public virtual void TestIntFieldCache()
{
Directory dir = NewDirectory();
IndexWriterConfig cfg = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
cfg.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, cfg);
Document doc = new Document();
Int32Field field = new Int32Field("f", 0, Field.Store.YES);
doc.Add(field);
int[] values = new int[TestUtil.NextInt32(Random, 1, 10)];
for (int i = 0; i < values.Length; ++i)
{
int v;
switch (Random.Next(10))
{
case 0:
v = int.MinValue;
break;
case 1:
v = 0;
break;
case 2:
v = int.MaxValue;
break;
default:
v = TestUtil.NextInt32(Random, -10, 10);
break;
}
values[i] = v;
if (v == 0 && Random.NextBoolean())
{
// missing
iw.AddDocument(new Document());
}
else
{
field.SetInt32Value(v);
iw.AddDocument(doc);
}
}
iw.ForceMerge(1);
DirectoryReader reader = iw.GetReader();
Int32s ints = FieldCache.DEFAULT.GetInt32s(GetOnlySegmentReader(reader), "f", false);
for (int i = 0; i < values.Length; ++i)
{
Assert.AreEqual(values[i], ints.Get(i));
}
reader.Dispose();
iw.Dispose();
dir.Dispose();
}
public override void BeforeClass()
{
base.BeforeClass();
NoDocs = AtLeast(4096);
Distance = (1 << 30) / NoDocs;
Directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), Directory, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMaxBufferedDocs(TestUtil.NextInt(Random(), 100, 1000)).SetMergePolicy(NewLogMergePolicy()));
FieldType storedInt = new FieldType(Int32Field.TYPE_NOT_STORED);
storedInt.IsStored = true;
storedInt.Freeze();
FieldType storedInt8 = new FieldType(storedInt);
storedInt8.NumericPrecisionStep = 8;
FieldType storedInt4 = new FieldType(storedInt);
storedInt4.NumericPrecisionStep = 4;
FieldType storedInt2 = new FieldType(storedInt);
storedInt2.NumericPrecisionStep = 2;
FieldType storedIntNone = new FieldType(storedInt);
storedIntNone.NumericPrecisionStep = int.MaxValue;
FieldType unstoredInt = Int32Field.TYPE_NOT_STORED;
FieldType unstoredInt8 = new FieldType(unstoredInt);
unstoredInt8.NumericPrecisionStep = 8;
FieldType unstoredInt4 = new FieldType(unstoredInt);
unstoredInt4.NumericPrecisionStep = 4;
FieldType unstoredInt2 = new FieldType(unstoredInt);
unstoredInt2.NumericPrecisionStep = 2;
Int32Field field8 = new Int32Field("field8", 0, storedInt8), field4 = new Int32Field("field4", 0, storedInt4), field2 = new Int32Field("field2", 0, storedInt2), fieldNoTrie = new Int32Field("field" + int.MaxValue, 0, storedIntNone), ascfield8 = new Int32Field("ascfield8", 0, unstoredInt8), ascfield4 = new Int32Field("ascfield4", 0, unstoredInt4), ascfield2 = new Int32Field("ascfield2", 0, unstoredInt2);
Document doc = new Document();
// add fields, that have a distance to test general functionality
doc.Add(field8);
doc.Add(field4);
doc.Add(field2);
doc.Add(fieldNoTrie);
// add ascending fields with a distance of 1, beginning at -noDocs/2 to test the correct splitting of range and inclusive/exclusive
doc.Add(ascfield8);
doc.Add(ascfield4);
doc.Add(ascfield2);
// Add a series of noDocs docs with increasing int values
for (int l = 0; l < NoDocs; l++)
{
int val = Distance * l + StartOffset;
field8.SetInt32Value(val);
field4.SetInt32Value(val);
field2.SetInt32Value(val);
fieldNoTrie.SetInt32Value(val);
val = l - (NoDocs / 2);
ascfield8.SetInt32Value(val);
ascfield4.SetInt32Value(val);
ascfield2.SetInt32Value(val);
writer.AddDocument(doc);
}
Reader = writer.Reader;
Searcher = NewSearcher(Reader);
writer.Dispose();
}
public void TestRandom()
{
int numberOfRuns = TestUtil.NextInt32(Random, 3, 6);
for (int iter = 0; iter < numberOfRuns; iter++)
{
if (VERBOSE)
{
Console.WriteLine(string.Format("TEST: iter={0} total={1}", iter, numberOfRuns));
}
int numDocs = TestUtil.NextInt32(Random, 100, 1000) * RANDOM_MULTIPLIER;
int numGroups = TestUtil.NextInt32(Random, 1, numDocs);
if (VERBOSE)
{
Console.WriteLine("TEST: numDocs=" + numDocs + " numGroups=" + numGroups);
}
List <BytesRef> groups = new List <BytesRef>();
for (int i = 0; i < numGroups; i++)
{
string randomValue;
do
{
// B/c of DV based impl we can't see the difference between an empty string and a null value.
// For that reason we don't generate empty string groups.
randomValue = TestUtil.RandomRealisticUnicodeString(Random);
} while ("".Equals(randomValue, StringComparison.Ordinal));
groups.Add(new BytesRef(randomValue));
}
string[] contentStrings = new string[TestUtil.NextInt32(Random, 2, 20)];
if (VERBOSE)
{
Console.WriteLine("TEST: create fake content");
}
for (int contentIDX = 0; contentIDX < contentStrings.Length; contentIDX++)
{
StringBuilder sb = new StringBuilder();
sb.append("real").append(Random.nextInt(3)).append(' ');
int fakeCount = Random.nextInt(10);
for (int fakeIDX = 0; fakeIDX < fakeCount; fakeIDX++)
{
sb.append("fake ");
}
contentStrings[contentIDX] = sb.toString();
if (VERBOSE)
{
Console.WriteLine(" content=" + sb.toString());
}
}
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
Random,
dir,
NewIndexWriterConfig(TEST_VERSION_CURRENT,
new MockAnalyzer(Random)));
bool preFlex = "Lucene3x".Equals(w.IndexWriter.Config.Codec.Name, StringComparison.Ordinal);
bool canUseIDV = !preFlex;
DocValuesType valueType = vts[Random.nextInt(vts.Length)];
Document doc = new Document();
Document docNoGroup = new Document();
Field group = NewStringField("group", "", Field.Store.NO);
doc.Add(group);
Field valuesField = null;
if (canUseIDV)
{
switch (valueType)
{
case DocValuesType.BINARY:
valuesField = new BinaryDocValuesField("group_dv", new BytesRef());
break;
case DocValuesType.SORTED:
valuesField = new SortedDocValuesField("group_dv", new BytesRef());
break;
default:
fail("unhandled type");
break;
}
doc.Add(valuesField);
}
Field sort1 = NewStringField("sort1", "", Field.Store.NO);
doc.Add(sort1);
docNoGroup.Add(sort1);
Field sort2 = NewStringField("sort2", "", Field.Store.NO);
doc.Add(sort2);
docNoGroup.Add(sort2);
Field sort3 = NewStringField("sort3", "", Field.Store.NO);
doc.Add(sort3);
docNoGroup.Add(sort3);
Field content = NewTextField("content", "", Field.Store.NO);
doc.Add(content);
docNoGroup.Add(content);
Int32Field id = new Int32Field("id", 0, Field.Store.NO);
doc.Add(id);
docNoGroup.Add(id);
GroupDoc[] groupDocs = new GroupDoc[numDocs];
for (int i = 0; i < numDocs; i++)
{
BytesRef groupValue;
if (Random.nextInt(24) == 17)
{
// So we test the "doc doesn't have the group'd
// field" case:
groupValue = null;
}
else
{
groupValue = groups[Random.nextInt(groups.size())];
}
GroupDoc groupDoc = new GroupDoc(
i,
groupValue,
groups[Random.nextInt(groups.size())],
groups[Random.nextInt(groups.size())],
new BytesRef(string.Format(CultureInfo.InvariantCulture, "{0:D5}", i)),
contentStrings[Random.nextInt(contentStrings.Length)]
);
if (VERBOSE)
{
Console.WriteLine(" doc content=" + groupDoc.content + " id=" + i + " group=" + (groupDoc.group == null ? "null" : groupDoc.group.Utf8ToString()) + " sort1=" + groupDoc.sort1.Utf8ToString() + " sort2=" + groupDoc.sort2.Utf8ToString() + " sort3=" + groupDoc.sort3.Utf8ToString());
}
groupDocs[i] = groupDoc;
if (groupDoc.group != null)
{
group.SetStringValue(groupDoc.group.Utf8ToString());
if (canUseIDV)
{
valuesField.SetBytesValue(new BytesRef(groupDoc.group.Utf8ToString()));
}
}
sort1.SetStringValue(groupDoc.sort1.Utf8ToString());
sort2.SetStringValue(groupDoc.sort2.Utf8ToString());
sort3.SetStringValue(groupDoc.sort3.Utf8ToString());
content.SetStringValue(groupDoc.content);
id.SetInt32Value(groupDoc.id);
if (groupDoc.group == null)
{
w.AddDocument(docNoGroup);
}
else
{
w.AddDocument(doc);
}
}
DirectoryReader r = w.GetReader();
w.Dispose();
// NOTE: intentional but temporary field cache insanity!
FieldCache.Int32s docIdToFieldId = FieldCache.DEFAULT.GetInt32s(SlowCompositeReaderWrapper.Wrap(r), "id", false);
int[] fieldIdToDocID = new int[numDocs];
for (int i = 0; i < numDocs; i++)
{
int fieldId = docIdToFieldId.Get(i);
fieldIdToDocID[fieldId] = i;
}
try
{
IndexSearcher s = NewSearcher(r);
if (typeof(SlowCompositeReaderWrapper).GetTypeInfo().IsAssignableFrom(s.IndexReader.GetType()))
{
canUseIDV = false;
}
else
{
canUseIDV = !preFlex;
}
for (int contentID = 0; contentID < 3; contentID++)
{
ScoreDoc[] hits = s.Search(new TermQuery(new Term("content", "real" + contentID)), numDocs).ScoreDocs;
foreach (ScoreDoc hit in hits)
{
GroupDoc gd = groupDocs[docIdToFieldId.Get(hit.Doc)];
assertTrue(gd.score == 0.0);
gd.score = hit.Score;
int docId = gd.id;
assertEquals(docId, docIdToFieldId.Get(hit.Doc));
}
}
foreach (GroupDoc gd in groupDocs)
{
assertTrue(gd.score != 0.0);
}
for (int searchIter = 0; searchIter < 100; searchIter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: searchIter=" + searchIter);
}
string searchTerm = "real" + Random.nextInt(3);
bool sortByScoreOnly = Random.nextBoolean();
Sort sortWithinGroup = GetRandomSort(sortByScoreOnly);
AbstractAllGroupHeadsCollector allGroupHeadsCollector = CreateRandomCollector("group", sortWithinGroup, canUseIDV, valueType);
s.Search(new TermQuery(new Term("content", searchTerm)), allGroupHeadsCollector);
int[] expectedGroupHeads = CreateExpectedGroupHeads(searchTerm, groupDocs, sortWithinGroup, sortByScoreOnly, fieldIdToDocID);
int[] actualGroupHeads = allGroupHeadsCollector.RetrieveGroupHeads();
// The actual group heads contains Lucene ids. Need to change them into our id value.
for (int i = 0; i < actualGroupHeads.Length; i++)
{
actualGroupHeads[i] = docIdToFieldId.Get(actualGroupHeads[i]);
}
// Allows us the easily iterate and assert the actual and expected results.
Array.Sort(expectedGroupHeads);
Array.Sort(actualGroupHeads);
if (VERBOSE)
{
Console.WriteLine("Collector: " + allGroupHeadsCollector.GetType().Name);
Console.WriteLine("Sort within group: " + sortWithinGroup);
Console.WriteLine("Num group: " + numGroups);
Console.WriteLine("Num doc: " + numDocs);
Console.WriteLine("\n=== Expected: \n");
foreach (int expectedDocId in expectedGroupHeads)
{
GroupDoc expectedGroupDoc = groupDocs[expectedDocId];
string expectedGroup = expectedGroupDoc.group == null ? null : expectedGroupDoc.group.Utf8ToString();
Console.WriteLine(
string.Format(CultureInfo.InvariantCulture,
"Group:{0,10} score{1:0.0#######,5} Sort1:{2,10} Sort2:{3,10} Sort3:{4,10} doc:{5,10}",
expectedGroup, expectedGroupDoc.score, expectedGroupDoc.sort1.Utf8ToString(),
expectedGroupDoc.sort2.Utf8ToString(), expectedGroupDoc.sort3.Utf8ToString(), expectedDocId)
);
}
Console.WriteLine("\n=== Actual: \n");
foreach (int actualDocId in actualGroupHeads)
{
GroupDoc actualGroupDoc = groupDocs[actualDocId];
string actualGroup = actualGroupDoc.group == null ? null : actualGroupDoc.group.Utf8ToString();
Console.WriteLine(
string.Format(CultureInfo.InvariantCulture,
"Group:{0,10} score{1:0.0#######,5} Sort1:{2,10} Sort2:{3,10} Sort3:{4,10} doc:{5,10}",
actualGroup, actualGroupDoc.score, actualGroupDoc.sort1.Utf8ToString(),
actualGroupDoc.sort2.Utf8ToString(), actualGroupDoc.sort3.Utf8ToString(), actualDocId)
);
}
Console.WriteLine("\n===================================================================================");
}
assertArrayEquals(expectedGroupHeads, actualGroupHeads);
}
}
finally
{
QueryUtils.PurgeFieldCache(r);
}
r.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 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();
}
Directory dir = NewDirectory();
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwConf.SetMaxBufferedDocs(RandomInts.NextIntBetween(Random(), 2, 30));
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, (IndexWriterConfig)iwConf.Clone());
int docCount = AtLeast(200);
var data = new byte[docCount][][];
for (int i = 0; i < docCount; ++i)
{
int fieldCount = Rarely() ? RandomInts.NextIntBetween(Random(), 1, 500) : RandomInts.NextIntBetween(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.w.AddDocument(doc);
if (Random().NextBoolean() && (i % (data.Length / 10) == 0))
{
iw.w.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();
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").GetNumericValue();
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);
iw.Dispose();
dir.Dispose();
}
