public void BeforeClass()
{
Dir = NewDirectory();
Sdir1 = NewDirectory();
Sdir2 = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), Dir, new MockAnalyzer(Random()), Similarity, TimeZone);
RandomIndexWriter swriter1 = new RandomIndexWriter(Random(), Sdir1, new MockAnalyzer(Random()), Similarity, TimeZone);
RandomIndexWriter swriter2 = new RandomIndexWriter(Random(), Sdir2, new MockAnalyzer(Random()), Similarity, TimeZone);
for (int i = 0; i < 10; i++)
{
Document doc = new Document();
doc.Add(NewStringField("data", Convert.ToString(i), Field.Store.NO));
writer.AddDocument(doc);
((i % 2 == 0) ? swriter1 : swriter2).AddDocument(doc);
}
writer.ForceMerge(1);
swriter1.ForceMerge(1);
swriter2.ForceMerge(1);
writer.Dispose();
swriter1.Dispose();
swriter2.Dispose();
Reader = DirectoryReader.Open(Dir);
Searcher = NewSearcher(Reader);
MultiReader = new MultiReader(new IndexReader[] { DirectoryReader.Open(Sdir1), DirectoryReader.Open(Sdir2) }, true);
MultiSearcher = NewSearcher(MultiReader);
MultiReaderDupls = new MultiReader(new IndexReader[] { DirectoryReader.Open(Sdir1), DirectoryReader.Open(Dir) }, true);
MultiSearcherDupls = NewSearcher(MultiReaderDupls);
}
public override void SetUp()
{
base.SetUp();
dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer
(Random()));
iwc.SetMergePolicy(NewLogMergePolicy());
var iw = new RandomIndexWriter(Random(), dir, iwc);
var doc = new Document
{
NewStringField("id", "1", Field.Store.YES),
NewTextField("body", "some contents and more contents", Field.Store.NO),
new NumericDocValuesField("popularity", 5)
};
iw.AddDocument(doc);
doc = new Document
{
NewStringField("id", "2", Field.Store.YES),
NewTextField("body", "another document with different contents", Field.Store
.NO),
new NumericDocValuesField("popularity", 20)
};
iw.AddDocument(doc);
doc = new Document
{
NewStringField("id", "3", Field.Store.YES),
NewTextField("body", "crappy contents", Field.Store.NO),
new NumericDocValuesField("popularity", 2)
};
iw.AddDocument(doc);
iw.ForceMerge(1);
reader = iw.Reader;
iw.Dispose();
}
public void TestFieldNotPresent()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
int num = AtLeast(3);
int skip = Random().Next(num);
var terms = new List<Term>();
for (int i = 0; i < num; i++)
{
terms.Add(new Term("field" + i, "content1"));
Document doc = new Document();
if (skip == i)
{
continue;
}
doc.Add(NewStringField("field" + i, "content1", Field.Store.YES));
w.AddDocument(doc);
}
w.ForceMerge(1);
IndexReader reader = w.Reader;
w.Dispose();
assertEquals(1, reader.Leaves.size());
AtomicReaderContext context = reader.Leaves.First();
TermsFilter tf = new TermsFilter(terms);
FixedBitSet bits = (FixedBitSet)tf.GetDocIdSet(context, context.AtomicReader.LiveDocs);
assertEquals("Must be num fields - 1 since we skip only one field", num - 1, bits.Cardinality());
reader.Dispose();
dir.Dispose();
}
public override void BeforeClass()
{
base.BeforeClass();
dir = NewDirectory();
sdir1 = NewDirectory();
sdir2 = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir, new MockAnalyzer(Random));
RandomIndexWriter swriter1 = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, sdir1, new MockAnalyzer(Random));
RandomIndexWriter swriter2 = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, sdir2, new MockAnalyzer(Random));
for (int i = 0; i < 10; i++)
{
Document doc = new Document();
doc.Add(NewStringField("data", Convert.ToString(i), Field.Store.NO));
writer.AddDocument(doc);
((i % 2 == 0) ? swriter1 : swriter2).AddDocument(doc);
}
writer.ForceMerge(1);
swriter1.ForceMerge(1);
swriter2.ForceMerge(1);
writer.Dispose();
swriter1.Dispose();
swriter2.Dispose();
reader = DirectoryReader.Open(dir);
searcher = NewSearcher(reader);
multiReader = new MultiReader(new IndexReader[] { DirectoryReader.Open(sdir1), DirectoryReader.Open(sdir2) }, true);
multiSearcher = NewSearcher(multiReader);
multiReaderDupls = new MultiReader(new IndexReader[] { DirectoryReader.Open(sdir1), DirectoryReader.Open(dir) }, true);
multiSearcherDupls = NewSearcher(multiReaderDupls);
}
public virtual void TestBooleanScorerMax()
{
Directory dir = NewDirectory();
RandomIndexWriter riw = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
int docCount = AtLeast(10000);
for (int i = 0; i < docCount; i++)
{
Document doc = new Document();
doc.Add(NewField("field", "a", TextField.TYPE_NOT_STORED));
riw.AddDocument(doc);
}
riw.ForceMerge(1);
IndexReader r = riw.Reader;
riw.Dispose();
IndexSearcher s = NewSearcher(r);
BooleanQuery bq = new BooleanQuery();
bq.Add(new TermQuery(new Term("field", "a")), BooleanClause.Occur.SHOULD);
bq.Add(new TermQuery(new Term("field", "a")), BooleanClause.Occur.SHOULD);
Weight w = s.CreateNormalizedWeight(bq);
Assert.AreEqual(1, s.IndexReader.Leaves.Count);
BulkScorer scorer = w.BulkScorer(s.IndexReader.Leaves[0], false, null);
FixedBitSet hits = new FixedBitSet(docCount);
AtomicInteger end = new AtomicInteger();
Collector c = new CollectorAnonymousInnerClassHelper(this, scorer, hits, end);
while (end.Get() < docCount)
{
int inc = TestUtil.NextInt(Random(), 1, 1000);
end.AddAndGet(inc);
scorer.Score(c, end.Get());
}
Assert.AreEqual(docCount, hits.Cardinality());
r.Dispose();
dir.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();
}
public void TestSkipField()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random, dir, Similarity, TimeZone);
int num = AtLeast(10);
var terms = new HashSet <Term>();
for (int i = 0; i < num; i++)
{
string field = "field" + Random.Next(100);
terms.Add(new Term(field, "content1"));
Document doc = new Document();
doc.Add(NewStringField(field, "content1", Field.Store.YES));
w.AddDocument(doc);
}
int randomFields = Random.Next(10);
for (int i = 0; i < randomFields; i++)
{
while (true)
{
string field = "field" + Random.Next(100);
Term t = new Term(field, "content1");
if (!terms.Contains(t))
{
terms.Add(t);
break;
}
}
}
w.ForceMerge(1);
IndexReader reader = w.GetReader();
w.Dispose();
assertEquals(1, reader.Leaves.size());
AtomicReaderContext context = reader.Leaves.First();
TermsFilter tf = new TermsFilter(terms.ToList());
FixedBitSet bits = (FixedBitSet)tf.GetDocIdSet(context, context.AtomicReader.LiveDocs);
assertEquals(context.Reader.NumDocs, bits.Cardinality());
reader.Dispose();
dir.Dispose();
}
public virtual void TestSimple()
{
Directory dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir);
int numDocs = AtLeast(100);
for (int i = 0; i < numDocs; i++)
{
iw.AddDocument(Doc());
}
IndexReader ir = iw.Reader;
VerifyCount(ir);
ir.Dispose();
iw.ForceMerge(1);
ir = iw.Reader;
VerifyCount(ir);
ir.Dispose();
iw.Dispose();
dir.Dispose();
}
public virtual void TestAcrossFields()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, new MockAnalyzer(Random(), MockTokenizer.WHITESPACE, true));
Document doc = new Document();
doc.Add(new TextField("hasMaybepayload", "here we go", Field.Store.YES));
writer.AddDocument(doc);
writer.Dispose();
writer = new RandomIndexWriter(Random(), dir, new MockAnalyzer(Random(), MockTokenizer.WHITESPACE, true));
doc = new Document();
doc.Add(new TextField("hasMaybepayload2", "here we go", Field.Store.YES));
writer.AddDocument(doc);
writer.AddDocument(doc);
writer.ForceMerge(1);
writer.Dispose();
dir.Dispose();
}
public virtual void TestBinary()
{
Directory dir = NewDirectory();
Document doc = new Document();
BytesRef @ref = new BytesRef();
Field field = new BinaryDocValuesField("bytes", @ref);
doc.Add(field);
IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, null);
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
int numDocs = AtLeast(500);
for (int i = 0; i < numDocs; i++)
{
@ref.CopyChars(TestUtil.RandomUnicodeString(Random()));
iw.AddDocument(doc);
if (Random().Next(17) == 0)
{
iw.Commit();
}
}
DirectoryReader ir = iw.Reader;
iw.ForceMerge(1);
DirectoryReader ir2 = iw.Reader;
AtomicReader merged = GetOnlySegmentReader(ir2);
iw.Dispose();
BinaryDocValues multi = MultiDocValues.GetBinaryValues(ir, "bytes");
BinaryDocValues single = merged.GetBinaryDocValues("bytes");
BytesRef actual = new BytesRef();
BytesRef expected = new BytesRef();
for (int i = 0; i < numDocs; i++)
{
single.Get(i, expected);
multi.Get(i, actual);
Assert.AreEqual(expected, actual);
}
ir.Dispose();
ir2.Dispose();
dir.Dispose();
}
public void TestFieldNotPresent()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
int num = AtLeast(3);
int skip = Random.Next(num);
var terms = new JCG.List <Term>();
for (int i = 0; i < num; i++)
{
terms.Add(new Term("field" + i, "content1"));
Document doc = new Document();
if (skip == i)
{
continue;
}
doc.Add(NewStringField("field" + i, "content1", Field.Store.YES));
w.AddDocument(doc);
}
w.ForceMerge(1);
IndexReader reader = w.GetReader();
w.Dispose();
assertEquals(1, reader.Leaves.size());
AtomicReaderContext context = reader.Leaves[0];
TermsFilter tf = new TermsFilter(terms);
FixedBitSet bits = (FixedBitSet)tf.GetDocIdSet(context, context.AtomicReader.LiveDocs);
assertEquals("Must be num fields - 1 since we skip only one field", num - 1, bits.Cardinality);
reader.Dispose();
dir.Dispose();
}
public override void SetUp()
{
base.SetUp();
directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random, directory, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetMergePolicy(NewLogMergePolicy()));
Document doc = new Document();
doc.Add(NewTextField("field", "one two three four five", Field.Store.YES));
doc.Add(NewTextField("sorter", "b", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two three four", Field.Store.YES));
doc.Add(NewTextField("sorter", "d", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two three y", Field.Store.YES));
doc.Add(NewTextField("sorter", "a", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two x", Field.Store.YES));
doc.Add(NewTextField("sorter", "c", Field.Store.YES));
writer.AddDocument(doc);
// tests here require single segment (eg try seed
// 8239472272678419952L), because SingleDocTestFilter(x)
// blindly accepts that docID in any sub-segment
writer.ForceMerge(1);
reader = writer.GetReader();
writer.Dispose();
searcher = NewSearcher(reader);
query = new TermQuery(new Term("field", "three"));
filter = NewStaticFilterB();
}
public override void BeforeClass()
{
base.BeforeClass();
dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
AddSome(doc, alwaysTerms);
if (Random.Next(100) < 90)
{
AddSome(doc, commonTerms);
}
if (Random.Next(100) < 50)
{
AddSome(doc, mediumTerms);
}
if (Random.Next(100) < 10)
{
AddSome(doc, rareTerms);
}
iw.AddDocument(doc);
}
iw.ForceMerge(1);
iw.Dispose();
r = DirectoryReader.Open(dir);
atomicReader = GetOnlySegmentReader(r);
searcher = new IndexSearcher(atomicReader);
searcher.Similarity = new DefaultSimilarityAnonymousInnerClassHelper();
}
public override void SetUp()
{
base.SetUp();
dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer
(Random));
iwc.SetMergePolicy(NewLogMergePolicy());
var iw = new RandomIndexWriter(Random, dir, iwc);
var doc = new Document
{
NewStringField("id", "1", Field.Store.YES),
NewTextField("body", "some contents and more contents", Field.Store.NO),
new NumericDocValuesField("popularity", 5)
};
iw.AddDocument(doc);
doc = new Document
{
NewStringField("id", "2", Field.Store.YES),
NewTextField("body", "another document with different contents", Field.Store.NO),
new NumericDocValuesField("popularity", 20)
};
iw.AddDocument(doc);
doc = new Document
{
NewStringField("id", "3", Field.Store.YES),
NewTextField("body", "crappy contents", Field.Store.NO),
new NumericDocValuesField("popularity", 2)
};
iw.AddDocument(doc);
iw.ForceMerge(1);
reader = iw.GetReader();
iw.Dispose();
}
public virtual void TestAddIndexes()
{
Directory d1 = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d1, Similarity, TimeZone);
Document doc = new Document();
doc.Add(NewStringField("id", "1", Field.Store.YES));
doc.Add(new NumericDocValuesField("dv", 1));
w.AddDocument(doc);
IndexReader r1 = w.Reader;
w.Dispose();
Directory d2 = NewDirectory();
w = new RandomIndexWriter(Random(), d2, Similarity, TimeZone);
doc = new Document();
doc.Add(NewStringField("id", "2", Field.Store.YES));
doc.Add(new NumericDocValuesField("dv", 2));
w.AddDocument(doc);
IndexReader r2 = w.Reader;
w.Dispose();
Directory d3 = NewDirectory();
w = new RandomIndexWriter(Random(), d3, Similarity, TimeZone);
w.AddIndexes(SlowCompositeReaderWrapper.Wrap(r1), SlowCompositeReaderWrapper.Wrap(r2));
r1.Dispose();
d1.Dispose();
r2.Dispose();
d2.Dispose();
w.ForceMerge(1);
DirectoryReader r3 = w.Reader;
w.Dispose();
AtomicReader sr = GetOnlySegmentReader(r3);
Assert.AreEqual(2, sr.NumDocs);
NumericDocValues docValues = sr.GetNumericDocValues("dv");
Assert.IsNotNull(docValues);
r3.Dispose();
d3.Dispose();
}
public virtual void TestNumerics()
{
Directory dir = NewDirectory();
Document doc = new Document();
Field field = new NumericDocValuesField("numbers", 0);
doc.Add(field);
IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, null);
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
int numDocs = AtLeast(500);
for (int i = 0; i < numDocs; i++)
{
field.LongValue = Random().NextLong();
iw.AddDocument(doc);
if (Random().Next(17) == 0)
{
iw.Commit();
}
}
DirectoryReader ir = iw.Reader;
iw.ForceMerge(1);
DirectoryReader ir2 = iw.Reader;
AtomicReader merged = GetOnlySegmentReader(ir2);
iw.Dispose();
NumericDocValues multi = MultiDocValues.GetNumericValues(ir, "numbers");
NumericDocValues single = merged.GetNumericDocValues("numbers");
for (int i = 0; i < numDocs; i++)
{
Assert.AreEqual(single.Get(i), multi.Get(i));
}
ir.Dispose();
ir2.Dispose();
dir.Dispose();
}
public void TestFieldNotPresent()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
int num = AtLeast(3);
int skip = Random().Next(num);
var terms = new List <Term>();
for (int i = 0; i < num; i++)
{
terms.Add(new Term("field" + i, "content1"));
Document doc = new Document();
if (skip == i)
{
continue;
}
doc.Add(NewStringField("field" + i, "content1", Field.Store.YES));
w.AddDocument(doc);
}
w.ForceMerge(1);
IndexReader reader = w.Reader;
w.Dispose();
assertEquals(1, reader.Leaves.size());
AtomicReaderContext context = reader.Leaves.First();
TermsFilter tf = new TermsFilter(terms);
FixedBitSet bits = (FixedBitSet)tf.GetDocIdSet(context, context.AtomicReader.LiveDocs);
assertEquals("Must be num fields - 1 since we skip only one field", num - 1, bits.Cardinality());
reader.Dispose();
dir.Dispose();
}
public override void SetUp()
{
base.SetUp();
directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), directory, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(NewLogMergePolicy()));
//Add series of docs with filterable fields : url, text and dates flags
AddDoc(writer, "http://lucene.apache.org", "lucene 1.4.3 available", "20040101");
AddDoc(writer, "http://lucene.apache.org", "New release pending", "20040102");
AddDoc(writer, "http://lucene.apache.org", "Lucene 1.9 out now", "20050101");
AddDoc(writer, "http://www.bar.com", "Local man bites dog", "20040101");
AddDoc(writer, "http://www.bar.com", "Dog bites local man", "20040102");
AddDoc(writer, "http://www.bar.com", "Dog uses Lucene", "20050101");
AddDoc(writer, "http://lucene.apache.org", "Lucene 2.0 out", "20050101");
AddDoc(writer, "http://lucene.apache.org", "Oops. Lucene 2.1 out", "20050102");
// Until we fix LUCENE-2348, the index must
// have only 1 segment:
writer.ForceMerge(1);
reader = writer.Reader;
writer.Dispose();
searcher = NewSearcher(reader);
}
[AwaitsFix(BugUrl = "https://github.com/apache/lucenenet/issues/269")] // LUCENENET TODO: this test fails on x86 on .NET Framework in Release mode only
#endif
public virtual void TestBS2DisjunctionNextVsAdvance()
{
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, d);
int numDocs = AtLeast(300);
for (int docUpto = 0; docUpto < numDocs; docUpto++)
{
string contents = "a";
if (Random.Next(20) <= 16)
{
contents += " b";
}
if (Random.Next(20) <= 8)
{
contents += " c";
}
if (Random.Next(20) <= 4)
{
contents += " d";
}
if (Random.Next(20) <= 2)
{
contents += " e";
}
if (Random.Next(20) <= 1)
{
contents += " f";
}
Document doc = new Document();
doc.Add(new TextField("field", contents, Field.Store.NO));
w.AddDocument(doc);
}
w.ForceMerge(1);
IndexReader r = w.GetReader();
IndexSearcher s = NewSearcher(r);
w.Dispose();
for (int iter = 0; iter < 10 * RandomMultiplier; iter++)
{
if (Verbose)
{
Console.WriteLine("iter=" + iter);
}
IList <string> terms = new List <string> {
"a", "b", "c", "d", "e", "f"
};
int numTerms = TestUtil.NextInt32(Random, 1, terms.Count);
while (terms.Count > numTerms)
{
terms.RemoveAt(Random.Next(terms.Count));
}
if (Verbose)
{
Console.WriteLine(" terms=" + terms);
}
BooleanQuery q = new BooleanQuery();
foreach (string term in terms)
{
q.Add(new BooleanClause(new TermQuery(new Term("field", term)), Occur.SHOULD));
}
Weight weight = s.CreateNormalizedWeight(q);
Scorer scorer = weight.GetScorer(s.m_leafContexts[0], null);
// First pass: just use .NextDoc() to gather all hits
IList <ScoreDoc> hits = new List <ScoreDoc>();
while (scorer.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
hits.Add(new ScoreDoc(scorer.DocID, scorer.GetScore()));
}
if (Verbose)
{
Console.WriteLine(" " + hits.Count + " hits");
}
// Now, randomly next/advance through the list and
// verify exact match:
for (int iter2 = 0; iter2 < 10; iter2++)
{
weight = s.CreateNormalizedWeight(q);
scorer = weight.GetScorer(s.m_leafContexts[0], null);
if (Verbose)
{
Console.WriteLine(" iter2=" + iter2);
}
int upto = -1;
while (upto < hits.Count)
{
int nextUpto;
int nextDoc;
int left = hits.Count - upto;
if (left == 1 || Random.nextBoolean())
{
// next
nextUpto = 1 + upto;
nextDoc = scorer.NextDoc();
}
else
{
// advance
int inc = TestUtil.NextInt32(Random, 1, left - 1);
nextUpto = inc + upto;
nextDoc = scorer.Advance(hits[nextUpto].Doc);
}
if (nextUpto == hits.Count)
{
Assert.AreEqual(DocIdSetIterator.NO_MORE_DOCS, nextDoc);
}
else
{
ScoreDoc hit = hits[nextUpto];
Assert.AreEqual(hit.Doc, nextDoc);
// Test for precise float equality:
Assert.IsTrue(hit.Score == scorer.GetScore(), "doc " + hit.Doc + " has wrong score: expected=" + hit.Score + " actual=" + scorer.GetScore());
}
upto = nextUpto;
}
}
}
r.Dispose();
d.Dispose();
}
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 TestAdvanceSingleParentNoChild()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(new LogDocMergePolicy()));
Document parentDoc = new Document();
parentDoc.Add(NewStringField("parent", "1", Field.Store.NO));
parentDoc.Add(NewStringField("isparent", "yes", Field.Store.NO));
w.AddDocuments(Arrays.AsList(parentDoc));
// Add another doc so scorer is not null
parentDoc = new Document();
parentDoc.Add(NewStringField("parent", "2", Field.Store.NO));
parentDoc.Add(NewStringField("isparent", "yes", Field.Store.NO));
Document childDoc = new Document();
childDoc.Add(NewStringField("child", "2", Field.Store.NO));
w.AddDocuments(Arrays.AsList(childDoc, parentDoc));
// Need single seg:
w.ForceMerge(1);
IndexReader r = w.Reader;
w.Dispose();
IndexSearcher s = NewSearcher(r);
Query tq = new TermQuery(new Term("child", "2"));
Filter parentFilter = new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("isparent", "yes"))));
ToParentBlockJoinQuery q = new ToParentBlockJoinQuery(tq, parentFilter, ScoreMode.Avg);
Weight weight = s.CreateNormalizedWeight(q);
DocIdSetIterator disi = weight.Scorer(s.IndexReader.Leaves.First(), null);
assertEquals(2, disi.Advance(0));
r.Dispose();
dir.Dispose();
}
public virtual void Test()
{
IList<string> postingsList = new List<string>();
int numTerms = AtLeast(300);
int maxTermsPerDoc = TestUtil.NextInt(Random(), 10, 20);
bool isSimpleText = "SimpleText".Equals(TestUtil.GetPostingsFormat("field"));
IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
if ((isSimpleText || iwc.MergePolicy is MockRandomMergePolicy) && (TEST_NIGHTLY || RANDOM_MULTIPLIER > 1))
{
// Otherwise test can take way too long (> 2 hours)
numTerms /= 2;
}
if (VERBOSE)
{
Console.WriteLine("maxTermsPerDoc=" + maxTermsPerDoc);
Console.WriteLine("numTerms=" + numTerms);
}
for (int i = 0; i < numTerms; i++)
{
string term = Convert.ToString(i);
for (int j = 0; j < i; j++)
{
postingsList.Add(term);
}
}
postingsList = CollectionsHelper.Shuffle(postingsList);
ConcurrentQueue<string> postings = new ConcurrentQueue<string>(postingsList);
Directory dir = NewFSDirectory(CreateTempDir(GetFullMethodName()));
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
int threadCount = TestUtil.NextInt(Random(), 1, 5);
if (VERBOSE)
{
Console.WriteLine("config: " + iw.w.Config);
Console.WriteLine("threadCount=" + threadCount);
}
Field prototype = NewTextField("field", "", Field.Store.NO);
FieldType fieldType = new FieldType((FieldType)prototype.FieldType);
if (Random().NextBoolean())
{
fieldType.OmitNorms = true;
}
int options = Random().Next(3);
if (options == 0)
{
fieldType.IndexOptions = FieldInfo.IndexOptions.DOCS_AND_FREQS; // we dont actually need positions
fieldType.StoreTermVectors = true; // but enforce term vectors when we do this so we check SOMETHING
}
else if (options == 1 && !DoesntSupportOffsets.Contains(TestUtil.GetPostingsFormat("field")))
{
fieldType.IndexOptions = FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS;
}
// else just positions
ThreadClass[] threads = new ThreadClass[threadCount];
CountdownEvent startingGun = new CountdownEvent(1);
for (int threadID = 0; threadID < threadCount; threadID++)
{
Random threadRandom = new Random(Random().Next());
Document document = new Document();
Field field = new Field("field", "", fieldType);
document.Add(field);
threads[threadID] = new ThreadAnonymousInnerClassHelper(this, numTerms, maxTermsPerDoc, postings, iw, startingGun, threadRandom, document, field);
threads[threadID].Start();
}
startingGun.Signal();
foreach (ThreadClass t in threads)
{
t.Join();
}
iw.ForceMerge(1);
DirectoryReader ir = iw.Reader;
Assert.AreEqual(1, ir.Leaves.Count);
AtomicReader air = (AtomicReader)ir.Leaves[0].Reader;
Terms terms = air.Terms("field");
// numTerms-1 because there cannot be a term 0 with 0 postings:
Assert.AreEqual(numTerms - 1, terms.Size());
TermsEnum termsEnum = terms.Iterator(null);
BytesRef termBR;
while ((termBR = termsEnum.Next()) != null)
{
int value = Convert.ToInt32(termBR.Utf8ToString());
Assert.AreEqual(value, termsEnum.TotalTermFreq());
// don't really need to check more than this, as CheckIndex
// will verify that totalTermFreq == total number of positions seen
// from a docsAndPositionsEnum.
}
ir.Dispose();
iw.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();
}
public override void SetUp()
{
base.SetUp();
Directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), Directory, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(NewLogMergePolicy()));
Document doc = new Document();
doc.Add(NewTextField("field", "one two three four five", Field.Store.YES));
doc.Add(NewTextField("sorter", "b", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two three four", Field.Store.YES));
doc.Add(NewTextField("sorter", "d", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two three y", Field.Store.YES));
doc.Add(NewTextField("sorter", "a", Field.Store.YES));
writer.AddDocument(doc);
doc = new Document();
doc.Add(NewTextField("field", "one two x", Field.Store.YES));
doc.Add(NewTextField("sorter", "c", Field.Store.YES));
writer.AddDocument(doc);
// tests here require single segment (eg try seed
// 8239472272678419952L), because SingleDocTestFilter(x)
// blindly accepts that docID in any sub-segment
writer.ForceMerge(1);
Reader = writer.Reader;
writer.Dispose();
Searcher = NewSearcher(Reader);
Query = new TermQuery(new Term("field", "three"));
Filter = NewStaticFilterB();
}
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();
}
// [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 TestRandomIndex()
{
Directory dir = NewDirectory();
MockAnalyzer analyzer = new MockAnalyzer(Random);
analyzer.MaxTokenLength = TestUtil.NextInt32(Random, 1, IndexWriter.MAX_TERM_LENGTH);
RandomIndexWriter w = new RandomIndexWriter(Random, dir, analyzer);
CreateRandomIndex(AtLeast(50), w, Random.NextInt64());
DirectoryReader reader = w.GetReader();
AtomicReader wrapper = SlowCompositeReaderWrapper.Wrap(reader);
string field = @"body";
Terms terms = wrapper.GetTerms(field);
var lowFreqQueue = new PriorityQueueAnonymousClass(5);
var highFreqQueue = new PriorityQueueAnonymousClass1(5);
try
{
TermsEnum iterator = terms.GetEnumerator();
while (iterator.MoveNext())
{
if (highFreqQueue.Count < 5)
{
highFreqQueue.Add(new TermAndFreq(
BytesRef.DeepCopyOf(iterator.Term), iterator.DocFreq));
lowFreqQueue.Add(new TermAndFreq(
BytesRef.DeepCopyOf(iterator.Term), iterator.DocFreq));
}
else
{
if (highFreqQueue.Top.freq < iterator.DocFreq)
{
highFreqQueue.Top.freq = iterator.DocFreq;
highFreqQueue.Top.term = BytesRef.DeepCopyOf(iterator.Term);
highFreqQueue.UpdateTop();
}
if (lowFreqQueue.Top.freq > iterator.DocFreq)
{
lowFreqQueue.Top.freq = iterator.DocFreq;
lowFreqQueue.Top.term = BytesRef.DeepCopyOf(iterator.Term);
lowFreqQueue.UpdateTop();
}
}
}
int lowFreq = lowFreqQueue.Top.freq;
int highFreq = highFreqQueue.Top.freq;
AssumeTrue(@"unlucky index", highFreq - 1 > lowFreq);
IList <TermAndFreq> highTerms = QueueToList(highFreqQueue);
IList <TermAndFreq> lowTerms = QueueToList(lowFreqQueue);
IndexSearcher searcher = NewSearcher(reader);
Occur lowFreqOccur = RandomOccur(Random);
BooleanQuery verifyQuery = new BooleanQuery();
CommonTermsQuery cq = new CommonTermsQuery(RandomOccur(Random),
lowFreqOccur, highFreq - 1, Random.NextBoolean());
foreach (TermAndFreq termAndFreq in lowTerms)
{
cq.Add(new Term(field, termAndFreq.term));
verifyQuery.Add(new BooleanClause(new TermQuery(new Term(field,
termAndFreq.term)), lowFreqOccur));
}
foreach (TermAndFreq termAndFreq in highTerms)
{
cq.Add(new Term(field, termAndFreq.term));
}
TopDocs cqSearch = searcher.Search(cq, reader.MaxDoc);
TopDocs verifySearch = searcher.Search(verifyQuery, reader.MaxDoc);
assertEquals(verifySearch.TotalHits, cqSearch.TotalHits);
var hits = new JCG.HashSet <int>();
foreach (ScoreDoc doc in verifySearch.ScoreDocs)
{
hits.Add(doc.Doc);
}
foreach (ScoreDoc doc in cqSearch.ScoreDocs)
{
assertTrue(hits.Remove(doc.Doc));
}
assertTrue(hits.Count == 0);
/*
* need to force merge here since QueryUtils adds checks based
* on leave readers which have different statistics than the top
* level reader if we have more than one segment. This could
* result in a different query / results.
*/
w.ForceMerge(1);
DirectoryReader reader2 = w.GetReader();
QueryUtils.Check(Random, cq, NewSearcher(reader2));
reader2.Dispose();
}
finally
{
reader.Dispose();
wrapper.Dispose();
w.Dispose();
dir.Dispose();
}
}
public static void BeforeClass()
{
Dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(Random(), Dir);
int numDocs = AtLeast(300);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
AddSome(doc, AlwaysTerms);
if (Random().Next(100) < 90)
{
AddSome(doc, CommonTerms);
}
if (Random().Next(100) < 50)
{
AddSome(doc, MediumTerms);
}
if (Random().Next(100) < 10)
{
AddSome(doc, RareTerms);
}
iw.AddDocument(doc);
}
iw.ForceMerge(1);
iw.Dispose();
r = DirectoryReader.Open(Dir);
atomicReader = GetOnlySegmentReader(r);
Searcher = new IndexSearcher(atomicReader);
Searcher.Similarity = new DefaultSimilarityAnonymousInnerClassHelper();
}
private void Populate(Directory directory, IndexWriterConfig config)
{
RandomIndexWriter writer = new RandomIndexWriter(Random(), directory, config);
for (int i = 0; i < NUMBER_OF_DOCUMENTS; i++)
{
Document document = new Document();
for (int f = 0; f < NUMBER_OF_FIELDS; f++)
{
document.Add(NewStringField("field" + f, Text, Field.Store.NO));
}
writer.AddDocument(document);
}
writer.ForceMerge(1);
writer.Dispose();
}
public virtual void TestDifferentTypedDocValuesField2()
{
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
// Index doc values are single-valued so we should not
// be able to add same field more than once:
Field f = new NumericDocValuesField("field", 17);
doc.Add(f);
doc.Add(new SortedDocValuesField("field", new BytesRef("hello")));
try
{
w.AddDocument(doc);
Assert.Fail("didn't hit expected exception");
}
catch (System.ArgumentException iae)
{
// expected
}
doc = new Document();
doc.Add(f);
w.AddDocument(doc);
w.ForceMerge(1);
DirectoryReader r = w.Reader;
Assert.AreEqual(17, GetOnlySegmentReader(r).GetNumericDocValues("field").Get(0));
r.Dispose();
w.Dispose();
d.Dispose();
}
/// <summary>
/// Indexes at least 1 document with f1, and at least 1 document with f2.
/// returns the norms for "field".
/// </summary>
internal virtual NumericDocValues GetNorms(string field, Field f1, Field f2)
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter riw = new RandomIndexWriter(Random(), dir, iwc);
// add f1
Document d = new Document();
d.Add(f1);
riw.AddDocument(d);
// add f2
d = new Document();
d.Add(f2);
riw.AddDocument(d);
// add a mix of f1's and f2's
int numExtraDocs = TestUtil.NextInt(Random(), 1, 1000);
for (int i = 0; i < numExtraDocs; i++)
{
d = new Document();
d.Add(Random().NextBoolean() ? f1 : f2);
riw.AddDocument(d);
}
IndexReader ir1 = riw.Reader;
// todo: generalize
NumericDocValues norms1 = MultiDocValues.GetNormValues(ir1, field);
// fully merge and validate MultiNorms against single segment.
riw.ForceMerge(1);
DirectoryReader ir2 = riw.Reader;
NumericDocValues norms2 = GetOnlySegmentReader(ir2).GetNormValues(field);
if (norms1 == null)
{
Assert.IsNull(norms2);
}
else
{
for (int docID = 0; docID < ir1.MaxDoc; docID++)
{
Assert.AreEqual(norms1.Get(docID), norms2.Get(docID));
}
}
ir1.Dispose();
ir2.Dispose();
riw.Dispose();
dir.Dispose();
return norms1;
}
public 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)
Directory dir = new MockDirectoryWrapper(Random(), new MMapDirectory(CreateTempDir("testBigDocuments")));
IndexWriterConfig iwConf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwConf.SetMaxBufferedDocs(RandomInts.NextIntBetween(Random(), 2, 30));
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwConf);
if (dir is MockDirectoryWrapper)
{
((MockDirectoryWrapper)dir).Throttling = MockDirectoryWrapper.Throttling_e.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.Indexed = false;
Field smallField = new Field("fld", RandomByteArray(Random().Next(10), 256), onlyStored);
int numFields = RandomInts.NextIntBetween(Random(), 500000, 1000000);
for (int i = 0; i < numFields; ++i)
{
bigDoc1.Add(smallField);
}
Field bigField = new Field("fld", RandomByteArray(RandomInts.NextIntBetween(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] = RandomInts.RandomFrom(Random(), Arrays.AsList(emptyDoc, bigDoc1, bigDoc2));
}
for (int i = 0; i < numDocs; ++i)
{
idField.StringValue = "" + 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
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);
IndexableField[] fieldValues = doc.GetFields("fld");
Assert.AreEqual(docs[i].GetFields("fld").Length, fieldValues.Length);
if (fieldValues.Length > 0)
{
Assert.AreEqual(docs[i].GetFields("fld")[0].BinaryValue(), fieldValues[0].BinaryValue());
}
}
rd.Dispose();
iw.Dispose();
dir.Dispose();
}
public virtual void TestIntersectStartTerm()
{
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
iwc.SetMergePolicy(new LogDocMergePolicy());
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, iwc);
Document doc = new Document();
doc.Add(NewStringField("field", "abc", Field.Store.NO));
w.AddDocument(doc);
doc = new Document();
doc.Add(NewStringField("field", "abd", Field.Store.NO));
w.AddDocument(doc);
doc = new Document();
doc.Add(NewStringField("field", "acd", Field.Store.NO));
w.AddDocument(doc);
doc = new Document();
doc.Add(NewStringField("field", "bcd", Field.Store.NO));
w.AddDocument(doc);
w.ForceMerge(1);
DirectoryReader r = w.Reader;
w.Dispose();
AtomicReader sub = GetOnlySegmentReader(r);
Terms terms = sub.Fields.Terms("field");
Automaton automaton = (new RegExp(".*d", RegExp.NONE)).ToAutomaton();
CompiledAutomaton ca = new CompiledAutomaton(automaton, false, false);
TermsEnum te;
// should seek to startTerm
te = terms.Intersect(ca, new BytesRef("aad"));
Assert.AreEqual("abd", te.Next().Utf8ToString());
Assert.AreEqual(1, te.Docs(null, null, DocsEnum.FLAG_NONE).NextDoc());
Assert.AreEqual("acd", te.Next().Utf8ToString());
Assert.AreEqual(2, te.Docs(null, null, DocsEnum.FLAG_NONE).NextDoc());
Assert.AreEqual("bcd", te.Next().Utf8ToString());
Assert.AreEqual(3, te.Docs(null, null, DocsEnum.FLAG_NONE).NextDoc());
Assert.IsNull(te.Next());
// should fail to find ceil label on second arc, rewind
te = terms.Intersect(ca, new BytesRef("add"));
Assert.AreEqual("bcd", te.Next().Utf8ToString());
Assert.AreEqual(3, te.Docs(null, null, DocsEnum.FLAG_NONE).NextDoc());
Assert.IsNull(te.Next());
// should reach end
te = terms.Intersect(ca, new BytesRef("bcd"));
Assert.IsNull(te.Next());
te = terms.Intersect(ca, new BytesRef("ddd"));
Assert.IsNull(te.Next());
r.Dispose();
dir.Dispose();
}
public virtual void Test2()
{
Random random = Random();
int NUM_DOCS = AtLeast(100);
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(random, dir);
bool allowDups = random.NextBoolean();
HashSet<string> seen = new HashSet<string>();
if (VERBOSE)
{
Console.WriteLine("TEST: NUM_DOCS=" + NUM_DOCS + " allowDups=" + allowDups);
}
int numDocs = 0;
IList<BytesRef> docValues = new List<BytesRef>();
// TODO: deletions
while (numDocs < NUM_DOCS)
{
string s;
if (random.NextBoolean())
{
s = TestUtil.RandomSimpleString(random);
}
else
{
s = TestUtil.RandomUnicodeString(random);
}
BytesRef br = new BytesRef(s);
if (!allowDups)
{
if (seen.Contains(s))
{
continue;
}
seen.Add(s);
}
if (VERBOSE)
{
Console.WriteLine(" " + numDocs + ": s=" + s);
}
Document doc = new Document();
doc.Add(new SortedDocValuesField("stringdv", br));
doc.Add(new NumericDocValuesField("id", numDocs));
docValues.Add(br);
writer.AddDocument(doc);
numDocs++;
if (random.Next(40) == 17)
{
// force flush
writer.Reader.Dispose();
}
}
writer.ForceMerge(1);
DirectoryReader r = writer.Reader;
writer.Dispose();
AtomicReader sr = GetOnlySegmentReader(r);
long END_TIME = DateTime.Now.Millisecond + (TEST_NIGHTLY ? 30 : 1);
int NUM_THREADS = TestUtil.NextInt(Random(), 1, 10);
ThreadClass[] threads = new ThreadClass[NUM_THREADS];
for (int thread = 0; thread < NUM_THREADS; thread++)
{
threads[thread] = new ThreadAnonymousInnerClassHelper2(this, random, docValues, sr, END_TIME);
threads[thread].Start();
}
foreach (ThreadClass thread in threads)
{
thread.Join();
}
r.Dispose();
dir.Dispose();
}
public virtual void TestCustomDoublesValueSource()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
Document doc = new Document();
writer.AddDocument(doc);
writer.AddDocument(doc);
writer.AddDocument(doc);
// Test wants 3 docs in one segment:
writer.ForceMerge(1);
var vs = new ValueSourceAnonymousInnerClassHelper(this, doc);
FacetsConfig config = new FacetsConfig();
FacetsCollector fc = new FacetsCollector();
IndexReader r = writer.Reader;
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
DoubleRange[] ranges = new DoubleRange[] { new DoubleRange("< 1", 0.0, true, 1.0, false), new DoubleRange("< 2", 0.0, true, 2.0, false), new DoubleRange("< 5", 0.0, true, 5.0, false), new DoubleRange("< 10", 0.0, true, 10.0, false), new DoubleRange("< 20", 0.0, true, 20.0, false), new DoubleRange("< 50", 0.0, true, 50.0, false) };
Filter fastMatchFilter;
AtomicBoolean filterWasUsed = new AtomicBoolean();
if (Random().NextBoolean())
{
// Sort of silly:
fastMatchFilter = new CachingWrapperFilterAnonymousInnerClassHelper(this, new QueryWrapperFilter(new MatchAllDocsQuery()), filterWasUsed);
}
else
{
fastMatchFilter = null;
}
if (VERBOSE)
{
Console.WriteLine("TEST: fastMatchFilter=" + fastMatchFilter);
}
Facets facets = new DoubleRangeFacetCounts("field", vs, fc, fastMatchFilter, ranges);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", facets.GetTopChildren(10, "field").ToString());
Assert.True(fastMatchFilter == null || filterWasUsed.Get());
DrillDownQuery ddq = new DrillDownQuery(config);
ddq.Add("field", ranges[1].GetFilter(fastMatchFilter, vs));
// Test simple drill-down:
Assert.AreEqual(1, s.Search(ddq, 10).TotalHits);
// Test drill-sideways after drill-down
DrillSideways ds = new DrillSidewaysAnonymousInnerClassHelper2(this, s, config, (TaxonomyReader)null, vs, ranges, fastMatchFilter);
DrillSidewaysResult dsr = ds.Search(ddq, 10);
Assert.AreEqual(1, dsr.Hits.TotalHits);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", dsr.Facets.GetTopChildren(10, "field").ToString());
IOUtils.Close(r, writer, dir);
}
public virtual void TestStressAdvance_Mem()
{
for (int iter = 0; iter < 3; iter++)
{
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter=" + iter);
}
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir);
HashSet<int> aDocs = new HashSet<int>();
Documents.Document doc = new Documents.Document();
Field f = NewStringField("field", "", Field.Store.NO);
doc.Add(f);
Field idField = NewStringField("id", "", Field.Store.YES);
doc.Add(idField);
int num = AtLeast(4097);
if (VERBOSE)
{
Console.WriteLine("\nTEST: numDocs=" + num);
}
for (int id = 0; id < num; id++)
{
if (Random().Next(4) == 3)
{
f.StringValue = "a";
aDocs.Add(id);
}
else
{
f.StringValue = "b";
}
idField.StringValue = "" + id;
w.AddDocument(doc);
if (VERBOSE)
{
Console.WriteLine("\nTEST: doc upto " + id);
}
}
w.ForceMerge(1);
IList<int> aDocIDs = new List<int>();
IList<int> bDocIDs = new List<int>();
DirectoryReader r = w.Reader;
int[] idToDocID = new int[r.MaxDoc];
for (int docID = 0; docID < idToDocID.Length; docID++)
{
int id = Convert.ToInt32(r.Document(docID).Get("id"));
if (aDocs.Contains(id))
{
aDocIDs.Add(docID);
}
else
{
bDocIDs.Add(docID);
}
}
TermsEnum te = GetOnlySegmentReader(r).Fields.Terms("field").Iterator(null);
DocsEnum de = null;
for (int iter2 = 0; iter2 < 10; iter2++)
{
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter=" + iter + " iter2=" + iter2);
}
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, te.SeekCeil(new BytesRef("a")));
de = TestUtil.Docs(Random(), te, null, de, DocsEnum.FLAG_NONE);
TestOne(de, aDocIDs);
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, te.SeekCeil(new BytesRef("b")));
de = TestUtil.Docs(Random(), te, null, de, DocsEnum.FLAG_NONE);
TestOne(de, bDocIDs);
}
w.Dispose();
r.Dispose();
dir.Dispose();
}
}
public void Test()
{
RandomIndexWriter writer;
DirectoryReader indexReader;
int numParents = AtLeast(200);
IndexWriterConfig cfg = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
cfg.SetMergePolicy(NewLogMergePolicy());
using (writer = new RandomIndexWriter(Random(), NewDirectory(), cfg))
{
Document parentDoc = new Document();
NumericDocValuesField parentVal = new NumericDocValuesField("parent_val", 0L);
parentDoc.Add(parentVal);
StringField parent = new StringField("parent", "true", Field.Store.YES);
parentDoc.Add(parent);
for (int i = 0; i < numParents; ++i)
{
List<Document> documents = new List<Document>();
int numChildren = Random().nextInt(10);
for (int j = 0; j < numChildren; ++j)
{
Document childDoc = new Document();
childDoc.Add(new NumericDocValuesField("child_val", Random().nextInt(5)));
documents.Add(childDoc);
}
parentVal.LongValue = (Random().nextInt(50));
documents.Add(parentDoc);
writer.AddDocuments(documents);
}
writer.ForceMerge(1);
indexReader = writer.Reader;
}
AtomicReader reader = GetOnlySegmentReader(indexReader);
Filter parentsFilter = new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("parent", "true"))));
FixedBitSet parentBits = (FixedBitSet)parentsFilter.GetDocIdSet(reader.AtomicContext, null);
NumericDocValues parentValues = reader.GetNumericDocValues("parent_val");
NumericDocValues childValues = reader.GetNumericDocValues("child_val");
Sort parentSort = new Sort(new SortField("parent_val", SortField.Type_e.LONG));
Sort childSort = new Sort(new SortField("child_val", SortField.Type_e.LONG));
Sort sort = new Sort(new SortField("custom", new BlockJoinComparatorSource(parentsFilter, parentSort, childSort)));
Sorter sorter = new Sorter(sort);
Sorter.DocMap docMap = sorter.Sort(reader);
assertEquals(reader.MaxDoc, docMap.Count);
int[] children = new int[1];
int numChildren2 = 0;
int previousParent = -1;
for (int i = 0; i < docMap.Count; ++i)
{
int oldID = docMap.NewToOld(i);
if (parentBits.Get(oldID))
{
// check that we have the right children
for (int j = 0; j < numChildren2; ++j)
{
assertEquals(oldID, parentBits.NextSetBit(children[j]));
}
// check that children are sorted
for (int j = 1; j < numChildren2; ++j)
{
int doc1 = children[j - 1];
int doc2 = children[j];
if (childValues.Get(doc1) == childValues.Get(doc2))
{
assertTrue(doc1 < doc2); // sort is stable
}
else
{
assertTrue(childValues.Get(doc1) < childValues.Get(doc2));
}
}
// check that parents are sorted
if (previousParent != -1)
{
if (parentValues.Get(previousParent) == parentValues.Get(oldID))
{
assertTrue(previousParent < oldID);
}
else
{
assertTrue(parentValues.Get(previousParent) < parentValues.Get(oldID));
}
}
// reset
previousParent = oldID;
numChildren2 = 0;
}
else
{
children = ArrayUtil.Grow(children, numChildren2 + 1);
children[numChildren2++] = oldID;
}
}
indexReader.Dispose();
writer.w.Directory.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 void TestInsideBooleanQuery()
{
const string idField = "id";
const string toField = "productId";
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random, dir,
NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random))
.SetMergePolicy(NewLogMergePolicy()));
// 0
Document doc = new Document();
doc.Add(new TextField("description", "random text", Field.Store.NO));
doc.Add(new TextField("name", "name1", Field.Store.NO));
doc.Add(new TextField(idField, "7", Field.Store.NO));
w.AddDocument(doc);
// 1
doc = new Document();
doc.Add(new TextField("price", "10.0", Field.Store.NO));
doc.Add(new TextField(idField, "2", Field.Store.NO));
doc.Add(new TextField(toField, "7", Field.Store.NO));
w.AddDocument(doc);
// 2
doc = new Document();
doc.Add(new TextField("price", "20.0", Field.Store.NO));
doc.Add(new TextField(idField, "3", Field.Store.NO));
doc.Add(new TextField(toField, "7", Field.Store.NO));
w.AddDocument(doc);
// 3
doc = new Document();
doc.Add(new TextField("description", "more random text", Field.Store.NO));
doc.Add(new TextField("name", "name2", Field.Store.NO));
doc.Add(new TextField(idField, "0", Field.Store.NO));
w.AddDocument(doc);
w.Commit();
// 4
doc = new Document();
doc.Add(new TextField("price", "10.0", Field.Store.NO));
doc.Add(new TextField(idField, "5", Field.Store.NO));
doc.Add(new TextField(toField, "0", Field.Store.NO));
w.AddDocument(doc);
// 5
doc = new Document();
doc.Add(new TextField("price", "20.0", Field.Store.NO));
doc.Add(new TextField(idField, "6", Field.Store.NO));
doc.Add(new TextField(toField, "0", Field.Store.NO));
w.AddDocument(doc);
w.ForceMerge(1);
IndexSearcher indexSearcher = new IndexSearcher(w.GetReader());
w.Dispose();
// Search for product
Query joinQuery = JoinUtil.CreateJoinQuery(idField, false, toField,
new TermQuery(new Term("description", "random")), indexSearcher, ScoreMode.Avg);
BooleanQuery bq = new BooleanQuery();
bq.Add(joinQuery, Occur.SHOULD);
bq.Add(new TermQuery(new Term("id", "3")), Occur.SHOULD);
indexSearcher.Search(bq, new CollectorAnonymousClass());
indexSearcher.IndexReader.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 virtual void TestSortedSetWithDups()
{
AssumeTrue("codec does not support SORTED_SET", DefaultCodecSupportsSortedSet());
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, null);
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
int numDocs = AtLeast(500);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
int numValues = Random().Next(5);
for (int j = 0; j < numValues; j++)
{
doc.Add(new SortedSetDocValuesField("bytes", new BytesRef(TestUtil.RandomSimpleString(Random(), 2))));
}
iw.AddDocument(doc);
if (Random().Next(17) == 0)
{
iw.Commit();
}
}
DirectoryReader ir = iw.Reader;
iw.ForceMerge(1);
DirectoryReader ir2 = iw.Reader;
AtomicReader merged = GetOnlySegmentReader(ir2);
iw.Dispose();
SortedSetDocValues multi = MultiDocValues.GetSortedSetValues(ir, "bytes");
SortedSetDocValues single = merged.GetSortedSetDocValues("bytes");
if (multi == null)
{
Assert.IsNull(single);
}
else
{
Assert.AreEqual(single.ValueCount, multi.ValueCount);
BytesRef actual = new BytesRef();
BytesRef expected = new BytesRef();
// check values
for (long i = 0; i < single.ValueCount; i++)
{
single.LookupOrd(i, expected);
multi.LookupOrd(i, actual);
Assert.AreEqual(expected, actual);
}
// check ord list
for (int i = 0; i < numDocs; i++)
{
single.Document = i;
List<long?> expectedList = new List<long?>();
long ord;
while ((ord = single.NextOrd()) != SortedSetDocValues.NO_MORE_ORDS)
{
expectedList.Add(ord);
}
multi.Document = i;
int upto = 0;
while ((ord = multi.NextOrd()) != SortedSetDocValues.NO_MORE_ORDS)
{
Assert.AreEqual((long)expectedList[upto], ord);
upto++;
}
Assert.AreEqual(expectedList.Count, upto);
}
}
ir.Dispose();
ir2.Dispose();
dir.Dispose();
}
public virtual void TestTermUTF16SortOrder()
{
Random rnd = Random();
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(rnd, dir, Similarity, TimeZone);
Document d = new Document();
// Single segment
Field f = NewStringField("f", "", Field.Store.NO);
d.Add(f);
char[] chars = new char[2];
HashSet<string> allTerms = new HashSet<string>();
int num = AtLeast(200);
for (int i = 0; i < num; i++)
{
string s;
if (rnd.NextBoolean())
{
// Single char
if (rnd.NextBoolean())
{
// Above surrogates
chars[0] = (char)GetInt(rnd, 1 + UnicodeUtil.UNI_SUR_LOW_END, 0xffff);
}
else
{
// Below surrogates
chars[0] = (char)GetInt(rnd, 0, UnicodeUtil.UNI_SUR_HIGH_START - 1);
}
s = new string(chars, 0, 1);
}
else
{
// Surrogate pair
chars[0] = (char)GetInt(rnd, UnicodeUtil.UNI_SUR_HIGH_START, UnicodeUtil.UNI_SUR_HIGH_END);
Assert.IsTrue(((int)chars[0]) >= UnicodeUtil.UNI_SUR_HIGH_START && ((int)chars[0]) <= UnicodeUtil.UNI_SUR_HIGH_END);
chars[1] = (char)GetInt(rnd, UnicodeUtil.UNI_SUR_LOW_START, UnicodeUtil.UNI_SUR_LOW_END);
s = new string(chars, 0, 2);
}
allTerms.Add(s);
f.StringValue = s;
writer.AddDocument(d);
if ((1 + i) % 42 == 0)
{
writer.Commit();
}
}
IndexReader r = writer.Reader;
// Test each sub-segment
foreach (AtomicReaderContext ctx in r.Leaves)
{
CheckTermsOrder(ctx.Reader, allTerms, false);
}
CheckTermsOrder(r, allTerms, true);
// Test multi segment
r.Dispose();
writer.ForceMerge(1);
// Test single segment
r = writer.Reader;
CheckTermsOrder(r, allTerms, true);
r.Dispose();
writer.Dispose();
dir.Dispose();
}
public void TestRandomIndex()
{
Directory dir = NewDirectory();
MockAnalyzer analyzer = new MockAnalyzer(Random);
analyzer.MaxTokenLength = TestUtil.NextInt32(Random, 1, IndexWriter.MAX_TERM_LENGTH);
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir, analyzer);
CreateRandomIndex(AtLeast(50), w, Random.NextInt64());
DirectoryReader reader = w.GetReader();
AtomicReader wrapper = SlowCompositeReaderWrapper.Wrap(reader);
string field = @"body";
Terms terms = wrapper.GetTerms(field);
var lowFreqQueue = new AnonymousPriorityQueue(this, 5);
Util.PriorityQueue<TermAndFreq> highFreqQueue = new AnonymousPriorityQueue1(this, 5);
try
{
TermsEnum iterator = terms.GetIterator(null);
while (iterator.Next() != null)
{
if (highFreqQueue.Count < 5)
{
highFreqQueue.Add(new TermAndFreq(BytesRef.DeepCopyOf(iterator.Term), iterator.DocFreq));
lowFreqQueue.Add(new TermAndFreq(BytesRef.DeepCopyOf(iterator.Term), iterator.DocFreq));
}
else
{
if (highFreqQueue.Top.freq < iterator.DocFreq)
{
highFreqQueue.Top.freq = iterator.DocFreq;
highFreqQueue.Top.term = BytesRef.DeepCopyOf(iterator.Term);
highFreqQueue.UpdateTop();
}
if (lowFreqQueue.Top.freq > iterator.DocFreq)
{
lowFreqQueue.Top.freq = iterator.DocFreq;
lowFreqQueue.Top.term = BytesRef.DeepCopyOf(iterator.Term);
lowFreqQueue.UpdateTop();
}
}
}
int lowFreq = lowFreqQueue.Top.freq;
int highFreq = highFreqQueue.Top.freq;
AssumeTrue(@"unlucky index", highFreq - 1 > lowFreq);
List<TermAndFreq> highTerms = QueueToList(highFreqQueue);
List<TermAndFreq> lowTerms = QueueToList(lowFreqQueue);
IndexSearcher searcher = NewSearcher(reader);
Occur lowFreqOccur = RandomOccur(Random);
BooleanQuery verifyQuery = new BooleanQuery();
CommonTermsQuery cq = new CommonTermsQuery(RandomOccur(Random), lowFreqOccur, highFreq - 1, Random.NextBoolean());
foreach (TermAndFreq termAndFreq in lowTerms)
{
cq.Add(new Term(field, termAndFreq.term));
verifyQuery.Add(new BooleanClause(new TermQuery(new Term(field, termAndFreq.term)), lowFreqOccur));
}
foreach (TermAndFreq termAndFreq in highTerms)
{
cq.Add(new Term(field, termAndFreq.term));
}
TopDocs cqSearch = searcher.Search(cq, reader.MaxDoc);
TopDocs verifySearch = searcher.Search(verifyQuery, reader.MaxDoc);
assertEquals(verifySearch.TotalHits, cqSearch.TotalHits);
var hits = new JCG.HashSet<int>();
foreach (ScoreDoc doc in verifySearch.ScoreDocs)
{
hits.Add(doc.Doc);
}
foreach (ScoreDoc doc in cqSearch.ScoreDocs)
{
assertTrue(hits.Remove(doc.Doc));
}
assertTrue(hits.Count == 0);
w.ForceMerge(1);
DirectoryReader reader2 = w.GetReader();
QueryUtils.Check(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, cq, NewSearcher(reader2));
reader2.Dispose();
}
finally
{
reader.Dispose();
wrapper.Dispose();
w.Dispose();
dir.Dispose();
}
}
public virtual void TestDocsWithField()
{
AssumeTrue("codec does not support docsWithField", DefaultCodecSupportsDocsWithField());
Directory dir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, null);
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc);
int numDocs = AtLeast(500);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
if (Random().Next(4) >= 0)
{
doc.Add(new NumericDocValuesField("numbers", Random().NextLong()));
}
doc.Add(new NumericDocValuesField("numbersAlways", Random().NextLong()));
iw.AddDocument(doc);
if (Random().Next(17) == 0)
{
iw.Commit();
}
}
DirectoryReader ir = iw.Reader;
iw.ForceMerge(1);
DirectoryReader ir2 = iw.Reader;
AtomicReader merged = GetOnlySegmentReader(ir2);
iw.Dispose();
Bits multi = MultiDocValues.GetDocsWithField(ir, "numbers");
Bits single = merged.GetDocsWithField("numbers");
if (multi == null)
{
Assert.IsNull(single);
}
else
{
Assert.AreEqual(single.Length(), multi.Length());
for (int i = 0; i < numDocs; i++)
{
Assert.AreEqual(single.Get(i), multi.Get(i));
}
}
multi = MultiDocValues.GetDocsWithField(ir, "numbersAlways");
single = merged.GetDocsWithField("numbersAlways");
Assert.AreEqual(single.Length(), multi.Length());
for (int i = 0; i < numDocs; i++)
{
Assert.AreEqual(single.Get(i), multi.Get(i));
}
ir.Dispose();
ir2.Dispose();
dir.Dispose();
}
public virtual void TestMultiValuedDocValuesField()
{
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
Field f = new NumericDocValuesField("field", 17);
// Index doc values are single-valued so we should not
// be able to add same field more than once:
doc.Add(f);
doc.Add(f);
try
{
w.AddDocument(doc);
Assert.Fail("didn't hit expected exception");
}
catch (System.ArgumentException iae)
{
// expected
}
doc = new Document();
doc.Add(f);
w.AddDocument(doc);
w.ForceMerge(1);
DirectoryReader r = w.Reader;
w.Dispose();
Assert.AreEqual(17, FieldCache.DEFAULT.GetInts(GetOnlySegmentReader(r), "field", false).Get(0));
r.Dispose();
d.Dispose();
}
public void Test()
{
RandomIndexWriter writer;
DirectoryReader indexReader;
int numParents = AtLeast(200);
IndexWriterConfig cfg = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
cfg.SetMergePolicy(NewLogMergePolicy());
using (writer = new RandomIndexWriter(Random(), NewDirectory(), cfg))
{
Document parentDoc = new Document();
NumericDocValuesField parentVal = new NumericDocValuesField("parent_val", 0L);
parentDoc.Add(parentVal);
StringField parent = new StringField("parent", "true", Field.Store.YES);
parentDoc.Add(parent);
for (int i = 0; i < numParents; ++i)
{
List <Document> documents = new List <Document>();
int numChildren = Random().nextInt(10);
for (int j = 0; j < numChildren; ++j)
{
Document childDoc = new Document();
childDoc.Add(new NumericDocValuesField("child_val", Random().nextInt(5)));
documents.Add(childDoc);
}
parentVal.SetInt64Value(Random().nextInt(50));
documents.Add(parentDoc);
writer.AddDocuments(documents);
}
writer.ForceMerge(1);
indexReader = writer.Reader;
}
AtomicReader reader = GetOnlySegmentReader(indexReader);
Filter parentsFilter = new FixedBitSetCachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Term("parent", "true"))));
FixedBitSet parentBits = (FixedBitSet)parentsFilter.GetDocIdSet(reader.AtomicContext, null);
NumericDocValues parentValues = reader.GetNumericDocValues("parent_val");
NumericDocValues childValues = reader.GetNumericDocValues("child_val");
Sort parentSort = new Sort(new SortField("parent_val", SortFieldType.INT64));
Sort childSort = new Sort(new SortField("child_val", SortFieldType.INT64));
Sort sort = new Sort(new SortField("custom", new BlockJoinComparerSource(parentsFilter, parentSort, childSort)));
Sorter sorter = new Sorter(sort);
Sorter.DocMap docMap = sorter.Sort(reader);
assertEquals(reader.MaxDoc, docMap.Count);
int[] children = new int[1];
int numChildren2 = 0;
int previousParent = -1;
for (int i = 0; i < docMap.Count; ++i)
{
int oldID = docMap.NewToOld(i);
if (parentBits.Get(oldID))
{
// check that we have the right children
for (int j = 0; j < numChildren2; ++j)
{
assertEquals(oldID, parentBits.NextSetBit(children[j]));
}
// check that children are sorted
for (int j = 1; j < numChildren2; ++j)
{
int doc1 = children[j - 1];
int doc2 = children[j];
if (childValues.Get(doc1) == childValues.Get(doc2))
{
assertTrue(doc1 < doc2); // sort is stable
}
else
{
assertTrue(childValues.Get(doc1) < childValues.Get(doc2));
}
}
// check that parents are sorted
if (previousParent != -1)
{
if (parentValues.Get(previousParent) == parentValues.Get(oldID))
{
assertTrue(previousParent < oldID);
}
else
{
assertTrue(parentValues.Get(previousParent) < parentValues.Get(oldID));
}
}
// reset
previousParent = oldID;
numChildren2 = 0;
}
else
{
children = ArrayUtil.Grow(children, numChildren2 + 1);
children[numChildren2++] = oldID;
}
}
indexReader.Dispose();
writer.w.Directory.Dispose();
}
