toDelete = new Term("id", Convert.ToString(random.Next(numDocs)));
if (random.NextBoolean())
{
iw.DeleteDocuments(toDelete);
}
else
{
iw.DeleteDocuments(new TermQuery(toDelete));
}
}
Reader = iw.Reader;
S1 = NewSearcher(Reader);
S2 = NewSearcher(Reader);
iw.Dispose();
}
private void CreateRandomIndexes(int maxSegments)
{
dir = NewDirectory();
numDocs = AtLeast(150);
int numTerms = TestUtil.NextInt(Random(), 1, numDocs / 5);
ISet<string> randomTerms = new HashSet<string>();
while (randomTerms.size() < numTerms)
{
randomTerms.add(TestUtil.RandomSimpleString(Random()));
}
terms = new List<string>(randomTerms);
int seed = Random().Next();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new Random(seed)));
iwc.SetMergePolicy(TestSortingMergePolicy.NewSortingMergePolicy(sort));
iw = new RandomIndexWriter(new Random(seed), dir, iwc);
for (int i = 0; i < numDocs; ++i)
{
Document doc = RandomDocument();
iw.AddDocument(doc);
if (i == numDocs / 2 || (i != numDocs - 1 && Random().nextInt(8) == 0))
{
iw.Commit();
}
if (Random().nextInt(15) == 0)
{
string term = RandomInts.RandomFrom(Random(), terms);
iw.DeleteDocuments(new Term("s", term));
}
}
reader = iw.Reader;
}
private void CreateRandomIndexes(int maxSegments)
{
dir = NewDirectory();
numDocs = AtLeast(150);
int numTerms = TestUtil.NextInt32(Random, 1, numDocs / 5);
ISet <string> randomTerms = new JCG.HashSet <string>();
while (randomTerms.size() < numTerms)
{
randomTerms.add(TestUtil.RandomSimpleString(Random));
}
terms = new JCG.List <string>(randomTerms);
long seed = Random.NextInt64();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new J2N.Randomizer(seed)));
iwc.SetMergePolicy(TestSortingMergePolicy.NewSortingMergePolicy(sort));
iw = new RandomIndexWriter(new J2N.Randomizer(seed), dir, iwc);
for (int i = 0; i < numDocs; ++i)
{
Document doc = RandomDocument();
iw.AddDocument(doc);
if (i == numDocs / 2 || (i != numDocs - 1 && Random.nextInt(8) == 0))
{
iw.Commit();
}
if (Random.nextInt(15) == 0)
{
string term = RandomPicks.RandomFrom(Random, terms);
iw.DeleteDocuments(new Term("s", term));
}
}
reader = iw.GetReader();
}
public virtual void Test()
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
conf.SetCodec(new Lucene46Codec());
RandomIndexWriter riw = new RandomIndexWriter(Random(), dir, conf);
Document doc = new Document();
// these fields should sometimes get term vectors, etc
Field idField = NewStringField("id", "", Field.Store.NO);
Field bodyField = NewTextField("body", "", Field.Store.NO);
Field dvField = new NumericDocValuesField("dv", 5);
doc.Add(idField);
doc.Add(bodyField);
doc.Add(dvField);
for (int i = 0; i < 100; i++)
{
idField.StringValue = Convert.ToString(i);
bodyField.StringValue = TestUtil.RandomUnicodeString(Random());
riw.AddDocument(doc);
if (Random().Next(7) == 0)
{
riw.Commit();
}
if (Random().Next(20) == 0)
{
riw.DeleteDocuments(new Term("id", Convert.ToString(i)));
}
}
riw.Dispose();
CheckHeaders(dir);
dir.Dispose();
}
public override void SetUp()
{
base.SetUp();
dir = NewDirectory();
iw = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
doc.Add(idField);
// add 500 docs with id 0..499
for (int i = 0; i < 500; i++)
{
idField.SetStringValue(Convert.ToString(i));
iw.AddDocument(doc);
}
// delete 20 of them
for (int i = 0; i < 20; i++)
{
iw.DeleteDocuments(new Term("id", Convert.ToString(Random.Next(iw.MaxDoc))));
}
ir = iw.GetReader();
@is = NewSearcher(ir);
}
private int[] BuildIndex(RandomIndexWriter writer, int docs)
{
int[] docStates = new int[docs];
for (int i = 0; i < docs; i++)
{
Document doc = new Document();
if (Random.NextBoolean())
{
docStates[i] = 1;
doc.Add(NewTextField("some", "value", Field.Store.YES));
}
doc.Add(NewTextField("all", "test", Field.Store.NO));
doc.Add(NewTextField("id", "" + i, Field.Store.YES));
writer.AddDocument(doc);
}
writer.Commit();
int numDeletes = Random.Next(docs);
for (int i = 0; i < numDeletes; i++)
{
int docID = Random.Next(docs);
writer.DeleteDocuments(new Term("id", "" + docID));
docStates[docID] = 2;
}
writer.Dispose();
return(docStates);
}
public virtual void TestRandom()
{
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, d);
w.IndexWriter.Config.SetMaxBufferedDocs(17);
int numDocs = AtLeast(100);
ISet <string> aDocs = new JCG.HashSet <string>();
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
string v;
if (Random.Next(5) == 4)
{
v = "a";
aDocs.Add("" + i);
}
else
{
v = "b";
}
Field f = NewStringField("field", v, Field.Store.NO);
doc.Add(f);
doc.Add(NewStringField("id", "" + i, Field.Store.YES));
w.AddDocument(doc);
}
int numDelDocs = AtLeast(10);
for (int i = 0; i < numDelDocs; i++)
{
string delID = "" + Random.Next(numDocs);
w.DeleteDocuments(new Term("id", delID));
aDocs.Remove(delID);
}
IndexReader r = w.GetReader();
w.Dispose();
TopDocs hits = NewSearcher(r).Search(new MatchAllDocsQuery(), new QueryWrapperFilter(new TermQuery(new Term("field", "a"))), numDocs);
Assert.AreEqual(aDocs.Count, hits.TotalHits);
foreach (ScoreDoc sd in hits.ScoreDocs)
{
Assert.IsTrue(aDocs.Contains(r.Document(sd.Doc).Get("id")));
}
r.Dispose();
d.Dispose();
}
public override void SetUp()
{
base.SetUp();
dir = NewDirectory();
fieldName = Random.NextBoolean() ? "field" : ""; // sometimes use an empty string as field name
RandomIndexWriter writer = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random, MockTokenizer.KEYWORD, false)).SetMaxBufferedDocs(TestUtil.NextInt32(Random, 50, 1000)));
JCG.List <string> terms = new JCG.List <string>();
int num = AtLeast(200);
for (int i = 0; i < num; i++)
{
Document doc = new Document();
doc.Add(NewStringField("id", Convert.ToString(i), Field.Store.NO));
int numTerms = Random.Next(4);
for (int j = 0; j < numTerms; j++)
{
string s = TestUtil.RandomUnicodeString(Random);
doc.Add(NewStringField(fieldName, s, Field.Store.NO));
// if the default codec doesn't support sortedset, we will uninvert at search time
if (DefaultCodecSupportsSortedSet)
{
doc.Add(new SortedSetDocValuesField(fieldName, new BytesRef(s)));
}
terms.Add(s);
}
writer.AddDocument(doc);
}
if (Verbose)
{
// utf16 order
terms.Sort(StringComparer.Ordinal);
Console.WriteLine("UTF16 order:");
foreach (string s in terms)
{
Console.WriteLine(" " + UnicodeUtil.ToHexString(s));
}
}
int numDeletions = Random.Next(num / 10);
for (int i = 0; i < numDeletions; i++)
{
writer.DeleteDocuments(new Term("id", Convert.ToString(Random.Next(num))));
}
reader = writer.GetReader();
searcher1 = NewSearcher(reader);
searcher2 = NewSearcher(reader);
writer.Dispose();
}
public virtual void TestSumDocFreq_Mem()
{
int numDocs = AtLeast(500);
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
Document doc = new Document();
Field id = NewStringField("id", "", Field.Store.NO);
Field field1 = NewTextField("foo", "", Field.Store.NO);
Field field2 = NewTextField("bar", "", Field.Store.NO);
doc.Add(id);
doc.Add(field1);
doc.Add(field2);
for (int i = 0; i < numDocs; i++)
{
id.StringValue = "" + i;
char ch1 = (char)TestUtil.NextInt(Random(), 'a', 'z');
char ch2 = (char)TestUtil.NextInt(Random(), 'a', 'z');
field1.StringValue = "" + ch1 + " " + ch2;
ch1 = (char)TestUtil.NextInt(Random(), 'a', 'z');
ch2 = (char)TestUtil.NextInt(Random(), 'a', 'z');
field2.StringValue = "" + ch1 + " " + ch2;
writer.AddDocument(doc);
}
IndexReader ir = writer.Reader;
AssertSumDocFreq(ir);
ir.Dispose();
int numDeletions = AtLeast(20);
for (int i = 0; i < numDeletions; i++)
{
writer.DeleteDocuments(new Term("id", "" + Random().Next(numDocs)));
}
writer.ForceMerge(1);
writer.Dispose();
ir = DirectoryReader.Open(dir);
AssertSumDocFreq(ir);
ir.Dispose();
dir.Dispose();
}
/** Creates an index for sorting. */
public void CreateIndex(Directory dir, int numDocs, Random random)
{
IList <int> ids = new List <int>();
for (int i = 0; i < numDocs; i++)
{
ids.Add(i * 10);
}
// shuffle them for indexing
// LUCENENET NOTE: Using LINQ, so we need to reassign the variable with the result
ids = CollectionsHelper.Shuffle(ids);
if (VERBOSE)
{
Console.WriteLine("Shuffled IDs for indexing: " + Arrays.ToString(ids.ToArray()));
}
PositionsTokenStream positions = new PositionsTokenStream();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random));
conf.SetMaxBufferedDocs(4); // create some segments
conf.SetSimilarity(new NormsSimilarity(conf.Similarity)); // for testing norms field
using (RandomIndexWriter writer = new RandomIndexWriter(random, dir, conf))
{
writer.RandomForceMerge = (false);
foreach (int id in ids)
{
writer.AddDocument(Doc(id, positions));
}
// delete some documents
writer.Commit();
foreach (int id in ids)
{
if (random.NextDouble() < 0.2)
{
if (VERBOSE)
{
Console.WriteLine("delete doc_id " + id);
}
writer.DeleteDocuments(new Term(ID_FIELD, id.ToString()));
}
}
}
}
public void TestAdvanceSingleDeletedParentNoChild()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir);
// First doc with 1 children
Document parentDoc = new Document();
parentDoc.Add(NewStringField("parent", "1", Field.Store.NO));
parentDoc.Add(NewStringField("isparent", "yes", Field.Store.NO));
Document childDoc = new Document();
childDoc.Add(NewStringField("child", "1", Field.Store.NO));
w.AddDocuments(Arrays.AsList(childDoc, parentDoc));
parentDoc = new Document();
parentDoc.Add(NewStringField("parent", "2", Field.Store.NO));
parentDoc.Add(NewStringField("isparent", "yes", Field.Store.NO));
w.AddDocuments(Arrays.AsList(parentDoc));
w.DeleteDocuments(new Term("parent", "2"));
parentDoc = new Document();
parentDoc.Add(NewStringField("parent", "2", Field.Store.NO));
parentDoc.Add(NewStringField("isparent", "yes", Field.Store.NO));
childDoc = new Document();
childDoc.Add(NewStringField("child", "2", Field.Store.NO));
w.AddDocuments(Arrays.AsList(childDoc, parentDoc));
IndexReader r = w.Reader;
w.Dispose();
IndexSearcher s = NewSearcher(r);
// Create a filter that defines "parent" documents in the index - in this case resumes
Filter parentsFilter = new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("isparent", "yes"))));
Query parentQuery = new TermQuery(new Term("parent", "2"));
ToChildBlockJoinQuery parentJoinQuery = new ToChildBlockJoinQuery(parentQuery, parentsFilter, Random().NextBoolean());
TopDocs topdocs = s.Search(parentJoinQuery, 3);
assertEquals(1, topdocs.TotalHits);
r.Dispose();
dir.Dispose();
}
public override void SetUp()
{
base.SetUp();
Dir = NewDirectory();
Iw = new RandomIndexWriter(Random(), Dir);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
doc.Add(idField);
// add 500 docs with id 0..499
for (int i = 0; i < 500; i++)
{
idField.StringValue = Convert.ToString(i);
Iw.AddDocument(doc);
}
// delete 20 of them
for (int i = 0; i < 20; i++)
{
Iw.DeleteDocuments(new Term("id", Convert.ToString(Random().Next(Iw.MaxDoc()))));
}
Ir = Iw.Reader;
@is = NewSearcher(Ir);
}
public override void SetUp()
{
base.SetUp();
Dir = NewDirectory();
Iw = new RandomIndexWriter(Random(), Dir, Similarity, TimeZone);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
doc.Add(idField);
// add 500 docs with id 0..499
for (int i = 0; i < 500; i++)
{
idField.SetStringValue(Convert.ToString(i));
Iw.AddDocument(doc);
}
// delete 20 of them
for (int i = 0; i < 20; i++)
{
Iw.DeleteDocuments(new Term("id", Convert.ToString(Random().Next(Iw.MaxDoc))));
}
Ir = Iw.Reader;
@is = NewSearcher(Ir);
}
/// <summary>
/// Tests dv against stored fields with threads (all types + missing)
/// </summary>
// [Test] // LUCENENET NOTE: For now, we are overriding this test in every subclass to pull it into the right context for the subclass
public virtual void TestThreads2()
{
AssumeTrue("Codec does not support GetDocsWithField", DefaultCodecSupportsDocsWithField());
AssumeTrue("Codec does not support SORTED_SET", DefaultCodecSupportsSortedSet());
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Field idField = new StringField("id", "", Field.Store.NO);
Field storedBinField = new StoredField("storedBin", new byte[0]);
Field dvBinField = new BinaryDocValuesField("dvBin", new BytesRef());
Field dvSortedField = new SortedDocValuesField("dvSorted", new BytesRef());
Field storedNumericField = new StoredField("storedNum", "");
Field dvNumericField = new NumericDocValuesField("dvNum", 0);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
int length = TestUtil.NextInt(Random(), 0, 8);
var buffer = new byte[length];
Random().NextBytes(buffer);
storedBinField.BytesValue = new BytesRef(buffer);
dvBinField.BytesValue = new BytesRef(buffer);
dvSortedField.BytesValue = new BytesRef(buffer);
long numericValue = Random().NextLong();
storedNumericField.StringValue = Convert.ToString(numericValue);
dvNumericField.LongValue = numericValue;
Document doc = new Document();
doc.Add(idField);
if (Random().Next(4) > 0)
{
doc.Add(storedBinField);
doc.Add(dvBinField);
doc.Add(dvSortedField);
}
if (Random().Next(4) > 0)
{
doc.Add(storedNumericField);
doc.Add(dvNumericField);
}
int numSortedSetFields = Random().Next(3);
SortedSet<string> values = new SortedSet<string>();
for (int j = 0; j < numSortedSetFields; j++)
{
values.Add(TestUtil.RandomSimpleString(Random()));
}
foreach (string v in values)
{
doc.Add(new SortedSetDocValuesField("dvSortedSet", new BytesRef(v)));
doc.Add(new StoredField("storedSortedSet", v));
}
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
int numThreads = TestUtil.NextInt(Random(), 2, 7);
ThreadClass[] threads = new ThreadClass[numThreads];
CountdownEvent startingGun = new CountdownEvent(1);
for (int i = 0; i < threads.Length; i++)
{
threads[i] = new ThreadAnonymousInnerClassHelper2(this, ir, startingGun);
threads[i].Start();
}
startingGun.Signal();
foreach (ThreadClass t in threads)
{
t.Join();
}
ir.Dispose();
dir.Dispose();
}
public virtual void TestDeleteAllSlowly()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir);
int NUM_DOCS = AtLeast(1000);
IList<int?> ids = new List<int?>(NUM_DOCS);
for (int id = 0; id < NUM_DOCS; id++)
{
ids.Add(id);
}
ids = CollectionsHelper.Shuffle(ids);
foreach (int id in ids)
{
Document doc = new Document();
doc.Add(NewStringField("id", "" + id, Field.Store.NO));
w.AddDocument(doc);
}
ids = CollectionsHelper.Shuffle(ids);
int upto = 0;
while (upto < ids.Count)
{
int left = ids.Count - upto;
int inc = Math.Min(left, TestUtil.NextInt(Random(), 1, 20));
int limit = upto + inc;
while (upto < limit)
{
w.DeleteDocuments(new Term("id", "" + ids[upto++]));
}
IndexReader r = w.Reader;
Assert.AreEqual(NUM_DOCS - upto, r.NumDocs);
r.Dispose();
}
w.Dispose();
dir.Dispose();
}
private void DoTestSortedVsFieldCache(int minLength, int maxLength)
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
Field indexedField = new StringField("indexed", "", Field.Store.NO);
Field dvField = new SortedDocValuesField("dv", new BytesRef());
doc.Add(idField);
doc.Add(indexedField);
doc.Add(dvField);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
int length;
if (minLength == maxLength)
{
length = minLength; // fixed length
}
else
{
length = TestUtil.NextInt(Random(), minLength, maxLength);
}
string value = TestUtil.RandomSimpleString(Random(), length);
indexedField.StringValue = value;
dvField.BytesValue = new BytesRef(value);
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
foreach (AtomicReaderContext context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
SortedDocValues expected = FieldCache.DEFAULT.GetTermsIndex(r, "indexed");
SortedDocValues actual = r.GetSortedDocValues("dv");
AssertEquals(r.MaxDoc, expected, actual);
}
ir.Dispose();
dir.Dispose();
}
private void DoTestSortedVsStoredFields(int minLength, int maxLength)
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
Field storedField = new StoredField("stored", new byte[0]);
Field dvField = new SortedDocValuesField("dv", new BytesRef());
doc.Add(idField);
doc.Add(storedField);
doc.Add(dvField);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
int length;
if (minLength == maxLength)
{
length = minLength; // fixed length
}
else
{
length = TestUtil.NextInt(Random(), minLength, maxLength);
}
var buffer = new byte[length];
Random().NextBytes(buffer);
storedField.BytesValue = new BytesRef(buffer);
dvField.BytesValue = new BytesRef(buffer);
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
foreach (AtomicReaderContext context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
BinaryDocValues docValues = r.GetSortedDocValues("dv");
for (int i = 0; i < r.MaxDoc; i++)
{
BytesRef binaryValue = r.Document(i).GetBinaryValue("stored");
BytesRef scratch = new BytesRef();
docValues.Get(i, scratch);
Assert.AreEqual(binaryValue, scratch);
}
}
ir.Dispose();
dir.Dispose();
}
public void BeforeClass()
{
Random random = Random();
Directory = NewDirectory();
Stopword = "" + RandomChar();
CharacterRunAutomaton stopset = new CharacterRunAutomaton(BasicAutomata.MakeString(Stopword));
Analyzer = new MockAnalyzer(random, MockTokenizer.WHITESPACE, false, stopset);
RandomIndexWriter iw = new RandomIndexWriter(random, Directory, Analyzer, ClassEnvRule.Similarity, ClassEnvRule.TimeZone);
Document doc = new Document();
Field id = new StringField("id", "", Field.Store.NO);
Field field = new TextField("field", "", Field.Store.NO);
doc.Add(id);
doc.Add(field);
// index some docs
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; i++)
{
id.StringValue = Convert.ToString(i);
field.StringValue = RandomFieldContents();
iw.AddDocument(doc);
}
// delete some docs
int numDeletes = numDocs / 20;
for (int i = 0; i < numDeletes; i++)
{
Term toDelete = new Term("id", Convert.ToString(random.Next(numDocs)));
if (random.NextBoolean())
{
iw.DeleteDocuments(toDelete);
}
else
{
iw.DeleteDocuments(new TermQuery(toDelete));
}
}
Reader = iw.Reader;
S1 = NewSearcher(Reader);
S2 = NewSearcher(Reader);
iw.Dispose();
}
// [Test] // LUCENENET NOTE: For now, we are overriding this test in every subclass to pull it into the right context for the subclass
public virtual void TestSortedSetMergeAwayAllValues()
{
AssumeTrue("Codec does not support SORTED_SET", DefaultCodecSupportsSortedSet());
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("id", "0", Field.Store.NO));
iwriter.AddDocument(doc);
doc = new Document();
doc.Add(new StringField("id", "1", Field.Store.NO));
doc.Add(new SortedSetDocValuesField("field", new BytesRef("hello")));
iwriter.AddDocument(doc);
iwriter.Commit();
iwriter.DeleteDocuments(new Term("id", "1"));
iwriter.ForceMerge(1);
DirectoryReader ireader = iwriter.Reader;
iwriter.Dispose();
SortedSetDocValues dv = GetOnlySegmentReader(ireader).GetSortedSetDocValues("field");
Assert.AreEqual(0, dv.ValueCount);
ireader.Dispose();
directory.Dispose();
}
// [Test] // LUCENENET NOTE: For now, we are overriding this test in every subclass to pull it into the right context for the subclass
public virtual void TestSortedMergeAwayAllValues()
{
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("id", "0", Field.Store.NO));
iwriter.AddDocument(doc);
doc = new Document();
doc.Add(new StringField("id", "1", Field.Store.NO));
doc.Add(new SortedDocValuesField("field", new BytesRef("hello")));
iwriter.AddDocument(doc);
iwriter.Commit();
iwriter.DeleteDocuments(new Term("id", "1"));
iwriter.ForceMerge(1);
DirectoryReader ireader = iwriter.Reader;
iwriter.Dispose();
SortedDocValues dv = GetOnlySegmentReader(ireader).GetSortedDocValues("field");
if (DefaultCodecSupportsDocsWithField())
{
Assert.AreEqual(-1, dv.GetOrd(0));
Assert.AreEqual(0, dv.ValueCount);
}
else
{
Assert.AreEqual(0, dv.GetOrd(0));
Assert.AreEqual(1, dv.ValueCount);
BytesRef @ref = new BytesRef();
dv.LookupOrd(0, @ref);
Assert.AreEqual(new BytesRef(), @ref);
}
ireader.Dispose();
directory.Dispose();
}
public void TestNestedDocScoringWithDeletes()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(NoMergePolicy.COMPOUND_FILES));
// Cannot assert this since we use NoMergePolicy:
w.DoRandomForceMergeAssert = false;
IList<Document> docs = new List<Document>();
docs.Add(MakeJob("java", 2007));
docs.Add(MakeJob("python", 2010));
docs.Add(MakeResume("Lisa", "United Kingdom"));
w.AddDocuments(docs);
docs.Clear();
docs.Add(MakeJob("c", 1999));
docs.Add(MakeJob("ruby", 2005));
docs.Add(MakeJob("java", 2006));
docs.Add(MakeResume("Frank", "United States"));
w.AddDocuments(docs);
w.Commit();
IndexSearcher s = NewSearcher(DirectoryReader.Open(dir));
ToParentBlockJoinQuery q = new ToParentBlockJoinQuery(NumericRangeQuery.NewIntRange("year", 1990, 2010, true, true), new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("docType", "resume")))), ScoreMode.Total);
TopDocs topDocs = s.Search(q, 10);
assertEquals(2, topDocs.TotalHits);
assertEquals(6, topDocs.ScoreDocs[0].Doc);
assertEquals(3.0f, topDocs.ScoreDocs[0].Score, 0.0f);
assertEquals(2, topDocs.ScoreDocs[1].Doc);
assertEquals(2.0f, topDocs.ScoreDocs[1].Score, 0.0f);
s.IndexReader.Dispose();
w.DeleteDocuments(new Term("skill", "java"));
w.Dispose();
s = NewSearcher(DirectoryReader.Open(dir));
topDocs = s.Search(q, 10);
assertEquals(2, topDocs.TotalHits);
assertEquals(6, topDocs.ScoreDocs[0].Doc);
assertEquals(2.0f, topDocs.ScoreDocs[0].Score, 0.0f);
assertEquals(2, topDocs.ScoreDocs[1].Doc);
assertEquals(1.0f, topDocs.ScoreDocs[1].Score, 0.0f);
s.IndexReader.Dispose();
dir.Dispose();
}
private void DoTestSortedSetVsStoredFields(int minLength, int maxLength, int maxValuesPerDoc)
{
Directory dir = NewDirectory();
IndexWriterConfig conf = new IndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
Field idField = new StringField("id", Convert.ToString(i), Field.Store.NO);
doc.Add(idField);
int length;
if (minLength == maxLength)
{
length = minLength; // fixed length
}
else
{
length = TestUtil.NextInt(Random(), minLength, maxLength);
}
int numValues = TestUtil.NextInt(Random(), 0, maxValuesPerDoc);
// create a random set of strings
SortedSet<string> values = new SortedSet<string>();
for (int v = 0; v < numValues; v++)
{
values.Add(TestUtil.RandomSimpleString(Random(), length));
}
// add ordered to the stored field
foreach (string v in values)
{
doc.Add(new StoredField("stored", v));
}
// add in any order to the dv field
IList<string> unordered = new List<string>(values);
unordered = CollectionsHelper.Shuffle(unordered);
foreach (string v in unordered)
{
doc.Add(new SortedSetDocValuesField("dv", new BytesRef(v)));
}
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
foreach (AtomicReaderContext context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
SortedSetDocValues docValues = r.GetSortedSetDocValues("dv");
BytesRef scratch = new BytesRef();
for (int i = 0; i < r.MaxDoc; i++)
{
string[] stringValues = r.Document(i).GetValues("stored");
if (docValues != null)
{
docValues.Document = i;
}
for (int j = 0; j < stringValues.Length; j++)
{
Debug.Assert(docValues != null);
long ord = docValues.NextOrd();
Debug.Assert(ord != SortedSetDocValues.NO_MORE_ORDS);
docValues.LookupOrd(ord, scratch);
Assert.AreEqual(stringValues[j], scratch.Utf8ToString());
}
Debug.Assert(docValues == null || docValues.NextOrd() == SortedSetDocValues.NO_MORE_ORDS);
}
}
ir.Dispose();
dir.Dispose();
}
public virtual void TestEnforceDeletions()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergeScheduler(new SerialMergeScheduler()).SetMergePolicy(NewLogMergePolicy(10)));
// asserts below requires no unexpected merges:
// NOTE: cannot use writer.getReader because RIW (on
// flipping a coin) may give us a newly opened reader,
// but we use .reopen on this reader below and expect to
// (must) get an NRT reader:
DirectoryReader reader = DirectoryReader.Open(writer.w, true);
// same reason we don't wrap?
IndexSearcher searcher = NewSearcher(reader, false);
// add a doc, refresh the reader, and check that it's there
Document doc = new Document();
doc.Add(NewStringField("id", "1", Field.Store.YES));
writer.AddDocument(doc);
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
TopDocs docs = searcher.Search(new MatchAllDocsQuery(), 1);
Assert.AreEqual(1, docs.TotalHits, "Should find a hit...");
Filter startFilter = new QueryWrapperFilter(new TermQuery(new Term("id", "1")));
CachingWrapperFilter filter = new CachingWrapperFilter(startFilter);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.IsTrue(filter.SizeInBytes() > 0);
Assert.AreEqual(1, docs.TotalHits, "[query + filter] Should find a hit...");
Query constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
// make sure we get a cache hit when we reopen reader
// that had no change to deletions
// fake delete (deletes nothing):
writer.DeleteDocuments(new Term("foo", "bar"));
IndexReader oldReader = reader;
reader = RefreshReader(reader);
Assert.IsTrue(reader == oldReader);
int missCount = filter.MissCount;
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
// cache hit:
Assert.AreEqual(missCount, filter.MissCount);
// now delete the doc, refresh the reader, and see that it's not there
writer.DeleteDocuments(new Term("id", "1"));
// NOTE: important to hold ref here so GC doesn't clear
// the cache entry! Else the assert below may sometimes
// fail:
oldReader = reader;
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
missCount = filter.MissCount;
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(0, docs.TotalHits, "[query + filter] Should *not* find a hit...");
// cache hit
Assert.AreEqual(missCount, filter.MissCount);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(0, docs.TotalHits, "[just filter] Should *not* find a hit...");
// apply deletes dynamically:
filter = new CachingWrapperFilter(startFilter);
writer.AddDocument(doc);
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(1, docs.TotalHits, "[query + filter] Should find a hit...");
missCount = filter.MissCount;
Assert.IsTrue(missCount > 0);
constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
Assert.AreEqual(missCount, filter.MissCount);
writer.AddDocument(doc);
// NOTE: important to hold ref here so GC doesn't clear
// the cache entry! Else the assert below may sometimes
// fail:
oldReader = reader;
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(2, docs.TotalHits, "[query + filter] Should find 2 hits...");
Assert.IsTrue(filter.MissCount > missCount);
missCount = filter.MissCount;
constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(2, docs.TotalHits, "[just filter] Should find a hit...");
Assert.AreEqual(missCount, filter.MissCount);
// now delete the doc, refresh the reader, and see that it's not there
writer.DeleteDocuments(new Term("id", "1"));
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(0, docs.TotalHits, "[query + filter] Should *not* find a hit...");
// CWF reused the same entry (it dynamically applied the deletes):
Assert.AreEqual(missCount, filter.MissCount);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(0, docs.TotalHits, "[just filter] Should *not* find a hit...");
// CWF reused the same entry (it dynamically applied the deletes):
Assert.AreEqual(missCount, filter.MissCount);
// NOTE: silliness to make sure JRE does not eliminate
// our holding onto oldReader to prevent
// CachingWrapperFilter's WeakHashMap from dropping the
// entry:
Assert.IsTrue(oldReader != null);
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
public virtual void TestZeroTerms()
{
var d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
doc.Add(NewTextField("field", "one two three", Field.Store.NO));
doc = new Document();
doc.Add(NewTextField("field2", "one two three", Field.Store.NO));
w.AddDocument(doc);
w.Commit();
w.DeleteDocuments(new Term("field", "one"));
w.ForceMerge(1);
IndexReader r = w.Reader;
w.Dispose();
Assert.AreEqual(1, r.NumDocs);
Assert.AreEqual(1, r.MaxDoc);
Terms terms = MultiFields.GetTerms(r, "field");
if (terms != null)
{
Assert.IsNull(terms.Iterator(null).Next());
}
r.Dispose();
d.Dispose();
}
public void TestWithDeletions()
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, iwc);
IDictionary<string, Document> docs = GenerateIndexDocuments(AtLeast(100));
Random rand = Random();
List<string> termsToDel = new List<string>();
foreach (Document doc in docs.Values)
{
if (rand.nextBoolean() && termsToDel.size() < docs.size() - 1)
{
termsToDel.Add(doc.Get(FIELD_NAME));
}
writer.AddDocument(doc);
}
writer.Commit();
Term[] delTerms = new Term[termsToDel.size()];
for (int i = 0; i < termsToDel.size(); i++)
{
delTerms[i] = new Term(FIELD_NAME, termsToDel[i]);
}
foreach (Term delTerm in delTerms)
{
writer.DeleteDocuments(delTerm);
}
writer.Commit();
writer.Dispose();
foreach (string termToDel in termsToDel)
{
var toDel = docs[termToDel];
docs.Remove(termToDel);
assertTrue(null != toDel);
}
IndexReader ir = DirectoryReader.Open(dir);
assertTrue("NumDocs should be > 0 but was " + ir.NumDocs, ir.NumDocs > 0);
assertEquals(ir.NumDocs, docs.size());
ValueSource[] toAdd = new ValueSource[] { new LongFieldSource(WEIGHT_FIELD_NAME_1), new LongFieldSource(WEIGHT_FIELD_NAME_2) };
IDictionary dictionary = new DocumentValueSourceDictionary(ir, FIELD_NAME, new SumFloatFunction(toAdd), PAYLOAD_FIELD_NAME);
IInputIterator inputIterator = dictionary.EntryIterator;
BytesRef f;
while ((f = inputIterator.Next()) != null)
{
string field = f.Utf8ToString();
Document doc = docs.ContainsKey(field) ? docs[field] : null;
docs.Remove(field);
long w1 = Convert.ToInt64(doc.GetField(WEIGHT_FIELD_NAME_1).NumericValue);
long w2 = Convert.ToInt64(doc.GetField(WEIGHT_FIELD_NAME_2).NumericValue);
assertTrue(f.equals(new BytesRef(doc.Get(FIELD_NAME))));
assertEquals(inputIterator.Weight, w2 + w1);
assertTrue(inputIterator.Payload.equals(doc.GetField(PAYLOAD_FIELD_NAME).BinaryValue));
}
assertTrue(!docs.Any());
ir.Dispose();
dir.Dispose();
}
public void TestWithDeletions()
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, iwc);
KeyValuePair<List<string>, IDictionary<string, Document>> res = GenerateIndexDocuments(AtLeast(1000), false, false);
IDictionary<string, Document> docs = res.Value;
List<String> invalidDocTerms = res.Key;
Random rand = Random();
List<string> termsToDel = new List<string>();
foreach (Document doc in docs.Values)
{
IndexableField f2 = doc.GetField(FIELD_NAME);
if (rand.nextBoolean() && f2 != null && !invalidDocTerms.Contains(f2.StringValue))
{
termsToDel.Add(doc.Get(FIELD_NAME));
}
writer.AddDocument(doc);
}
writer.Commit();
Term[] delTerms = new Term[termsToDel.size()];
for (int i = 0; i < termsToDel.size(); i++)
{
delTerms[i] = new Term(FIELD_NAME, termsToDel[i]);
}
foreach (Term delTerm in delTerms)
{
writer.DeleteDocuments(delTerm);
}
writer.Commit();
writer.Dispose();
foreach (string termToDel in termsToDel)
{
var toDel = docs[termToDel];
assertTrue(toDel != null);
docs.Remove(termToDel);
}
IndexReader ir = DirectoryReader.Open(dir);
assertEquals(ir.NumDocs, docs.size());
IDictionary dictionary = new DocumentDictionary(ir, FIELD_NAME, WEIGHT_FIELD_NAME);
IInputIterator inputIterator = dictionary.EntryIterator;
BytesRef f;
while ((f = inputIterator.Next()) != null)
{
var field = f.Utf8ToString();
Document doc = docs.ContainsKey(field) ? docs[field] : null;
docs.Remove(field);
assertTrue(f.equals(new BytesRef(doc.Get(FIELD_NAME))));
IndexableField weightField = doc.GetField(WEIGHT_FIELD_NAME);
assertEquals(inputIterator.Weight, (weightField != null) ? Convert.ToInt64(weightField.NumericValue) : 0);
assertEquals(inputIterator.Payload, null);
}
foreach (string invalidTerm in invalidDocTerms)
{
var invalid = docs[invalidTerm];
docs.Remove(invalidTerm);
assertNotNull(invalid);
}
assertTrue(!docs.Any());
ir.Dispose();
dir.Dispose();
}
public void TestMVGroupedFacetingWithDeletes()
{
string groupField = "hotel";
FieldType customType = new FieldType();
customType.IsStored = (true);
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
Random,
dir,
NewIndexWriterConfig(TEST_VERSION_CURRENT,
new MockAnalyzer(Random)).SetMergePolicy(NoMergePolicy.COMPOUND_FILES));
bool useDv = false;
// Cannot assert this since we use NoMergePolicy:
w.DoRandomForceMergeAssert = (false);
// 0
Document doc = new Document();
doc.Add(new StringField("x", "x", Field.Store.NO));
w.AddDocument(doc);
// 1
doc = new Document();
AddField(doc, groupField, "a", useDv);
doc.Add(new StringField("airport", "ams", Field.Store.NO));
w.AddDocument(doc);
w.Commit();
w.DeleteDocuments(new TermQuery(new Term("airport", "ams")));
// 2
doc = new Document();
AddField(doc, groupField, "a", useDv);
doc.Add(new StringField("airport", "ams", Field.Store.NO));
w.AddDocument(doc);
// 3
doc = new Document();
AddField(doc, groupField, "a", useDv);
doc.Add(new StringField("airport", "dus", Field.Store.NO));
w.AddDocument(doc);
// 4
doc = new Document();
AddField(doc, groupField, "b", useDv);
doc.Add(new StringField("airport", "ams", Field.Store.NO));
w.AddDocument(doc);
// 5
doc = new Document();
AddField(doc, groupField, "b", useDv);
doc.Add(new StringField("airport", "ams", Field.Store.NO));
w.AddDocument(doc);
// 6
doc = new Document();
AddField(doc, groupField, "b", useDv);
doc.Add(new StringField("airport", "ams", Field.Store.NO));
w.AddDocument(doc);
w.Commit();
// 7
doc = new Document();
doc.Add(new StringField("x", "x", Field.Store.NO));
w.AddDocument(doc);
w.Commit();
w.Dispose();
IndexSearcher indexSearcher = NewSearcher(DirectoryReader.Open(dir));
AbstractGroupFacetCollector groupedAirportFacetCollector = CreateRandomCollector(groupField, "airport", null, true);
indexSearcher.Search(new MatchAllDocsQuery(), groupedAirportFacetCollector);
TermGroupFacetCollector.GroupedFacetResult airportResult = groupedAirportFacetCollector.MergeSegmentResults(10, 0, false);
assertEquals(3, airportResult.TotalCount);
assertEquals(1, airportResult.TotalMissingCount);
IList <TermGroupFacetCollector.FacetEntry> entries = airportResult.GetFacetEntries(0, 10);
assertEquals(2, entries.size());
assertEquals("ams", entries[0].Value.Utf8ToString());
assertEquals(2, entries[0].Count);
assertEquals("dus", entries[1].Value.Utf8ToString());
assertEquals(1, entries[1].Count);
indexSearcher.IndexReader.Dispose();
dir.Dispose();
}
public void TestThreads()
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
Field storedBinField = new StoredField("storedBin", new byte[0]);
Field dvBinField = new BinaryDocValuesField("dvBin", new BytesRef());
Field dvSortedField = new SortedDocValuesField("dvSorted", new BytesRef());
Field storedNumericField = new StoredField("storedNum", "");
Field dvNumericField = new NumericDocValuesField("dvNum", 0);
doc.Add(idField);
doc.Add(storedBinField);
doc.Add(dvBinField);
doc.Add(dvSortedField);
doc.Add(storedNumericField);
doc.Add(dvNumericField);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
int length = TestUtil.NextInt(Random(), 0, 8);
var buffer = new byte[length];
Random().NextBytes(buffer);
storedBinField.BytesValue = new BytesRef(buffer);
dvBinField.BytesValue = new BytesRef(buffer);
dvSortedField.BytesValue = new BytesRef(buffer);
long numericValue = Random().NextLong();
storedNumericField.StringValue = Convert.ToString(numericValue);
dvNumericField.LongValue = numericValue;
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
int numThreads = TestUtil.NextInt(Random(), 2, 7);
ThreadClass[] threads = new ThreadClass[numThreads];
CountdownEvent startingGun = new CountdownEvent(1);
for (int i = 0; i < threads.Length; i++)
{
threads[i] = new ThreadAnonymousInnerClassHelper(this, ir, startingGun);
threads[i].Start();
}
startingGun.Signal();
foreach (ThreadClass t in threads)
{
t.Join();
}
ir.Dispose();
dir.Dispose();
}
public virtual void TestForceMergeDeletesMaxSegSize()
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
TieredMergePolicy tmp = new TieredMergePolicy();
tmp.MaxMergedSegmentMB = 0.01;
tmp.ForceMergeDeletesPctAllowed = 0.0;
conf.SetMergePolicy(tmp);
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, conf);
w.RandomForceMerge = false;
int numDocs = AtLeast(200);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
doc.Add(NewStringField("id", "" + i, Field.Store.NO));
doc.Add(NewTextField("content", "aaa " + i, Field.Store.NO));
w.AddDocument(doc);
}
w.ForceMerge(1);
IndexReader r = w.Reader;
Assert.AreEqual(numDocs, r.MaxDoc);
Assert.AreEqual(numDocs, r.NumDocs);
r.Dispose();
if (VERBOSE)
{
Console.WriteLine("\nTEST: delete doc");
}
w.DeleteDocuments(new Term("id", "" + (42 + 17)));
r = w.Reader;
Assert.AreEqual(numDocs, r.MaxDoc);
Assert.AreEqual(numDocs - 1, r.NumDocs);
r.Dispose();
w.ForceMergeDeletes();
r = w.Reader;
Assert.AreEqual(numDocs - 1, r.MaxDoc);
Assert.AreEqual(numDocs - 1, r.NumDocs);
r.Dispose();
w.Dispose();
dir.Dispose();
}
// [Test] // LUCENENET NOTE: For now, we are overriding this test in every subclass to pull it into the right context for the subclass
public virtual void TestMerge()
{
RandomDocumentFactory docFactory = new RandomDocumentFactory(this, 5, 20);
int numDocs = AtLeast(100);
int numDeletes = Random().Next(numDocs);
HashSet<int?> deletes = new HashSet<int?>();
while (deletes.Count < numDeletes)
{
deletes.Add(Random().Next(numDocs));
}
foreach (Options options in ValidOptions())
{
RandomDocument[] docs = new RandomDocument[numDocs];
for (int i = 0; i < numDocs; ++i)
{
docs[i] = docFactory.NewDocument(TestUtil.NextInt(Random(), 1, 3), AtLeast(10), options);
}
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, ClassEnvRule.Similarity, ClassEnvRule.TimeZone);
for (int i = 0; i < numDocs; ++i)
{
writer.AddDocument(AddId(docs[i].ToDocument(), "" + i));
if (Rarely())
{
writer.Commit();
}
}
foreach (int delete in deletes)
{
writer.DeleteDocuments(new Term("id", "" + delete));
}
// merge with deletes
writer.ForceMerge(1);
IndexReader reader = writer.Reader;
for (int i = 0; i < numDocs; ++i)
{
if (!deletes.Contains(i))
{
int docID = DocID(reader, "" + i);
AssertEquals(docs[i], reader.GetTermVectors(docID));
}
}
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
}
public void TestRandom()
{
// We build two indices at once: one normalized (which
// ToParentBlockJoinQuery/Collector,
// ToChildBlockJoinQuery can query) and the other w/
// the same docs, just fully denormalized:
Directory dir = NewDirectory();
Directory joinDir = NewDirectory();
int numParentDocs = TestUtil.NextInt(Random(), 100 * RANDOM_MULTIPLIER, 300 * RANDOM_MULTIPLIER);
//final int numParentDocs = 30;
// Values for parent fields:
string[][] parentFields = GetRandomFields(numParentDocs / 2);
// Values for child fields:
string[][] childFields = GetRandomFields(numParentDocs);
bool doDeletes = Random().NextBoolean();
IList<int> toDelete = new List<int>();
// TODO: parallel star join, nested join cases too!
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
RandomIndexWriter joinW = new RandomIndexWriter(Random(), joinDir, Similarity, TimeZone);
for (int parentDocID = 0; parentDocID < numParentDocs; parentDocID++)
{
Document parentDoc = new Document();
Document parentJoinDoc = new Document();
Field id = NewStringField("parentID", "" + parentDocID, Field.Store.YES);
parentDoc.Add(id);
parentJoinDoc.Add(id);
parentJoinDoc.Add(NewStringField("isParent", "x", Field.Store.NO));
for (int field = 0; field < parentFields.Length; field++)
{
if (Random().NextDouble() < 0.9)
{
Field f = NewStringField("parent" + field, parentFields[field][Random().Next(parentFields[field].Length)], Field.Store.NO);
parentDoc.Add(f);
parentJoinDoc.Add(f);
}
}
if (doDeletes)
{
parentDoc.Add(NewStringField("blockID", "" + parentDocID, Field.Store.NO));
parentJoinDoc.Add(NewStringField("blockID", "" + parentDocID, Field.Store.NO));
}
IList<Document> joinDocs = new List<Document>();
if (VERBOSE)
{
StringBuilder sb = new StringBuilder();
sb.Append("parentID=").Append(parentDoc.Get("parentID"));
for (int fieldID = 0; fieldID < parentFields.Length; fieldID++)
{
string parent = parentDoc.Get("parent" + fieldID);
if (parent != null)
{
sb.Append(" parent" + fieldID + "=" + parent);
}
}
Console.WriteLine(" " + sb);
}
int numChildDocs = TestUtil.NextInt(Random(), 1, 20);
for (int childDocID = 0; childDocID < numChildDocs; childDocID++)
{
// Denormalize: copy all parent fields into child doc:
Document childDoc = TestUtil.CloneDocument(parentDoc);
Document joinChildDoc = new Document();
joinDocs.Add(joinChildDoc);
Field childID = NewStringField("childID", "" + childDocID, Field.Store.YES);
childDoc.Add(childID);
joinChildDoc.Add(childID);
for (int childFieldID = 0; childFieldID < childFields.Length; childFieldID++)
{
if (Random().NextDouble() < 0.9)
{
Field f = NewStringField("child" + childFieldID, childFields[childFieldID][Random().Next(childFields[childFieldID].Length)], Field.Store.NO);
childDoc.Add(f);
joinChildDoc.Add(f);
}
}
if (VERBOSE)
{
StringBuilder sb = new StringBuilder();
sb.Append("childID=").Append(joinChildDoc.Get("childID"));
for (int fieldID = 0; fieldID < childFields.Length; fieldID++)
{
string child = joinChildDoc.Get("child" + fieldID);
if (child != null)
{
sb.Append(" child" + fieldID + "=" + child);
}
}
Console.WriteLine(" " + sb);
}
if (doDeletes)
{
joinChildDoc.Add(NewStringField("blockID", "" + parentDocID, Field.Store.NO));
}
w.AddDocument(childDoc);
}
// Parent last:
joinDocs.Add(parentJoinDoc);
joinW.AddDocuments(joinDocs);
if (doDeletes && Random().Next(30) == 7)
{
toDelete.Add(parentDocID);
}
}
foreach (int deleteID in toDelete)
{
if (VERBOSE)
{
Console.WriteLine("DELETE parentID=" + deleteID);
}
w.DeleteDocuments(new Term("blockID", "" + deleteID));
joinW.DeleteDocuments(new Term("blockID", "" + deleteID));
}
IndexReader r = w.Reader;
w.Dispose();
IndexReader joinR = joinW.Reader;
joinW.Dispose();
if (VERBOSE)
{
Console.WriteLine("TEST: reader=" + r);
Console.WriteLine("TEST: joinReader=" + joinR);
for (int docIDX = 0; docIDX < joinR.MaxDoc; docIDX++)
{
Console.WriteLine(" docID=" + docIDX + " doc=" + joinR.Document(docIDX));
}
}
IndexSearcher s = NewSearcher(r);
IndexSearcher joinS = new IndexSearcher(joinR);
Filter parentsFilter = new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("isParent", "x"))));
int iters = 200 * RANDOM_MULTIPLIER;
for (int iter = 0; iter < iters; iter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: iter=" + (1 + iter) + " of " + iters);
}
Query childQuery;
if (Random().Next(3) == 2)
{
int childFieldID = Random().Next(childFields.Length);
childQuery = new TermQuery(new Term("child" + childFieldID, childFields[childFieldID][Random().Next(childFields[childFieldID].Length)]));
}
else if (Random().Next(3) == 2)
{
BooleanQuery bq = new BooleanQuery();
childQuery = bq;
int numClauses = TestUtil.NextInt(Random(), 2, 4);
bool didMust = false;
for (int clauseIDX = 0; clauseIDX < numClauses; clauseIDX++)
{
Query clause;
BooleanClause.Occur occur;
if (!didMust && Random().NextBoolean())
{
occur = Random().NextBoolean() ? BooleanClause.Occur.MUST : BooleanClause.Occur.MUST_NOT;
clause = new TermQuery(RandomChildTerm(childFields[0]));
didMust = true;
}
else
{
occur = BooleanClause.Occur.SHOULD;
int childFieldID = TestUtil.NextInt(Random(), 1, childFields.Length - 1);
clause = new TermQuery(new Term("child" + childFieldID, childFields[childFieldID][Random().Next(childFields[childFieldID].Length)]));
}
bq.Add(clause, occur);
}
}
else
{
BooleanQuery bq = new BooleanQuery();
childQuery = bq;
bq.Add(new TermQuery(RandomChildTerm(childFields[0])), BooleanClause.Occur.MUST);
int childFieldID = TestUtil.NextInt(Random(), 1, childFields.Length - 1);
bq.Add(new TermQuery(new Term("child" + childFieldID, childFields[childFieldID][Random().Next(childFields[childFieldID].Length)])), Random().NextBoolean() ? BooleanClause.Occur.MUST : BooleanClause.Occur.MUST_NOT);
}
int x = Random().Next(4);
ScoreMode agg;
if (x == 0)
{
agg = ScoreMode.None;
}
else if (x == 1)
{
agg = ScoreMode.Max;
}
else if (x == 2)
{
agg = ScoreMode.Total;
}
else
{
agg = ScoreMode.Avg;
}
ToParentBlockJoinQuery childJoinQuery = new ToParentBlockJoinQuery(childQuery, parentsFilter, agg);
// To run against the block-join index:
Query parentJoinQuery;
// Same query as parentJoinQuery, but to run against
// the fully denormalized index (so we can compare
// results):
Query parentQuery;
if (Random().NextBoolean())
{
parentQuery = childQuery;
parentJoinQuery = childJoinQuery;
}
else
{
// AND parent field w/ child field
BooleanQuery bq = new BooleanQuery();
parentJoinQuery = bq;
Term parentTerm = RandomParentTerm(parentFields[0]);
if (Random().NextBoolean())
{
bq.Add(childJoinQuery, BooleanClause.Occur.MUST);
bq.Add(new TermQuery(parentTerm), BooleanClause.Occur.MUST);
}
else
{
bq.Add(new TermQuery(parentTerm), BooleanClause.Occur.MUST);
bq.Add(childJoinQuery, BooleanClause.Occur.MUST);
}
BooleanQuery bq2 = new BooleanQuery();
parentQuery = bq2;
if (Random().NextBoolean())
{
bq2.Add(childQuery, BooleanClause.Occur.MUST);
bq2.Add(new TermQuery(parentTerm), BooleanClause.Occur.MUST);
}
else
{
bq2.Add(new TermQuery(parentTerm), BooleanClause.Occur.MUST);
bq2.Add(childQuery, BooleanClause.Occur.MUST);
}
}
Sort parentSort = GetRandomSort("parent", parentFields.Length);
Sort childSort = GetRandomSort("child", childFields.Length);
if (VERBOSE)
{
Console.WriteLine("\nTEST: query=" + parentQuery + " joinQuery=" + parentJoinQuery + " parentSort=" + parentSort + " childSort=" + childSort);
}
// Merge both sorts:
IList<SortField> sortFields = new List<SortField>(Arrays.AsList(parentSort.GetSort()));
sortFields.AddRange(Arrays.AsList(childSort.GetSort()));
Sort parentAndChildSort = new Sort(sortFields.ToArray());
TopDocs results = s.Search(parentQuery, null, r.NumDocs, parentAndChildSort);
if (VERBOSE)
{
Console.WriteLine("\nTEST: normal index gets " + results.TotalHits + " hits");
ScoreDoc[] hits = results.ScoreDocs;
for (int hitIDX = 0; hitIDX < hits.Length; hitIDX++)
{
Document doc = s.Doc(hits[hitIDX].Doc);
//System.out.println(" score=" + hits[hitIDX].Score + " parentID=" + doc.Get("parentID") + " childID=" + doc.Get("childID") + " (docID=" + hits[hitIDX].Doc + ")");
Console.WriteLine(" parentID=" + doc.Get("parentID") + " childID=" + doc.Get("childID") + " (docID=" + hits[hitIDX].Doc + ")");
FieldDoc fd = (FieldDoc)hits[hitIDX];
if (fd.Fields != null)
{
Console.Write(" ");
foreach (object o in fd.Fields)
{
if (o is BytesRef)
{
Console.Write(((BytesRef)o).Utf8ToString() + " ");
}
else
{
Console.Write(o + " ");
}
}
Console.WriteLine();
}
}
}
bool trackScores;
bool trackMaxScore;
if (agg == ScoreMode.None)
{
trackScores = false;
trackMaxScore = false;
}
else
{
trackScores = Random().NextBoolean();
trackMaxScore = Random().NextBoolean();
}
ToParentBlockJoinCollector c = new ToParentBlockJoinCollector(parentSort, 10, trackScores, trackMaxScore);
joinS.Search(parentJoinQuery, c);
int hitsPerGroup = TestUtil.NextInt(Random(), 1, 20);
//final int hitsPerGroup = 100;
TopGroups<int> joinResults = c.GetTopGroups(childJoinQuery, childSort, 0, hitsPerGroup, 0, true);
if (VERBOSE)
{
Console.WriteLine("\nTEST: block join index gets " + (joinResults == null ? 0 : joinResults.Groups.Length) + " groups; hitsPerGroup=" + hitsPerGroup);
if (joinResults != null)
{
IGroupDocs<int>[] groups = joinResults.Groups;
for (int groupIDX = 0; groupIDX < groups.Length; groupIDX++)
{
IGroupDocs<int> group = groups[groupIDX];
if (group.GroupSortValues != null)
{
Console.Write(" ");
foreach (object o in group.GroupSortValues)
{
if (o is BytesRef)
{
Console.Write(((BytesRef)o).Utf8ToString() + " ");
}
else
{
Console.Write(o + " ");
}
}
Console.WriteLine();
}
assertNotNull(group.GroupValue);
Document parentDoc = joinS.Doc(group.GroupValue);
Console.WriteLine(" group parentID=" + parentDoc.Get("parentID") + " (docID=" + group.GroupValue + ")");
for (int hitIDX = 0; hitIDX < group.ScoreDocs.Length; hitIDX++)
{
Document doc = joinS.Doc(group.ScoreDocs[hitIDX].Doc);
//System.out.println(" score=" + group.ScoreDocs[hitIDX].Score + " childID=" + doc.Get("childID") + " (docID=" + group.ScoreDocs[hitIDX].Doc + ")");
Console.WriteLine(" childID=" + doc.Get("childID") + " child0=" + doc.Get("child0") + " (docID=" + group.ScoreDocs[hitIDX].Doc + ")");
}
}
}
}
if (results.TotalHits == 0)
{
assertNull(joinResults);
}
else
{
CompareHits(r, joinR, results, joinResults);
TopDocs b = joinS.Search(childJoinQuery, 10);
foreach (ScoreDoc hit in b.ScoreDocs)
{
Explanation explanation = joinS.Explain(childJoinQuery, hit.Doc);
Document document = joinS.Doc(hit.Doc - 1);
int childId = Convert.ToInt32(document.Get("childID"));
assertTrue(explanation.IsMatch);
assertEquals(hit.Score, explanation.Value, 0.0f);
assertEquals(string.Format("Score based on child doc range from {0} to {1}", hit.Doc - 1 - childId, hit.Doc - 1), explanation.Description);
}
}
// Test joining in the opposite direction (parent to
// child):
// Get random query against parent documents:
Query parentQuery2;
if (Random().Next(3) == 2)
{
int fieldID = Random().Next(parentFields.Length);
parentQuery2 = new TermQuery(new Term("parent" + fieldID, parentFields[fieldID][Random().Next(parentFields[fieldID].Length)]));
}
else if (Random().Next(3) == 2)
{
BooleanQuery bq = new BooleanQuery();
parentQuery2 = bq;
int numClauses = TestUtil.NextInt(Random(), 2, 4);
bool didMust = false;
for (int clauseIDX = 0; clauseIDX < numClauses; clauseIDX++)
{
Query clause;
BooleanClause.Occur occur;
if (!didMust && Random().NextBoolean())
{
occur = Random().NextBoolean() ? BooleanClause.Occur.MUST : BooleanClause.Occur.MUST_NOT;
clause = new TermQuery(RandomParentTerm(parentFields[0]));
didMust = true;
}
else
{
occur = BooleanClause.Occur.SHOULD;
int fieldID = TestUtil.NextInt(Random(), 1, parentFields.Length - 1);
clause = new TermQuery(new Term("parent" + fieldID, parentFields[fieldID][Random().Next(parentFields[fieldID].Length)]));
}
bq.Add(clause, occur);
}
}
else
{
BooleanQuery bq = new BooleanQuery();
parentQuery2 = bq;
bq.Add(new TermQuery(RandomParentTerm(parentFields[0])), BooleanClause.Occur.MUST);
int fieldID = TestUtil.NextInt(Random(), 1, parentFields.Length - 1);
bq.Add(new TermQuery(new Term("parent" + fieldID, parentFields[fieldID][Random().Next(parentFields[fieldID].Length)])), Random().NextBoolean() ? BooleanClause.Occur.MUST : BooleanClause.Occur.MUST_NOT);
}
if (VERBOSE)
{
Console.WriteLine("\nTEST: top down: parentQuery2=" + parentQuery2);
}
// Maps parent query to child docs:
ToChildBlockJoinQuery parentJoinQuery2 = new ToChildBlockJoinQuery(parentQuery2, parentsFilter, Random().NextBoolean());
// To run against the block-join index:
Query childJoinQuery2;
// Same query as parentJoinQuery, but to run against
// the fully denormalized index (so we can compare
// results):
Query childQuery2;
// apply a filter to children
Filter childFilter2, childJoinFilter2;
if (Random().NextBoolean())
{
childQuery2 = parentQuery2;
childJoinQuery2 = parentJoinQuery2;
childFilter2 = null;
childJoinFilter2 = null;
}
else
{
Term childTerm = RandomChildTerm(childFields[0]);
if (Random().NextBoolean()) // filtered case
{
childJoinQuery2 = parentJoinQuery2;
Filter f = new QueryWrapperFilter(new TermQuery(childTerm));
childJoinFilter2 = Random().NextBoolean() ? new FixedBitSetCachingWrapperFilter(f) : f;
}
else
{
childJoinFilter2 = null;
// AND child field w/ parent query:
BooleanQuery bq = new BooleanQuery();
childJoinQuery2 = bq;
if (Random().NextBoolean())
{
bq.Add(parentJoinQuery2, BooleanClause.Occur.MUST);
bq.Add(new TermQuery(childTerm), BooleanClause.Occur.MUST);
}
else
{
bq.Add(new TermQuery(childTerm), BooleanClause.Occur.MUST);
bq.Add(parentJoinQuery2, BooleanClause.Occur.MUST);
}
}
if (Random().NextBoolean()) // filtered case
{
childQuery2 = parentQuery2;
Filter f = new QueryWrapperFilter(new TermQuery(childTerm));
childFilter2 = Random().NextBoolean() ? new FixedBitSetCachingWrapperFilter(f) : f;
}
else
{
childFilter2 = null;
BooleanQuery bq2 = new BooleanQuery();
childQuery2 = bq2;
if (Random().NextBoolean())
{
bq2.Add(parentQuery2, BooleanClause.Occur.MUST);
bq2.Add(new TermQuery(childTerm), BooleanClause.Occur.MUST);
}
else
{
bq2.Add(new TermQuery(childTerm), BooleanClause.Occur.MUST);
bq2.Add(parentQuery2, BooleanClause.Occur.MUST);
}
}
}
Sort childSort2 = GetRandomSort("child", childFields.Length);
// Search denormalized index:
if (VERBOSE)
{
Console.WriteLine("TEST: run top down query=" + childQuery2 + " filter=" + childFilter2 + " sort=" + childSort2);
}
TopDocs results2 = s.Search(childQuery2, childFilter2, r.NumDocs, childSort2);
if (VERBOSE)
{
Console.WriteLine(" " + results2.TotalHits + " totalHits:");
foreach (ScoreDoc sd in results2.ScoreDocs)
{
Document doc = s.Doc(sd.Doc);
Console.WriteLine(" childID=" + doc.Get("childID") + " parentID=" + doc.Get("parentID") + " docID=" + sd.Doc);
}
}
// Search join index:
if (VERBOSE)
{
Console.WriteLine("TEST: run top down join query=" + childJoinQuery2 + " filter=" + childJoinFilter2 + " sort=" + childSort2);
}
TopDocs joinResults2 = joinS.Search(childJoinQuery2, childJoinFilter2, joinR.NumDocs, childSort2);
if (VERBOSE)
{
Console.WriteLine(" " + joinResults2.TotalHits + " totalHits:");
foreach (ScoreDoc sd in joinResults2.ScoreDocs)
{
Document doc = joinS.Doc(sd.Doc);
Document parentDoc = GetParentDoc(joinR, parentsFilter, sd.Doc);
Console.WriteLine(" childID=" + doc.Get("childID") + " parentID=" + parentDoc.Get("parentID") + " docID=" + sd.Doc);
}
}
CompareChildHits(r, joinR, results2, joinResults2);
}
r.Dispose();
joinR.Dispose();
dir.Dispose();
joinDir.Dispose();
}
public virtual void TestPostings()
{
Directory dir = NewFSDirectory(CreateTempDir("postings"));
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetCodec(Codec.ForName("Lucene40"));
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
Document doc = new Document();
// id field
FieldType idType = new FieldType(StringField.TYPE_NOT_STORED);
idType.StoreTermVectors = true;
Field idField = new Field("id", "", idType);
doc.Add(idField);
// title field: short text field
FieldType titleType = new FieldType(TextField.TYPE_NOT_STORED);
titleType.StoreTermVectors = true;
titleType.StoreTermVectorPositions = true;
titleType.StoreTermVectorOffsets = true;
titleType.IndexOptions = IndexOptions();
Field titleField = new Field("title", "", titleType);
doc.Add(titleField);
// body field: long text field
FieldType bodyType = new FieldType(TextField.TYPE_NOT_STORED);
bodyType.StoreTermVectors = true;
bodyType.StoreTermVectorPositions = true;
bodyType.StoreTermVectorOffsets = true;
bodyType.IndexOptions = IndexOptions();
Field bodyField = new Field("body", "", bodyType);
doc.Add(bodyField);
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
titleField.StringValue = FieldValue(1);
bodyField.StringValue = FieldValue(3);
iw.AddDocument(doc);
if (Random().Next(20) == 0)
{
iw.DeleteDocuments(new Term("id", Convert.ToString(i)));
}
}
if (Random().NextBoolean())
{
// delete 1-100% of docs
iw.DeleteDocuments(new Term("title", Terms[Random().Next(Terms.Length)]));
}
iw.Dispose();
dir.Dispose(); // checkindex
}
private void CreateRandomIndexes()
{
dir1 = NewDirectory();
dir2 = NewDirectory();
int numDocs = AtLeast(150);
int numTerms = TestUtil.NextInt(Random(), 1, numDocs / 5);
ISet<string> randomTerms = new HashSet<string>();
while (randomTerms.size() < numTerms)
{
randomTerms.add(TestUtil.RandomSimpleString(Random()));
}
terms = new List<string>(randomTerms);
long seed = Random().NextLong();
IndexWriterConfig iwc1 = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new Random((int)seed)));
IndexWriterConfig iwc2 = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new Random((int)seed)));
iwc2.SetMergePolicy(NewSortingMergePolicy(sort));
RandomIndexWriter iw1 = new RandomIndexWriter(new Random((int)seed), dir1, iwc1);
RandomIndexWriter iw2 = new RandomIndexWriter(new Random((int)seed), dir2, iwc2);
for (int i = 0; i < numDocs; ++i)
{
if (Random().nextInt(5) == 0 && i != numDocs - 1)
{
string term = RandomInts.RandomFrom(Random(), terms);
iw1.DeleteDocuments(new Term("s", term));
iw2.DeleteDocuments(new Term("s", term));
}
Document doc = randomDocument();
iw1.AddDocument(doc);
iw2.AddDocument(doc);
if (Random().nextInt(8) == 0)
{
iw1.Commit();
iw2.Commit();
}
}
// Make sure we have something to merge
iw1.Commit();
iw2.Commit();
Document doc2 = randomDocument();
// NOTE: don't use RIW.addDocument directly, since it sometimes commits
// which may trigger a merge, at which case forceMerge may not do anything.
// With field updates this is a problem, since the updates can go into the
// single segment in the index, and threefore the index won't be sorted.
// This hurts the assumption of the test later on, that the index is sorted
// by SortingMP.
iw1.w.AddDocument(doc2);
iw2.w.AddDocument(doc2);
if (DefaultCodecSupportsFieldUpdates())
{
// update NDV of docs belonging to one term (covers many documents)
long value = Random().NextLong();
string term = RandomInts.RandomFrom(Random(), terms);
iw1.w.UpdateNumericDocValue(new Term("s", term), "ndv", value);
iw2.w.UpdateNumericDocValue(new Term("s", term), "ndv", value);
}
iw1.ForceMerge(1);
iw2.ForceMerge(1);
iw1.Dispose();
iw2.Dispose();
reader = DirectoryReader.Open(dir1);
sortedReader = DirectoryReader.Open(dir2);
}
private void CreateRandomIndexes()
{
dir1 = NewDirectory();
dir2 = NewDirectory();
int numDocs = AtLeast(150);
int numTerms = TestUtil.NextInt32(Random, 1, numDocs / 5);
ISet <string> randomTerms = new JCG.HashSet <string>();
while (randomTerms.size() < numTerms)
{
randomTerms.add(TestUtil.RandomSimpleString(Random));
}
terms = new JCG.List <string>(randomTerms);
long seed = Random.NextInt64();
IndexWriterConfig iwc1 = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new Random((int)seed)));
IndexWriterConfig iwc2 = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(new Random((int)seed)));
iwc2.SetMergePolicy(NewSortingMergePolicy(sort));
RandomIndexWriter iw1 = new RandomIndexWriter(new Random((int)seed), dir1, iwc1);
RandomIndexWriter iw2 = new RandomIndexWriter(new Random((int)seed), dir2, iwc2);
for (int i = 0; i < numDocs; ++i)
{
if (Random.nextInt(5) == 0 && i != numDocs - 1)
{
string term = RandomPicks.RandomFrom(Random, terms);
iw1.DeleteDocuments(new Term("s", term));
iw2.DeleteDocuments(new Term("s", term));
}
Document doc = randomDocument();
iw1.AddDocument(doc);
iw2.AddDocument(doc);
if (Random.nextInt(8) == 0)
{
iw1.Commit();
iw2.Commit();
}
}
// Make sure we have something to merge
iw1.Commit();
iw2.Commit();
Document doc2 = randomDocument();
// NOTE: don't use RIW.addDocument directly, since it sometimes commits
// which may trigger a merge, at which case forceMerge may not do anything.
// With field updates this is a problem, since the updates can go into the
// single segment in the index, and threefore the index won't be sorted.
// This hurts the assumption of the test later on, that the index is sorted
// by SortingMP.
iw1.IndexWriter.AddDocument(doc2);
iw2.IndexWriter.AddDocument(doc2);
if (DefaultCodecSupportsFieldUpdates)
{
// update NDV of docs belonging to one term (covers many documents)
long value = Random.NextInt64();
string term = RandomPicks.RandomFrom(Random, terms);
iw1.IndexWriter.UpdateNumericDocValue(new Term("s", term), "ndv", value);
iw2.IndexWriter.UpdateNumericDocValue(new Term("s", term), "ndv", value);
}
iw1.ForceMerge(1);
iw2.ForceMerge(1);
iw1.Dispose();
iw2.Dispose();
reader = DirectoryReader.Open(dir1);
sortedReader = DirectoryReader.Open(dir2);
}
public void TestWithDeletions()
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, iwc);
IDictionary <string, Document> docs = GenerateIndexDocuments(AtLeast(100));
Random rand = Random();
List <string> termsToDel = new List <string>();
foreach (Document doc in docs.Values)
{
if (rand.nextBoolean() && termsToDel.size() < docs.size() - 1)
{
termsToDel.Add(doc.Get(FIELD_NAME));
}
writer.AddDocument(doc);
}
writer.Commit();
Term[] delTerms = new Term[termsToDel.size()];
for (int i = 0; i < termsToDel.size(); i++)
{
delTerms[i] = new Term(FIELD_NAME, termsToDel[i]);
}
foreach (Term delTerm in delTerms)
{
writer.DeleteDocuments(delTerm);
}
writer.Commit();
writer.Dispose();
foreach (string termToDel in termsToDel)
{
var toDel = docs[termToDel];
docs.Remove(termToDel);
assertTrue(null != toDel);
}
IndexReader ir = DirectoryReader.Open(dir);
assertTrue("NumDocs should be > 0 but was " + ir.NumDocs, ir.NumDocs > 0);
assertEquals(ir.NumDocs, docs.size());
ValueSource[] toAdd = new ValueSource[] { new Int64FieldSource(WEIGHT_FIELD_NAME_1), new Int64FieldSource(WEIGHT_FIELD_NAME_2) };
IDictionary dictionary = new DocumentValueSourceDictionary(ir, FIELD_NAME, new SumSingleFunction(toAdd), PAYLOAD_FIELD_NAME);
IInputIterator inputIterator = dictionary.GetEntryIterator();
BytesRef f;
while ((f = inputIterator.Next()) != null)
{
string field = f.Utf8ToString();
Document doc = docs.ContainsKey(field) ? docs[field] : null;
docs.Remove(field);
long w1 = doc.GetField(WEIGHT_FIELD_NAME_1).GetInt64ValueOrDefault();
long w2 = doc.GetField(WEIGHT_FIELD_NAME_2).GetInt64ValueOrDefault();
assertTrue(f.equals(new BytesRef(doc.Get(FIELD_NAME))));
assertEquals(inputIterator.Weight, w2 + w1);
assertTrue(inputIterator.Payload.equals(doc.GetField(PAYLOAD_FIELD_NAME).GetBinaryValue()));
}
assertTrue(!docs.Any());
ir.Dispose();
dir.Dispose();
}
private void DoTestMissingVsFieldCache(LongProducer longs)
{
AssumeTrue("Codec does not support GetDocsWithField", DefaultCodecSupportsDocsWithField());
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Field idField = new StringField("id", "", Field.Store.NO);
Field indexedField = NewStringField("indexed", "", Field.Store.NO);
Field dvField = new NumericDocValuesField("dv", 0);
// index some docs
int numDocs = AtLeast(300);
// numDocs should be always > 256 so that in case of a codec that optimizes
// for numbers of values <= 256, all storage layouts are tested
Debug.Assert(numDocs > 256);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
long value = longs.Next();
indexedField.StringValue = Convert.ToString(value);
dvField.LongValue = value;
Document doc = new Document();
doc.Add(idField);
// 1/4 of the time we neglect to add the fields
if (Random().Next(4) > 0)
{
doc.Add(indexedField);
doc.Add(dvField);
}
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
// merge some segments and ensure that at least one of them has more than
// 256 values
writer.ForceMerge(numDocs / 256);
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
foreach (var context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
Bits expected = FieldCache.DEFAULT.GetDocsWithField(r, "indexed");
Bits actual = FieldCache.DEFAULT.GetDocsWithField(r, "dv");
AssertEquals(expected, actual);
}
ir.Dispose();
dir.Dispose();
}
public void TestWithDeletions()
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, iwc);
KeyValuePair <List <string>, IDictionary <string, Document> > res = GenerateIndexDocuments(AtLeast(1000), false, false);
IDictionary <string, Document> docs = res.Value;
List <String> invalidDocTerms = res.Key;
Random rand = Random();
List <string> termsToDel = new List <string>();
foreach (Document doc in docs.Values)
{
IIndexableField f2 = doc.GetField(FIELD_NAME);
if (rand.nextBoolean() && f2 != null && !invalidDocTerms.Contains(f2.GetStringValue()))
{
termsToDel.Add(doc.Get(FIELD_NAME));
}
writer.AddDocument(doc);
}
writer.Commit();
Term[] delTerms = new Term[termsToDel.size()];
for (int i = 0; i < termsToDel.size(); i++)
{
delTerms[i] = new Term(FIELD_NAME, termsToDel[i]);
}
foreach (Term delTerm in delTerms)
{
writer.DeleteDocuments(delTerm);
}
writer.Commit();
writer.Dispose();
foreach (string termToDel in termsToDel)
{
var toDel = docs[termToDel];
assertTrue(toDel != null);
docs.Remove(termToDel);
}
IndexReader ir = DirectoryReader.Open(dir);
assertEquals(ir.NumDocs, docs.size());
IDictionary dictionary = new DocumentDictionary(ir, FIELD_NAME, WEIGHT_FIELD_NAME);
IInputIterator inputIterator = dictionary.GetEntryIterator();
BytesRef f;
while ((f = inputIterator.Next()) != null)
{
var field = f.Utf8ToString();
Document doc = docs.ContainsKey(field) ? docs[field] : null;
docs.Remove(field);
assertTrue(f.equals(new BytesRef(doc.Get(FIELD_NAME))));
IIndexableField weightField = doc.GetField(WEIGHT_FIELD_NAME);
assertEquals(inputIterator.Weight, (weightField != null) ? Convert.ToInt64(weightField.GetNumericValue()) : 0);
assertEquals(inputIterator.Payload, null);
}
foreach (string invalidTerm in invalidDocTerms)
{
var invalid = docs[invalidTerm];
docs.Remove(invalidTerm);
assertNotNull(invalid);
}
assertTrue(!docs.Any());
ir.Dispose();
dir.Dispose();
}
public virtual void TestPostings()
{
Directory dir = NewFSDirectory(CreateTempDir("postings"));
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetCodec(Codec.ForName("Lucene40"));
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
Document doc = new Document();
// id field
FieldType idType = new FieldType(StringField.TYPE_NOT_STORED);
idType.StoreTermVectors = true;
Field idField = new Field("id", "", idType);
doc.Add(idField);
// title field: short text field
FieldType titleType = new FieldType(TextField.TYPE_NOT_STORED);
titleType.StoreTermVectors = true;
titleType.StoreTermVectorPositions = true;
titleType.StoreTermVectorOffsets = true;
titleType.IndexOptions = IndexOptions();
Field titleField = new Field("title", "", titleType);
doc.Add(titleField);
// body field: long text field
FieldType bodyType = new FieldType(TextField.TYPE_NOT_STORED);
bodyType.StoreTermVectors = true;
bodyType.StoreTermVectorPositions = true;
bodyType.StoreTermVectorOffsets = true;
bodyType.IndexOptions = IndexOptions();
Field bodyField = new Field("body", "", bodyType);
doc.Add(bodyField);
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
titleField.StringValue = FieldValue(1);
bodyField.StringValue = FieldValue(3);
iw.AddDocument(doc);
if (Random().Next(20) == 0)
{
iw.DeleteDocuments(new Term("id", Convert.ToString(i)));
}
}
if (Random().NextBoolean())
{
// delete 1-100% of docs
iw.DeleteDocuments(new Term("title", Terms[Random().Next(Terms.Length)]));
}
iw.Dispose();
dir.Dispose(); // checkindex
}
public virtual void TestTotalBytesSize()
{
Directory d = NewDirectory();
if (d is MockDirectoryWrapper)
{
((MockDirectoryWrapper)d).Throttling = MockDirectoryWrapper.Throttling_e.NEVER;
}
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMaxBufferedDocs(5);
iwc.SetMergeScheduler(new TrackingCMS());
if (TestUtil.GetPostingsFormat("id").Equals("SimpleText"))
{
// no
iwc.SetCodec(TestUtil.AlwaysPostingsFormat(new Lucene41PostingsFormat()));
}
RandomIndexWriter w = new RandomIndexWriter(Random(), d, iwc);
for (int i = 0; i < 1000; i++)
{
Document doc = new Document();
doc.Add(new StringField("id", "" + i, Field.Store.NO));
w.AddDocument(doc);
if (Random().NextBoolean())
{
w.DeleteDocuments(new Term("id", "" + Random().Next(i + 1)));
}
}
Assert.IsTrue(((TrackingCMS)w.w.Config.MergeScheduler).TotMergedBytes != 0);
w.Dispose();
d.Dispose();
}
public void TestBulkMergeWithDeletes()
{
int numDocs = AtLeast(200);
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(NoMergePolicy.COMPOUND_FILES));
for (int i = 0; i < numDocs; ++i)
{
Document doc = new Document();
doc.Add(new StringField("id", Convert.ToString(i), Field.Store.YES));
doc.Add(new StoredField("f", TestUtil.RandomSimpleString(Random())));
w.AddDocument(doc);
}
int deleteCount = TestUtil.NextInt(Random(), 5, numDocs);
for (int i = 0; i < deleteCount; ++i)
{
int id = Random().Next(numDocs);
w.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
w.Commit();
w.Dispose();
w = new RandomIndexWriter(Random(), dir);
w.ForceMerge(TestUtil.NextInt(Random(), 1, 3));
w.Commit();
w.Dispose();
TestUtil.CheckIndex(dir);
dir.Dispose();
}
public virtual void TestEnforceDeletions()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetMergeScheduler(new SerialMergeScheduler()).SetMergePolicy(NewLogMergePolicy(10)));
// asserts below requires no unexpected merges:
// NOTE: cannot use writer.getReader because RIW (on
// flipping a coin) may give us a newly opened reader,
// but we use .reopen on this reader below and expect to
// (must) get an NRT reader:
DirectoryReader reader = DirectoryReader.Open(writer.IndexWriter, true);
// same reason we don't wrap?
IndexSearcher searcher = NewSearcher(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
reader, false);
// add a doc, refresh the reader, and check that it's there
Document doc = new Document();
doc.Add(NewStringField("id", "1", Field.Store.YES));
writer.AddDocument(doc);
reader = RefreshReader(reader);
searcher = NewSearcher(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
reader, false);
TopDocs docs = searcher.Search(new MatchAllDocsQuery(), 1);
Assert.AreEqual(1, docs.TotalHits, "Should find a hit...");
Filter startFilter = new QueryWrapperFilter(new TermQuery(new Term("id", "1")));
CachingWrapperFilter filter = new CachingWrapperFilter(startFilter);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.IsTrue(filter.GetSizeInBytes() > 0);
Assert.AreEqual(1, docs.TotalHits, "[query + filter] Should find a hit...");
Query constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
// make sure we get a cache hit when we reopen reader
// that had no change to deletions
// fake delete (deletes nothing):
writer.DeleteDocuments(new Term("foo", "bar"));
IndexReader oldReader = reader;
reader = RefreshReader(reader);
Assert.IsTrue(reader == oldReader);
int missCount = filter.missCount;
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
// cache hit:
Assert.AreEqual(missCount, filter.missCount);
// now delete the doc, refresh the reader, and see that it's not there
writer.DeleteDocuments(new Term("id", "1"));
// NOTE: important to hold ref here so GC doesn't clear
// the cache entry! Else the assert below may sometimes
// fail:
oldReader = reader;
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
missCount = filter.missCount;
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(0, docs.TotalHits, "[query + filter] Should *not* find a hit...");
// cache hit
Assert.AreEqual(missCount, filter.missCount);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(0, docs.TotalHits, "[just filter] Should *not* find a hit...");
// apply deletes dynamically:
filter = new CachingWrapperFilter(startFilter);
writer.AddDocument(doc);
reader = RefreshReader(reader);
searcher = NewSearcher(reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(1, docs.TotalHits, "[query + filter] Should find a hit...");
missCount = filter.missCount;
Assert.IsTrue(missCount > 0);
constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(1, docs.TotalHits, "[just filter] Should find a hit...");
Assert.AreEqual(missCount, filter.missCount);
writer.AddDocument(doc);
// NOTE: important to hold ref here so GC doesn't clear
// the cache entry! Else the assert below may sometimes
// fail:
oldReader = reader;
reader = RefreshReader(reader);
searcher = NewSearcher(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(2, docs.TotalHits, "[query + filter] Should find 2 hits...");
Assert.IsTrue(filter.missCount > missCount);
missCount = filter.missCount;
constantScore = new ConstantScoreQuery(filter);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(2, docs.TotalHits, "[just filter] Should find a hit...");
Assert.AreEqual(missCount, filter.missCount);
// now delete the doc, refresh the reader, and see that it's not there
writer.DeleteDocuments(new Term("id", "1"));
reader = RefreshReader(reader);
searcher = NewSearcher(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
reader, false);
docs = searcher.Search(new MatchAllDocsQuery(), filter, 1);
Assert.AreEqual(0, docs.TotalHits, "[query + filter] Should *not* find a hit...");
// CWF reused the same entry (it dynamically applied the deletes):
Assert.AreEqual(missCount, filter.missCount);
docs = searcher.Search(constantScore, 1);
Assert.AreEqual(0, docs.TotalHits, "[just filter] Should *not* find a hit...");
// CWF reused the same entry (it dynamically applied the deletes):
Assert.AreEqual(missCount, filter.missCount);
// NOTE: silliness to make sure JRE does not eliminate
// our holding onto oldReader to prevent
// CachingWrapperFilter's WeakHashMap from dropping the
// entry:
Assert.IsTrue(oldReader != null);
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
public virtual void TestRandomStoredFields()
{
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);
int docCount = AtLeast(200);
int fieldCount = TestUtil.NextInt(rand, 1, 5);
IList<int?> fieldIDs = new List<int?>();
FieldType customType = new FieldType(TextField.TYPE_STORED);
customType.Tokenized = false;
Field idField = NewField("id", "", customType);
for (int i = 0; i < fieldCount; i++)
{
fieldIDs.Add(i);
}
IDictionary<string, Document> docs = new Dictionary<string, Document>();
if (VERBOSE)
{
Console.WriteLine("TEST: build index docCount=" + docCount);
}
FieldType customType2 = new FieldType();
customType2.Stored = true;
for (int i = 0; i < docCount; i++)
{
Document doc = new Document();
doc.Add(idField);
string id = "" + i;
idField.StringValue = id;
docs[id] = doc;
if (VERBOSE)
{
Console.WriteLine("TEST: add doc id=" + id);
}
foreach (int field in fieldIDs)
{
string s;
if (rand.Next(4) != 3)
{
s = TestUtil.RandomUnicodeString(rand, 1000);
doc.Add(NewField("f" + field, s, customType2));
}
else
{
s = null;
}
}
w.AddDocument(doc);
if (rand.Next(50) == 17)
{
// mixup binding of field name -> Number every so often
fieldIDs = CollectionsHelper.Shuffle(fieldIDs);
}
if (rand.Next(5) == 3 && i > 0)
{
string delID = "" + rand.Next(i);
if (VERBOSE)
{
Console.WriteLine("TEST: delete doc id=" + delID);
}
w.DeleteDocuments(new Term("id", delID));
docs.Remove(delID);
}
}
if (VERBOSE)
{
Console.WriteLine("TEST: " + docs.Count + " docs in index; now load fields");
}
if (docs.Count > 0)
{
string[] idsList = docs.Keys.ToArray(/*new string[docs.Count]*/);
for (int x = 0; x < 2; x++)
{
IndexReader r = w.Reader;
IndexSearcher s = NewSearcher(r);
if (VERBOSE)
{
Console.WriteLine("TEST: cycle x=" + x + " r=" + r);
}
int num = AtLeast(1000);
for (int iter = 0; iter < num; iter++)
{
string testID = idsList[rand.Next(idsList.Length)];
if (VERBOSE)
{
Console.WriteLine("TEST: test id=" + testID);
}
TopDocs hits = s.Search(new TermQuery(new Term("id", testID)), 1);
Assert.AreEqual(1, hits.TotalHits);
Document doc = r.Document(hits.ScoreDocs[0].Doc);
Document docExp = docs[testID];
for (int i = 0; i < fieldCount; i++)
{
Assert.AreEqual("doc " + testID + ", field f" + fieldCount + " is wrong", docExp.Get("f" + i), doc.Get("f" + i));
}
}
r.Dispose();
w.ForceMerge(1);
}
}
w.Dispose();
dir.Dispose();
}
private void DoTestNumericsVsStoredFields(LongProducer longs)
{
Directory dir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
Document doc = new Document();
Field idField = new StringField("id", "", Field.Store.NO);
Field storedField = NewStringField("stored", "", Field.Store.YES);
Field dvField = new NumericDocValuesField("dv", 0);
doc.Add(idField);
doc.Add(storedField);
doc.Add(dvField);
// index some docs
int numDocs = AtLeast(300);
// numDocs should be always > 256 so that in case of a codec that optimizes
// for numbers of values <= 256, all storage layouts are tested
Debug.Assert(numDocs > 256);
for (int i = 0; i < numDocs; i++)
{
idField.StringValue = Convert.ToString(i);
long value = longs.Next();
storedField.StringValue = Convert.ToString(value);
dvField.LongValue = value;
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
// merge some segments and ensure that at least one of them has more than
// 256 values
writer.ForceMerge(numDocs / 256);
writer.Dispose();
// compare
DirectoryReader ir = DirectoryReader.Open(dir);
foreach (AtomicReaderContext context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
NumericDocValues docValues = r.GetNumericDocValues("dv");
for (int i = 0; i < r.MaxDoc; i++)
{
long storedValue = Convert.ToInt64(r.Document(i).Get("stored"));
Assert.AreEqual(storedValue, docValues.Get(i));
}
}
ir.Dispose();
dir.Dispose();
}
private void DoTestSortedSetVsUninvertedField(int minLength, int maxLength)
{
Directory dir = NewDirectory();
IndexWriterConfig conf = new IndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, conf);
// index some docs
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
Field idField = new StringField("id", Convert.ToString(i), Field.Store.NO);
doc.Add(idField);
int length;
if (minLength == maxLength)
{
length = minLength; // fixed length
}
else
{
length = TestUtil.NextInt(Random(), minLength, maxLength);
}
int numValues = Random().Next(17);
// create a random list of strings
IList<string> values = new List<string>();
for (int v = 0; v < numValues; v++)
{
values.Add(TestUtil.RandomSimpleString(Random(), length));
}
// add in any order to the indexed field
IList<string> unordered = new List<string>(values);
unordered = CollectionsHelper.Shuffle(unordered);
foreach (string v in unordered)
{
doc.Add(NewStringField("indexed", v, Field.Store.NO));
}
// add in any order to the dv field
IList<string> unordered2 = new List<string>(values);
unordered2 = CollectionsHelper.Shuffle(unordered2);
foreach (string v in unordered2)
{
doc.Add(new SortedSetDocValuesField("dv", new BytesRef(v)));
}
writer.AddDocument(doc);
if (Random().Next(31) == 0)
{
writer.Commit();
}
}
// delete some docs
int numDeletions = Random().Next(numDocs / 10);
for (int i = 0; i < numDeletions; i++)
{
int id = Random().Next(numDocs);
writer.DeleteDocuments(new Term("id", Convert.ToString(id)));
}
// compare per-segment
DirectoryReader ir = writer.Reader;
foreach (AtomicReaderContext context in ir.Leaves)
{
AtomicReader r = context.AtomicReader;
SortedSetDocValues expected = FieldCache.DEFAULT.GetDocTermOrds(r, "indexed");
SortedSetDocValues actual = r.GetSortedSetDocValues("dv");
AssertEquals(r.MaxDoc, expected, actual);
}
ir.Dispose();
writer.ForceMerge(1);
// now compare again after the merge
ir = writer.Reader;
AtomicReader ar = GetOnlySegmentReader(ir);
SortedSetDocValues expected_ = FieldCache.DEFAULT.GetDocTermOrds(ar, "indexed");
SortedSetDocValues actual_ = ar.GetSortedSetDocValues("dv");
AssertEquals(ir.MaxDoc, expected_, actual_);
ir.Dispose();
writer.Dispose();
dir.Dispose();
}
