public virtual void TestEmptyDocs()
{
using (Directory dir = NewDirectory())
{
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
using (RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwConf))
{
// make sure that the fact that documents might be empty is not a problem
Document emptyDoc = new Document();
int numDocs = Random.NextBoolean() ? 1 : AtLeast(1000);
for (int i = 0; i < numDocs; ++i)
{
iw.AddDocument(emptyDoc);
}
iw.Commit();
using (DirectoryReader rd = DirectoryReader.Open(dir))
{
for (int i = 0; i < numDocs; ++i)
{
Document doc = rd.Document(i);
Assert.IsNotNull(doc);
Assert.IsTrue(doc.Fields.Count <= 0);
}
} // rd.Dispose();
} // iw.Dispose();
} // dir.Dispose();
}
public virtual void TestReadSkip()
{
using (Directory dir = NewDirectory())
{
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
using (RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwConf))
{
FieldType ft = new FieldType();
ft.IsStored = true;
ft.Freeze();
string @string = TestUtil.RandomSimpleString(Random, 50);
var bytes = @string.GetBytes(Encoding.UTF8);
long l = Random.NextBoolean() ? Random.Next(42) : Random.NextInt64();
int i = Random.NextBoolean() ? Random.Next(42) : Random.Next();
float f = Random.NextSingle();
double d = Random.NextDouble();
IList <Field> fields = Arrays.AsList(new Field("bytes", bytes, ft), new Field("string", @string, ft), new Int64Field("long", l, Field.Store.YES), new Int32Field("int", i, Field.Store.YES), new SingleField("float", f, Field.Store.YES), new DoubleField("double", d, Field.Store.YES)
);
for (int k = 0; k < 100; ++k)
{
Document doc = new Document();
foreach (Field fld in fields)
{
doc.Add(fld);
}
iw.IndexWriter.AddDocument(doc);
}
iw.Commit();
using (DirectoryReader reader = DirectoryReader.Open(dir))
{
int docID = Random.Next(100);
foreach (Field fld in fields)
{
string fldName = fld.Name;
Document sDoc = reader.Document(docID, Collections.Singleton(fldName));
IIndexableField sField = sDoc.GetField(fldName);
if (typeof(Field) == fld.GetType())
{
Assert.AreEqual(fld.GetBinaryValue(), sField.GetBinaryValue());
Assert.AreEqual(fld.GetStringValue(), sField.GetStringValue());
}
else
{
#pragma warning disable 612, 618
Assert.AreEqual(fld.GetNumericValue(), sField.GetNumericValue());
#pragma warning restore 612, 618
}
}
} // reader.Dispose();
} // iw.Dispose();
} // dir.Dispose();
}
internal static byte[] RandomArray(int length, int max)
{
var arr = new byte[length];
for (int i = 0; i < arr.Length; ++i)
{
arr[i] = (byte)RandomInts.RandomInt32Between(Random, 0, max);
}
return(arr);
}
public virtual void TestIncompressible()
{
var decompressed = new byte[RandomInts.RandomInt32Between(Random, 20, 256)];
for (int i = 0; i < decompressed.Length; ++i)
{
decompressed[i] = (byte)i;
}
Test(decompressed);
}
public virtual void TestLongLiterals()
{
// long literals (length >= 16) which are not the last literals
var decompressed = RandomArray(RandomInts.RandomInt32Between(Random, 400, 1024), 256);
int matchRef = Random.Next(30);
int matchOff = RandomInts.RandomInt32Between(Random, decompressed.Length - 40, decompressed.Length - 20);
int matchLength = RandomInts.RandomInt32Between(Random, 4, 10);
Array.Copy(decompressed, matchRef, decompressed, matchOff, matchLength);
Test(decompressed);
}
public virtual void TestLongMatchs()
{
// match length >= 20
var decompressed = new byte[RandomInts.RandomInt32Between(Random, 300, 1024)];
for (int i = 0; i < decompressed.Length; ++i)
{
decompressed[i] = (byte)i;
}
Test(decompressed);
}
public virtual void TestStringUnion()
{
List <BytesRef> strings = new List <BytesRef>();
for (int i = RandomInts.RandomInt32Between(Random, 0, 1000); --i >= 0;)
{
strings.Add(new BytesRef(TestUtil.RandomUnicodeString(Random)));
}
strings.Sort();
Automaton union = BasicAutomata.MakeStringUnion(strings);
Assert.IsTrue(union.IsDeterministic);
Assert.IsTrue(BasicOperations.SameLanguage(union, NaiveUnion(strings)));
}
public virtual void TestDeletePartiallyWrittenFilesIfAbort()
{
Directory dir = NewDirectory();
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
iwConf.SetCodec(CompressingCodec.RandomInstance(Random));
// disable CFS because this test checks file names
iwConf.SetMergePolicy(NewLogMergePolicy(false));
iwConf.SetUseCompoundFile(false);
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwConf);
Document validDoc = new Document();
validDoc.Add(new Int32Field("id", 0, Field.Store.YES));
iw.AddDocument(validDoc);
iw.Commit();
// make sure that #writeField will fail to trigger an abort
Document invalidDoc = new Document();
FieldType fieldType = new FieldType();
fieldType.IsStored = true;
invalidDoc.Add(new FieldAnonymousInnerClassHelper(this, fieldType));
try
{
Assert.Throws <ArgumentException>(() => {
iw.AddDocument(invalidDoc);
iw.Commit();
});
}
finally
{
int counter = 0;
foreach (string fileName in dir.ListAll())
{
if (fileName.EndsWith(".fdt", StringComparison.Ordinal) || fileName.EndsWith(".fdx", StringComparison.Ordinal))
{
counter++;
}
}
// Only one .fdt and one .fdx files must have been found
Assert.AreEqual(2, counter);
iw.Dispose();
dir.Dispose();
}
}
public virtual void TestConcurrentReads()
{
using (Directory dir = NewDirectory())
{
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
using (RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwConf))
{
// make sure the readers are properly cloned
Document doc = new Document();
Field field = new StringField("fld", "", Field.Store.YES);
doc.Add(field);
int numDocs = AtLeast(1000);
for (int i = 0; i < numDocs; ++i)
{
field.SetStringValue("" + i);
iw.AddDocument(doc);
}
iw.Commit();
AtomicObject <Exception> ex = new AtomicObject <Exception>();
using (DirectoryReader rd = DirectoryReader.Open(dir))
{
IndexSearcher searcher = new IndexSearcher(rd);
int concurrentReads = AtLeast(5);
int readsPerThread = AtLeast(50);
IList <ThreadClass> readThreads = new List <ThreadClass>();
for (int i = 0; i < concurrentReads; ++i)
{
readThreads.Add(new ThreadAnonymousInnerClassHelper(numDocs, rd, searcher, readsPerThread, ex, i));
}
foreach (ThreadClass thread in readThreads)
{
thread.Start();
}
foreach (ThreadClass thread in readThreads)
{
thread.Join();
}
} // rd.Dispose();
if (ex.Value != null)
{
throw ex.Value;
}
} // iw.Dispose();
} // dir.Dispose();
}
/// <summary>
/// Creates a random <see cref="CompressingCodec"/> that is using a segment suffix.
/// </summary>
public static CompressingCodec RandomInstance(Random random, bool withSegmentSuffix)
{
return(RandomInstance(random, RandomInts.RandomInt32Between(random, 1, 500), withSegmentSuffix));
}
/// <summary>
/// Creates a random <see cref="CompressingCodec"/> that is using an empty segment
/// suffix.
/// </summary>
public static CompressingCodec RandomInstance(Random random)
{
return(RandomInstance(random, RandomInts.RandomInt32Between(random, 1, 500), false));
}
public virtual void Test([ValueSource(typeof(ConcurrentMergeSchedulerFactories), "Values")] Func <IConcurrentMergeScheduler> newScheduler)
{
MockDirectoryWrapper dir = new MockDirectoryWrapper(Random, new MMapDirectory(CreateTempDir("4GBStoredFields")));
dir.Throttling = Throttling.NEVER;
var config = new IndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random))
.SetMaxBufferedDocs(IndexWriterConfig.DISABLE_AUTO_FLUSH)
.SetRAMBufferSizeMB(256.0)
.SetMergeScheduler(newScheduler())
.SetMergePolicy(NewLogMergePolicy(false, 10))
.SetOpenMode(OpenMode.CREATE);
IndexWriter w = new IndexWriter(dir, config);
MergePolicy mp = w.Config.MergePolicy;
if (mp is LogByteSizeMergePolicy)
{
// 1 petabyte:
((LogByteSizeMergePolicy)mp).MaxMergeMB = 1024 * 1024 * 1024;
}
Document doc = new Document();
FieldType ft = new FieldType();
ft.IsIndexed = false;
ft.IsStored = true;
ft.Freeze();
int valueLength = RandomInts.RandomInt32Between(Random, 1 << 13, 1 << 20);
var value = new byte[valueLength];
for (int i = 0; i < valueLength; ++i)
{
// random so that even compressing codecs can't compress it
value[i] = (byte)Random.Next(256);
}
Field f = new Field("fld", value, ft);
doc.Add(f);
int numDocs = (int)((1L << 32) / valueLength + 100);
for (int i = 0; i < numDocs; ++i)
{
w.AddDocument(doc);
if (VERBOSE && i % (numDocs / 10) == 0)
{
Console.WriteLine(i + " of " + numDocs + "...");
}
}
w.ForceMerge(1);
w.Dispose();
if (VERBOSE)
{
bool found = false;
foreach (string file in dir.ListAll())
{
if (file.EndsWith(".fdt", StringComparison.Ordinal))
{
long fileLength = dir.FileLength(file);
if (fileLength >= 1L << 32)
{
found = true;
}
Console.WriteLine("File length of " + file + " : " + fileLength);
}
}
if (!found)
{
Console.WriteLine("No .fdt file larger than 4GB, test bug?");
}
}
DirectoryReader rd = DirectoryReader.Open(dir);
Document sd = rd.Document(numDocs - 1);
Assert.IsNotNull(sd);
Assert.AreEqual(1, sd.Fields.Count);
BytesRef valueRef = sd.GetBinaryValue("fld");
Assert.IsNotNull(valueRef);
Assert.AreEqual(new BytesRef(value), valueRef);
rd.Dispose();
dir.Dispose();
}
public virtual void TestEncodeDecode()
{
int iterations = RandomInts.RandomInt32Between(Random, 1, 1000);
float AcceptableOverheadRatio = (float)Random.NextDouble();
int[] values = new int[(iterations - 1) * Lucene41PostingsFormat.BLOCK_SIZE + ForUtil.MAX_DATA_SIZE];
for (int i = 0; i < iterations; ++i)
{
int bpv = Random.Next(32);
if (bpv == 0)
{
int value = RandomInts.RandomInt32Between(Random, 0, int.MaxValue);
for (int j = 0; j < Lucene41PostingsFormat.BLOCK_SIZE; ++j)
{
values[i * Lucene41PostingsFormat.BLOCK_SIZE + j] = value;
}
}
else
{
for (int j = 0; j < Lucene41PostingsFormat.BLOCK_SIZE; ++j)
{
values[i * Lucene41PostingsFormat.BLOCK_SIZE + j] = RandomInts.RandomInt32Between(Random, 0, (int)PackedInt32s.MaxValue(bpv));
}
}
}
Directory d = new RAMDirectory();
long endPointer;
{
// encode
IndexOutput @out = d.CreateOutput("test.bin", IOContext.DEFAULT);
ForUtil forUtil = new ForUtil(AcceptableOverheadRatio, @out);
for (int i = 0; i < iterations; ++i)
{
forUtil.WriteBlock(Arrays.CopyOfRange(values, i * Lucene41PostingsFormat.BLOCK_SIZE, values.Length), new byte[Lucene41.ForUtil.MAX_ENCODED_SIZE], @out);
}
endPointer = @out.GetFilePointer();
@out.Dispose();
}
{
// decode
IndexInput @in = d.OpenInput("test.bin", IOContext.READ_ONCE);
ForUtil forUtil = new ForUtil(@in);
for (int i = 0; i < iterations; ++i)
{
if (Random.NextBoolean())
{
forUtil.SkipBlock(@in);
continue;
}
int[] restored = new int[Lucene41.ForUtil.MAX_DATA_SIZE];
forUtil.ReadBlock(@in, new byte[Lucene41.ForUtil.MAX_ENCODED_SIZE], restored);
Assert.AreEqual(Arrays.CopyOfRange(values, i * Lucene41PostingsFormat.BLOCK_SIZE, (i + 1) * Lucene41PostingsFormat.BLOCK_SIZE), Arrays.CopyOf(restored, Lucene41PostingsFormat.BLOCK_SIZE));
}
assertEquals(endPointer, @in.GetFilePointer());
@in.Dispose();
}
}
private void DoTest(DocValuesType type)
{
Directory d = NewDirectory();
IndexWriterConfig iwConfig = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
int nDocs = AtLeast(50);
Field id = new NumericDocValuesField("id", 0);
Field f;
switch (type)
{
case DocValuesType.BINARY:
f = new BinaryDocValuesField("dv", new BytesRef());
break;
case DocValuesType.SORTED:
f = new SortedDocValuesField("dv", new BytesRef());
break;
case DocValuesType.NUMERIC:
f = new NumericDocValuesField("dv", 0);
break;
default:
throw new InvalidOperationException();
}
Document document = new Document();
document.Add(id);
document.Add(f);
object[] vals = new object[nDocs];
RandomIndexWriter iw = new RandomIndexWriter(Random, d, iwConfig);
for (int i = 0; i < nDocs; ++i)
{
id.SetInt64Value(i);
switch (type)
{
case DocValuesType.SORTED:
case DocValuesType.BINARY:
do
{
vals[i] = TestUtil.RandomSimpleString(Random, 20);
} while (((string)vals[i]).Length == 0);
f.SetBytesValue(new BytesRef((string)vals[i]));
break;
case DocValuesType.NUMERIC:
int bitsPerValue = RandomInts.RandomInt32Between(Random, 1, 31); // keep it an int
vals[i] = (long)Random.Next((int)PackedInt32s.MaxValue(bitsPerValue));
f.SetInt64Value((long)vals[i]);
break;
}
iw.AddDocument(document);
if (Random.NextBoolean() && i % 10 == 9)
{
iw.Commit();
}
}
iw.Dispose();
DirectoryReader rd = DirectoryReader.Open(d);
foreach (AtomicReaderContext leave in rd.Leaves)
{
FunctionValues ids = (new Int64FieldSource("id")).GetValues(null, leave);
ValueSource vs;
switch (type)
{
case DocValuesType.BINARY:
case DocValuesType.SORTED:
vs = new BytesRefFieldSource("dv");
break;
case DocValuesType.NUMERIC:
vs = new Int64FieldSource("dv");
break;
default:
throw new InvalidOperationException();
}
FunctionValues values = vs.GetValues(null, leave);
BytesRef bytes = new BytesRef();
for (int i = 0; i < leave.AtomicReader.MaxDoc; ++i)
{
assertTrue(values.Exists(i));
if (vs is BytesRefFieldSource)
{
assertTrue(values.ObjectVal(i) is string);
}
else if (vs is Int64FieldSource)
{
assertTrue(values.ObjectVal(i) is long?);
assertTrue(values.BytesVal(i, bytes));
}
else
{
throw new InvalidOperationException();
}
object expected = vals[ids.Int32Val(i)];
switch (type)
{
case DocValuesType.SORTED:
values.OrdVal(i); // no exception
assertTrue(values.NumOrd >= 1);
goto case DocValuesType.BINARY;
case DocValuesType.BINARY:
assertEquals(expected, values.ObjectVal(i));
assertEquals(expected, values.StrVal(i));
assertEquals(expected, values.ObjectVal(i));
assertEquals(expected, values.StrVal(i));
assertTrue(values.BytesVal(i, bytes));
assertEquals(new BytesRef((string)expected), bytes);
break;
case DocValuesType.NUMERIC:
assertEquals(Convert.ToInt64(expected, CultureInfo.InvariantCulture), values.Int64Val(i));
break;
}
}
}
rd.Dispose();
d.Dispose();
}
public virtual void TestBigDocuments()
{
// "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)
using (Directory dir = new MockDirectoryWrapper(Random, new MMapDirectory(CreateTempDir("testBigDocuments"))))
{
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
using (RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwConf))
{
if (dir is MockDirectoryWrapper)
{
((MockDirectoryWrapper)dir).Throttling = Throttling.NEVER;
}
Document emptyDoc = new Document(); // emptyDoc
Document bigDoc1 = new Document(); // lot of small fields
Document bigDoc2 = new Document(); // 1 very big field
Field idField = new StringField("id", "", Field.Store.NO);
emptyDoc.Add(idField);
bigDoc1.Add(idField);
bigDoc2.Add(idField);
FieldType onlyStored = new FieldType(StringField.TYPE_STORED);
onlyStored.IsIndexed = false;
Field smallField = new Field("fld", RandomByteArray(Random.Next(10), 256), onlyStored);
int numFields = RandomInts.RandomInt32Between(Random, 500000, 1000000);
for (int i = 0; i < numFields; ++i)
{
bigDoc1.Add(smallField);
}
Field bigField = new Field("fld", RandomByteArray(RandomInts.RandomInt32Between(Random, 1000000, 5000000), 2), onlyStored);
bigDoc2.Add(bigField);
int numDocs = AtLeast(5);
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.Next(numDocs) == 0)
{
iw.Commit();
}
}
iw.Commit();
iw.ForceMerge(1); // look at what happens when big docs are merged
using (DirectoryReader rd = DirectoryReader.Open(dir))
{
IndexSearcher searcher = new IndexSearcher(rd);
for (int i = 0; i < numDocs; ++i)
{
Query query = new TermQuery(new Term("id", "" + i));
TopDocs topDocs = searcher.Search(query, 1);
Assert.AreEqual(1, topDocs.TotalHits, "" + i);
Document doc = rd.Document(topDocs.ScoreDocs[0].Doc);
Assert.IsNotNull(doc);
IIndexableField[] fieldValues = doc.GetFields("fld");
Assert.AreEqual(docs[i].GetFields("fld").Length, fieldValues.Length);
if (fieldValues.Length > 0)
{
Assert.AreEqual(docs[i].GetFields("fld")[0].GetBinaryValue(), fieldValues[0].GetBinaryValue());
}
}
} // rd.Dispose();
} // iw.Dispose();
} // dir.Dispose();
}
public virtual void TestWriteReadMerge()
{
// get another codec, other than the default: so we are merging segments across different codecs
Codec otherCodec;
if ("SimpleText".Equals(Codec.Default.Name, StringComparison.Ordinal))
{
otherCodec = new Lucene46Codec();
}
else
{
otherCodec = new SimpleTextCodec();
}
using (Directory dir = NewDirectory())
{
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwConf.SetMaxBufferedDocs(RandomInts.RandomInt32Between(Random, 2, 30));
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, (IndexWriterConfig)iwConf.Clone());
try
{
int docCount = AtLeast(200);
var data = new byte[docCount][][];
for (int i = 0; i < docCount; ++i)
{
int fieldCount = Rarely() ? RandomInts.RandomInt32Between(Random, 1, 500) : RandomInts.RandomInt32Between(Random, 1, 5);
data[i] = new byte[fieldCount][];
for (int j = 0; j < fieldCount; ++j)
{
int length = Rarely() ? Random.Next(1000) : Random.Next(10);
int max = Rarely() ? 256 : 2;
data[i][j] = RandomByteArray(length, max);
}
}
FieldType type = new FieldType(StringField.TYPE_STORED);
type.IsIndexed = false;
type.Freeze();
Int32Field id = new Int32Field("id", 0, Field.Store.YES);
for (int i = 0; i < data.Length; ++i)
{
Document doc = new Document();
doc.Add(id);
id.SetInt32Value(i);
for (int j = 0; j < data[i].Length; ++j)
{
Field f = new Field("bytes" + j, data[i][j], type);
doc.Add(f);
}
iw.IndexWriter.AddDocument(doc);
if (Random.NextBoolean() && (i % (data.Length / 10) == 0))
{
iw.IndexWriter.Dispose();
// test merging against a non-compressing codec
if (iwConf.Codec == otherCodec)
{
iwConf.SetCodec(Codec.Default);
}
else
{
iwConf.SetCodec(otherCodec);
}
iw = new RandomIndexWriter(Random, dir, (IndexWriterConfig)iwConf.Clone());
}
}
for (int i = 0; i < 10; ++i)
{
int min = Random.Next(data.Length);
int max = min + Random.Next(20);
iw.DeleteDocuments(NumericRangeQuery.NewInt32Range("id", min, max, true, false));
}
iw.ForceMerge(2); // force merges with deletions
iw.Commit();
using (DirectoryReader ir = DirectoryReader.Open(dir))
{
Assert.IsTrue(ir.NumDocs > 0);
int numDocs = 0;
for (int i = 0; i < ir.MaxDoc; ++i)
{
Document doc = ir.Document(i);
if (doc == null)
{
continue;
}
++numDocs;
int docId = (int)doc.GetField("id").GetInt32Value();
Assert.AreEqual(data[docId].Length + 1, doc.Fields.Count);
for (int j = 0; j < data[docId].Length; ++j)
{
var arr = data[docId][j];
BytesRef arr2Ref = doc.GetBinaryValue("bytes" + j);
var arr2 = Arrays.CopyOfRange(arr2Ref.Bytes, arr2Ref.Offset, arr2Ref.Offset + arr2Ref.Length);
Assert.AreEqual(arr, arr2);
}
}
Assert.IsTrue(ir.NumDocs <= numDocs);
} // ir.Dispose();
iw.DeleteAll();
iw.Commit();
iw.ForceMerge(1);
}
finally
{
iw.Dispose();
}
} // dir.Dispose();
}