| public SetNorm ( int doc, String field, byte value ) : void | ||
| doc | int | |
| field | String | |
| value | byte | |
| return | void | |
public virtual void TestIsCurrent()
{
Directory dir1 = GetDir1();
Directory dir2 = GetDir2();
ParallelReader pr = new ParallelReader();
pr.Add(IndexReader.Open(dir1));
pr.Add(IndexReader.Open(dir2));
Assert.IsTrue(pr.IsCurrent());
IndexReader modifier = IndexReader.Open(dir1);
modifier.SetNorm(0, "f1", 100);
modifier.Close();
// one of the two IndexReaders which ParallelReader is using
// is not current anymore
Assert.IsFalse(pr.IsCurrent());
modifier = IndexReader.Open(dir2);
modifier.SetNorm(0, "f3", 100);
modifier.Close();
// now both are not current anymore
Assert.IsFalse(pr.IsCurrent());
}
/// <summary> 1. Get a norm from the original reader 2. Clone the original reader 3.
/// Delete a document and set the norm of the cloned reader 4. Verify the norms
/// are not the same on each reader 5. Verify the doc deleted is only in the
/// cloned reader 6. Try to delete a document in the original reader, an
/// exception should be thrown
///
/// </summary>
/// <param name="r1">IndexReader to perform tests on
/// </param>
/// <throws> Exception </throws>
private void PerformDefaultTests(IndexReader r1)
{
float norm1 = Similarity.DecodeNorm(r1.Norms("field1")[4]);
IndexReader pr1Clone = (IndexReader)r1.Clone();
pr1Clone.DeleteDocument(10);
pr1Clone.SetNorm(4, "field1", 0.5f);
Assert.IsTrue(Similarity.DecodeNorm(r1.Norms("field1")[4]) == norm1);
Assert.IsTrue(Similarity.DecodeNorm(pr1Clone.Norms("field1")[4]) != norm1);
Assert.IsTrue(!r1.IsDeleted(10));
Assert.IsTrue(pr1Clone.IsDeleted(10));
// try to update the original reader, which should throw an exception
try
{
r1.DeleteDocument(11);
Assert.Fail("Tried to delete doc 11 and an exception should have been thrown");
}
catch (System.Exception exception)
{
// expectted
}
pr1Clone.Close();
}
public virtual void TestCloneWithSetNorm()
{
Directory dir1 = new MockRAMDirectory();
TestIndexReaderReopen.CreateIndex(dir1, false);
IndexReader orig = IndexReader.Open(dir1, false);
orig.SetNorm(1, "field1", 17.0f);
byte encoded = Similarity.EncodeNorm(17.0f);
Assert.AreEqual(encoded, orig.Norms("field1")[1]);
// the cloned segmentreader should have 2 references, 1 to itself, and 1 to
// the original segmentreader
IndexReader clonedReader = (IndexReader)orig.Clone();
orig.Close();
clonedReader.Close();
IndexReader r = IndexReader.Open(dir1, false);
Assert.AreEqual(encoded, r.Norms("field1")[1]);
r.Close();
dir1.Close();
}
public virtual void TestNormsRefCounting()
{
Directory dir1 = new MockRAMDirectory();
TestIndexReaderReopen.CreateIndex(dir1, false);
IndexReader reader1 = IndexReader.Open(dir1, false);
IndexReader reader2C = (IndexReader)reader1.Clone();
SegmentReader segmentReader2C = SegmentReader.GetOnlySegmentReader(reader2C);
segmentReader2C.Norms("field1"); // load the norms for the field
Norm reader2CNorm = segmentReader2C.norms_ForNUnit["field1"];
Assert.IsTrue(reader2CNorm.BytesRef().RefCount() == 2, "reader2CNorm.bytesRef()=" + reader2CNorm.BytesRef());
IndexReader reader3C = (IndexReader)reader2C.Clone();
SegmentReader segmentReader3C = SegmentReader.GetOnlySegmentReader(reader3C);
Norm reader3CCNorm = segmentReader3C.norms_ForNUnit["field1"];
Assert.AreEqual(3, reader3CCNorm.BytesRef().RefCount());
// edit a norm and the refcount should be 1
IndexReader reader4C = (IndexReader)reader3C.Clone();
SegmentReader segmentReader4C = SegmentReader.GetOnlySegmentReader(reader4C);
Assert.AreEqual(4, reader3CCNorm.BytesRef().RefCount());
reader4C.SetNorm(5, "field1", 0.33f);
// generate a cannot update exception in reader1
Assert.Throws <LockObtainFailedException>(() => reader3C.SetNorm(1, "field1", 0.99f), "did not hit expected exception");
// norm values should be different
Assert.IsTrue(Similarity.DecodeNorm(segmentReader3C.Norms("field1")[5]) != Similarity.DecodeNorm(segmentReader4C.Norms("field1")[5]));
Norm reader4CCNorm = segmentReader4C.norms_ForNUnit["field1"];
Assert.AreEqual(3, reader3CCNorm.BytesRef().RefCount());
Assert.AreEqual(1, reader4CCNorm.BytesRef().RefCount());
IndexReader reader5C = (IndexReader)reader4C.Clone();
SegmentReader segmentReader5C = SegmentReader.GetOnlySegmentReader(reader5C);
Norm reader5CCNorm = segmentReader5C.norms_ForNUnit["field1"];
reader5C.SetNorm(5, "field1", 0.7f);
Assert.AreEqual(1, reader5CCNorm.BytesRef().RefCount());
reader5C.Close();
reader4C.Close();
reader3C.Close();
reader2C.Close();
reader1.Close();
dir1.Close();
}
private void ModifyNormsForF1(Directory dir)
{
IndexReader ir = IndexReader.Open(dir, false);
int n = ir.MaxDoc;
for (int i = 0; i < n; i += 3)
{
// modify for every third doc
int k = (i * 3) % modifiedNorms.Count;
float origNorm = (float)modifiedNorms[i];
float newNorm = (float)modifiedNorms[k];
//System.out.println("Modifying: for "+i+" from "+origNorm+" to "+newNorm);
//System.out.println(" and: for "+k+" from "+newNorm+" to "+origNorm);
modifiedNorms[i] = newNorm;
modifiedNorms[k] = origNorm;
ir.SetNorm(i, "f" + 1, newNorm);
ir.SetNorm(k, "f" + 1, origNorm);
}
ir.Close();
}
/* Open pre-lockless index, add docs, do a delete &
* setNorm, and search */
public virtual void ChangeIndexNoAdds(System.String dirName)
{
dirName = FullDir(dirName);
Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo(dirName));
// make sure searching sees right # hits
IndexSearcher searcher = new IndexSearcher(dir, true);
ScoreDoc[] hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Assert.AreEqual(34, hits.Length, "wrong number of hits");
Document d = searcher.Doc(hits[0].Doc);
Assert.AreEqual("21", d.Get("id"), "wrong first document");
searcher.Close();
// make sure we can do a delete & setNorm against this
// pre-lockless segment:
IndexReader reader = IndexReader.Open(dir, false);
Term searchTerm = new Term("id", "6");
int delCount = reader.DeleteDocuments(searchTerm);
Assert.AreEqual(1, delCount, "wrong delete count");
reader.SetNorm(22, "content", (float)2.0);
reader.Close();
// make sure they "took":
searcher = new IndexSearcher(dir, true);
hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Assert.AreEqual(33, hits.Length, "wrong number of hits");
d = searcher.Doc(hits[0].Doc);
Assert.AreEqual("22", d.Get("id"), "wrong first document");
TestHits(hits, 33, searcher.IndexReader);
searcher.Close();
// optimize
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, IndexWriter.MaxFieldLength.UNLIMITED);
writer.Optimize();
writer.Close();
searcher = new IndexSearcher(dir, true);
hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Assert.AreEqual(33, hits.Length, "wrong number of hits");
d = searcher.Doc(hits[0].Doc);
Assert.AreEqual("22", d.Get("id"), "wrong first document");
TestHits(hits, 33, searcher.IndexReader);
searcher.Close();
dir.Close();
}
public virtual void TestSetNorm_Renamed()
{
RAMDirectory store = new RAMDirectory();
IndexWriter writer = new IndexWriter(store, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED, null);
// add the same document four times
IFieldable f1 = new Field("field", "word", Field.Store.YES, Field.Index.ANALYZED);
Document d1 = new Document();
d1.Add(f1);
writer.AddDocument(d1, null);
writer.AddDocument(d1, null);
writer.AddDocument(d1, null);
writer.AddDocument(d1, null);
writer.Close();
// reset the boost of each instance of this document
IndexReader reader = IndexReader.Open((Directory)store, false, null);
reader.SetNorm(0, "field", 1.0f, null);
reader.SetNorm(1, "field", 2.0f, null);
reader.SetNorm(2, "field", 4.0f, null);
reader.SetNorm(3, "field", 16.0f, null);
reader.Close();
// check that searches are ordered by this boost
float[] scores = new float[4];
new IndexSearcher(store, true, null).Search(new TermQuery(new Term("field", "word")), new AnonymousClassCollector(scores, this), null);
float lastScore = 0.0f;
for (int i = 0; i < 4; i++)
{
Assert.IsTrue(scores[i] > lastScore);
lastScore = scores[i];
}
}
private void ModifyNormsForF1(IndexReader ir)
{
int n = ir.MaxDoc();
// System.out.println("modifyNormsForF1 maxDoc: "+n);
for (int i = 0; i < n; i += 3)
{
// modify for every third doc
int k = (i * 3) % modifiedNorms.Count;
float origNorm = (float)((System.Single)modifiedNorms[i]);
float newNorm = (float)((System.Single)modifiedNorms[k]);
// System.out.println("Modifying: for "+i+" from "+origNorm+" to
// "+newNorm);
// System.out.println(" and: for "+k+" from "+newNorm+" to "+origNorm);
modifiedNorms[i] = (float)newNorm;
modifiedNorms[k] = (float)origNorm;
ir.SetNorm(i, "f" + 1, newNorm);
ir.SetNorm(k, "f" + 1, origNorm);
// System.out.println("setNorm i: "+i);
// break;
}
// ir.close();
}
/// <summary> 1. Get a norm from the original reader 2. Clone the original reader 3.
/// Delete a document and set the norm of the cloned reader 4. Verify the norms
/// are not the same on each reader 5. Verify the doc deleted is only in the
/// cloned reader 6. Try to delete a document in the original reader, an
/// exception should be thrown
///
/// </summary>
/// <param name="r1">IndexReader to perform tests on
/// </param>
/// <throws> Exception </throws>
private void PerformDefaultTests(IndexReader r1)
{
float norm1 = Similarity.DecodeNorm(r1.Norms("field1")[4]);
IndexReader pr1Clone = (IndexReader)r1.Clone();
pr1Clone.DeleteDocument(10);
pr1Clone.SetNorm(4, "field1", 0.5f);
Assert.IsTrue(Similarity.DecodeNorm(r1.Norms("field1")[4]) == norm1);
Assert.IsTrue(Similarity.DecodeNorm(pr1Clone.Norms("field1")[4]) != norm1);
Assert.IsTrue(!r1.IsDeleted(10));
Assert.IsTrue(pr1Clone.IsDeleted(10));
// try to update the original reader, which should throw an exception
Assert.Throws <LockObtainFailedException>(() => r1.DeleteDocument(11),
"Tried to delete doc 11 and an exception should have been thrown");
pr1Clone.Close();
}
public virtual void CreateIndex(System.String dirName, bool doCFS)
{
RmDir(dirName);
dirName = FullDir(dirName);
Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo(dirName));
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
writer.UseCompoundFile = doCFS;
writer.SetMaxBufferedDocs(10);
for (int i = 0; i < 35; i++)
{
AddDoc(writer, i);
}
Assert.AreEqual(35, writer.MaxDoc(), "wrong doc count");
writer.Close();
// open fresh writer so we get no prx file in the added segment
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), IndexWriter.MaxFieldLength.LIMITED);
writer.UseCompoundFile = doCFS;
writer.SetMaxBufferedDocs(10);
AddNoProxDoc(writer);
writer.Close();
// Delete one doc so we get a .del file:
IndexReader reader = IndexReader.Open(dir, false);
Term searchTerm = new Term("id", "7");
int delCount = reader.DeleteDocuments(searchTerm);
Assert.AreEqual(1, delCount, "didn't delete the right number of documents");
// Set one norm so we get a .s0 file:
reader.SetNorm(21, "content", (float)1.5);
reader.Close();
}
public virtual void TestKeepLastNDeletionPolicyWithReader()
{
int N = 10;
for (int pass = 0; pass < 2; pass++)
{
bool useCompoundFile = (pass % 2) != 0;
KeepLastNDeletionPolicy policy = new KeepLastNDeletionPolicy(this, N);
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, policy, IndexWriter.MaxFieldLength.UNLIMITED);
writer.UseCompoundFile = useCompoundFile;
writer.Close();
Term searchTerm = new Term("content", "aaa");
Query query = new TermQuery(searchTerm);
for (int i = 0; i < N + 1; i++)
{
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, policy, IndexWriter.MaxFieldLength.UNLIMITED);
writer.UseCompoundFile = useCompoundFile;
for (int j = 0; j < 17; j++)
{
AddDoc(writer);
}
// this is a commit
writer.Close();
IndexReader reader = IndexReader.Open(dir, policy, false);
reader.DeleteDocument(3 * i + 1);
reader.SetNorm(4 * i + 1, "content", 2.0F);
IndexSearcher searcher = new IndexSearcher(reader);
ScoreDoc[] hits = searcher.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(16 * (1 + i), hits.Length);
// this is a commit
reader.Close();
searcher.Close();
}
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, policy, IndexWriter.MaxFieldLength.UNLIMITED);
writer.UseCompoundFile = useCompoundFile;
writer.Optimize();
// this is a commit
writer.Close();
Assert.AreEqual(2 * (N + 2), policy.numOnInit);
Assert.AreEqual(2 * (N + 2) - 1, policy.numOnCommit);
IndexSearcher searcher2 = new IndexSearcher(dir, false);
ScoreDoc[] hits2 = searcher2.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(176, hits2.Length);
// Simplistic check: just verify only the past N segments_N's still
// exist, and, I can open a reader on each:
long gen = SegmentInfos.GetCurrentSegmentGeneration(dir);
dir.DeleteFile(IndexFileNames.SEGMENTS_GEN);
int expectedCount = 176;
for (int i = 0; i < N + 1; i++)
{
try
{
IndexReader reader = IndexReader.Open(dir, true);
// Work backwards in commits on what the expected
// count should be.
searcher2 = new IndexSearcher(reader);
hits2 = searcher2.Search(query, null, 1000).ScoreDocs;
if (i > 1)
{
if (i % 2 == 0)
{
expectedCount += 1;
}
else
{
expectedCount -= 17;
}
}
Assert.AreEqual(expectedCount, hits2.Length);
searcher2.Close();
reader.Close();
if (i == N)
{
Assert.Fail("should have failed on commits before last 5");
}
}
catch (System.IO.IOException e)
{
if (i != N)
{
throw e;
}
}
if (i < N)
{
dir.DeleteFile(IndexFileNames.FileNameFromGeneration(IndexFileNames.SEGMENTS, "", gen));
}
gen--;
}
dir.Close();
}
}
public virtual void TestKeepLastNDeletionPolicyWithCreates()
{
int N = 10;
for (int pass = 0; pass < 4; pass++)
{
bool autoCommit = pass < 2;
bool useCompoundFile = (pass % 2) > 0;
KeepLastNDeletionPolicy policy = new KeepLastNDeletionPolicy(this, N);
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, autoCommit, new WhitespaceAnalyzer(), true, policy);
writer.SetMaxBufferedDocs(10);
writer.SetUseCompoundFile(useCompoundFile);
writer.Close();
Term searchTerm = new Term("content", "aaa");
Query query = new TermQuery(searchTerm);
for (int i = 0; i < N + 1; i++)
{
writer = new IndexWriter(dir, autoCommit, new WhitespaceAnalyzer(), false, policy);
writer.SetMaxBufferedDocs(10);
writer.SetUseCompoundFile(useCompoundFile);
for (int j = 0; j < 17; j++)
{
AddDoc(writer);
}
// this is a commit when autoCommit=false:
writer.Close();
IndexReader reader = IndexReader.Open(dir, policy);
reader.DeleteDocument(3);
reader.SetNorm(5, "content", 2.0F);
IndexSearcher searcher = new IndexSearcher(reader);
ScoreDoc[] hits = searcher.Search(query, null, 1000).scoreDocs;
Assert.AreEqual(16, hits.Length);
// this is a commit when autoCommit=false:
reader.Close();
searcher.Close();
writer = new IndexWriter(dir, autoCommit, new WhitespaceAnalyzer(), true, policy);
// This will not commit: there are no changes
// pending because we opened for "create":
writer.Close();
}
Assert.AreEqual(1 + 3 * (N + 1), policy.numOnInit);
if (!autoCommit)
{
Assert.AreEqual(3 * (N + 1), policy.numOnCommit);
}
IndexSearcher searcher2 = new IndexSearcher(dir);
ScoreDoc[] hits2 = searcher2.Search(query, null, 1000).scoreDocs;
Assert.AreEqual(0, hits2.Length);
// Simplistic check: just verify only the past N segments_N's still
// exist, and, I can open a reader on each:
long gen = SegmentInfos.GetCurrentSegmentGeneration(dir);
dir.DeleteFile(IndexFileNames.SEGMENTS_GEN);
int expectedCount = 0;
for (int i = 0; i < N + 1; i++)
{
try
{
IndexReader reader = IndexReader.Open(dir);
// Work backwards in commits on what the expected
// count should be. Only check this in the
// autoCommit false case:
if (!autoCommit)
{
searcher2 = new IndexSearcher(reader);
hits2 = searcher2.Search(query, null, 1000).scoreDocs;
Assert.AreEqual(expectedCount, hits2.Length);
searcher2.Close();
if (expectedCount == 0)
{
expectedCount = 16;
}
else if (expectedCount == 16)
{
expectedCount = 17;
}
else if (expectedCount == 17)
{
expectedCount = 0;
}
}
reader.Close();
if (i == N)
{
Assert.Fail("should have failed on commits before last " + N);
}
}
catch (System.IO.IOException e)
{
if (i != N)
{
throw e;
}
}
if (i < N)
{
dir.DeleteFile(IndexFileNames.FileNameFromGeneration(IndexFileNames.SEGMENTS, "", gen));
}
gen--;
}
dir.Close();
}
}
public virtual void TestExactFileNames()
{
System.String outputDir = "lucene.backwardscompat0.index";
RmDir(outputDir);
try
{
Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo(FullDir(outputDir)));
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true,
IndexWriter.MaxFieldLength.UNLIMITED);
writer.SetRAMBufferSizeMB(16.0);
for (int i = 0; i < 35; i++)
{
AddDoc(writer, i);
}
Assert.AreEqual(35, writer.MaxDoc(), "wrong doc count");
writer.Close();
// Delete one doc so we get a .del file:
IndexReader reader = IndexReader.Open(dir, false);
Term searchTerm = new Term("id", "7");
int delCount = reader.DeleteDocuments(searchTerm);
Assert.AreEqual(1, delCount, "didn't delete the right number of documents");
// Set one norm so we get a .s0 file:
reader.SetNorm(21, "content", (float)1.5);
reader.Close();
// The numbering of fields can vary depending on which
// JRE is in use. On some JREs we see content bound to
// field 0; on others, field 1. So, here we have to
// figure out which field number corresponds to
// "content", and then set our expected file names below
// accordingly:
CompoundFileReader cfsReader = new CompoundFileReader(dir, "_0.cfs");
FieldInfos fieldInfos = new FieldInfos(cfsReader, "_0.fnm");
int contentFieldIndex = -1;
for (int i = 0; i < fieldInfos.Size(); i++)
{
FieldInfo fi = fieldInfos.FieldInfo(i);
if (fi.name_ForNUnit.Equals("content"))
{
contentFieldIndex = i;
break;
}
}
cfsReader.Close();
Assert.IsTrue(contentFieldIndex != -1,
"could not locate the 'content' field number in the _2.cfs segment");
// Now verify file names:
System.String[] expected;
expected = new System.String[]
{ "_0.cfs", "_0_1.del", "_0_1.s" + contentFieldIndex, "segments_3", "segments.gen" };
System.String[] actual = dir.ListAll();
System.Array.Sort(expected);
System.Array.Sort(actual);
if (!CollectionsHelper.Equals(expected, actual))
{
Assert.Fail("incorrect filenames in index: expected:\n " + AsString(expected) +
"\n actual:\n " + AsString(actual));
}
dir.Close();
}
finally
{
RmDir(outputDir);
}
}
protected internal override void DoSetNorm(int d, System.String f, byte b, IState state)
{
in_Renamed.SetNorm(d, f, b, state);
}
/* Open pre-lockless index, add docs, do a delete &
* setNorm, and search */
public virtual void ChangeIndexWithAdds(System.String dirName)
{
System.String origDirName = dirName;
dirName = FullDir(dirName);
Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo(dirName));
// open writer
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, IndexWriter.MaxFieldLength.UNLIMITED);
// add 10 docs
for (int i = 0; i < 10; i++)
{
AddDoc(writer, 35 + i);
}
// make sure writer sees right total -- writer seems not to know about deletes in .del?
int expected;
if (Compare(origDirName, "24") < 0)
{
expected = 45;
}
else
{
expected = 46;
}
Assert.AreEqual(expected, writer.MaxDoc(), "wrong doc count");
writer.Close();
// make sure searching sees right # hits
IndexSearcher searcher = new IndexSearcher(dir, true);
ScoreDoc[] hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Document d = searcher.Doc(hits[0].Doc);
Assert.AreEqual("21", d.Get("id"), "wrong first document");
TestHits(hits, 44, searcher.IndexReader);
searcher.Close();
// make sure we can do delete & setNorm against this
// pre-lockless segment:
IndexReader reader = IndexReader.Open(dir, false);
Term searchTerm = new Term("id", "6");
int delCount = reader.DeleteDocuments(searchTerm);
Assert.AreEqual(1, delCount, "wrong delete count");
reader.SetNorm(22, "content", (float)2.0);
reader.Close();
// make sure they "took":
searcher = new IndexSearcher(dir, true);
hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Assert.AreEqual(43, hits.Length, "wrong number of hits");
d = searcher.Doc(hits[0].Doc);
Assert.AreEqual("22", d.Get("id"), "wrong first document");
TestHits(hits, 43, searcher.IndexReader);
searcher.Close();
// optimize
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, IndexWriter.MaxFieldLength.UNLIMITED);
writer.Optimize();
writer.Close();
searcher = new IndexSearcher(dir, true);
hits = searcher.Search(new TermQuery(new Term("content", "aaa")), null, 1000).ScoreDocs;
Assert.AreEqual(43, hits.Length, "wrong number of hits");
d = searcher.Doc(hits[0].Doc);
TestHits(hits, 43, searcher.IndexReader);
Assert.AreEqual("22", d.Get("id"), "wrong first document");
searcher.Close();
dir.Close();
}
public virtual void TestDeleteLeftoverFiles()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
writer.SetMaxBufferedDocs(10);
int i;
for (i = 0; i < 35; i++)
{
AddDoc(writer, i);
}
writer.SetUseCompoundFile(false);
for (; i < 45; i++)
{
AddDoc(writer, i);
}
writer.Close();
// Delete one doc so we get a .del file:
IndexReader reader = IndexReader.Open(dir);
Term searchTerm = new Term("id", "7");
int delCount = reader.DeleteDocuments(searchTerm);
Assert.AreEqual(1, delCount, "didn't delete the right number of documents");
// Set one norm so we get a .s0 file:
reader.SetNorm(21, "content", (float)1.5);
reader.Close();
// Now, artificially create an extra .del file & extra
// .s0 file:
System.String[] files = dir.ListAll();
/*
* for(int j=0;j<files.length;j++) {
* System.out.println(j + ": " + files[j]);
* }
*/
// The numbering of fields can vary depending on which
// JRE is in use. On some JREs we see content bound to
// field 0; on others, field 1. So, here we have to
// figure out which field number corresponds to
// "content", and then set our expected file names below
// accordingly:
CompoundFileReader cfsReader = new CompoundFileReader(dir, "_2.cfs");
FieldInfos fieldInfos = new FieldInfos(cfsReader, "_2.fnm");
int contentFieldIndex = -1;
for (i = 0; i < fieldInfos.Size(); i++)
{
FieldInfo fi = fieldInfos.FieldInfo(i);
if (fi.name_ForNUnit.Equals("content"))
{
contentFieldIndex = i;
break;
}
}
cfsReader.Close();
Assert.IsTrue(contentFieldIndex != -1, "could not locate the 'content' field number in the _2.cfs segment");
System.String normSuffix = "s" + contentFieldIndex;
// Create a bogus separate norms file for a
// segment/field that actually has a separate norms file
// already:
CopyFile(dir, "_2_1." + normSuffix, "_2_2." + normSuffix);
// Create a bogus separate norms file for a
// segment/field that actually has a separate norms file
// already, using the "not compound file" extension:
CopyFile(dir, "_2_1." + normSuffix, "_2_2.f" + contentFieldIndex);
// Create a bogus separate norms file for a
// segment/field that does not have a separate norms
// file already:
CopyFile(dir, "_2_1." + normSuffix, "_1_1." + normSuffix);
// Create a bogus separate norms file for a
// segment/field that does not have a separate norms
// file already using the "not compound file" extension:
CopyFile(dir, "_2_1." + normSuffix, "_1_1.f" + contentFieldIndex);
// Create a bogus separate del file for a
// segment that already has a separate del file:
CopyFile(dir, "_0_1.del", "_0_2.del");
// Create a bogus separate del file for a
// segment that does not yet have a separate del file:
CopyFile(dir, "_0_1.del", "_1_1.del");
// Create a bogus separate del file for a
// non-existent segment:
CopyFile(dir, "_0_1.del", "_188_1.del");
// Create a bogus segment file:
CopyFile(dir, "_0.cfs", "_188.cfs");
// Create a bogus fnm file when the CFS already exists:
CopyFile(dir, "_0.cfs", "_0.fnm");
// Create a deletable file:
CopyFile(dir, "_0.cfs", "deletable");
// Create some old segments file:
CopyFile(dir, "segments_3", "segments");
CopyFile(dir, "segments_3", "segments_2");
// Create a bogus cfs file shadowing a non-cfs segment:
CopyFile(dir, "_2.cfs", "_3.cfs");
System.String[] filesPre = dir.ListAll();
// Open & close a writer: it should delete the above 4
// files and nothing more:
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, IndexWriter.MaxFieldLength.LIMITED);
writer.Close();
System.String[] files2 = dir.ListAll();
dir.Close();
System.Array.Sort(files);
System.Array.Sort(files2);
System.Collections.Hashtable dif = DifFiles(files, files2);
if (!SupportClass.CollectionsHelper.Equals(files, files2))
{
Assert.Fail("IndexFileDeleter failed to delete unreferenced extra files: should have deleted " + (filesPre.Length - files.Length) + " files but only deleted " + (filesPre.Length - files2.Length) + "; expected files:\n " + AsString(files) + "\n actual files:\n " + AsString(files2) + "\ndif: " + SupportClass.CollectionsHelper.CollectionToString(dif));
}
}
protected internal override void DoSetNorm(int d, string f, byte b)
{
in_Renamed.SetNorm(d, f, b);
}
private void ModifyNormsForF1(IndexReader ir)
{
int n = ir.MaxDoc;
// System.out.println("modifyNormsForF1 maxDoc: "+n);
for (int i = 0; i < n; i += 3)
{
// modify for every third doc
int k = (i * 3) % modifiedNorms.Count;
float origNorm = (float) (modifiedNorms[i]);
float newNorm = (float) (modifiedNorms[k]);
// System.out.println("Modifying: for "+i+" from "+origNorm+" to
// "+newNorm);
// System.out.println(" and: for "+k+" from "+newNorm+" to "+origNorm);
modifiedNorms[i] = newNorm;
modifiedNorms[k] = origNorm;
ir.SetNorm(i, "f" + 1, newNorm);
ir.SetNorm(k, "f" + 1, origNorm);
// System.out.println("setNorm i: "+i);
// break;
}
// ir.close();
}
public virtual void TestQuery()
{
RAMDirectory dir = new RAMDirectory();
IndexWriter iw = new IndexWriter(dir, analyzer, true, IndexWriter.MaxFieldLength.LIMITED, null);
iw.SetMaxBufferedDocs(2); // force multi-segment
AddDoc("one", iw, 1f);
AddDoc("two", iw, 20f);
AddDoc("three four", iw, 300f);
iw.Close();
IndexReader ir = IndexReader.Open((Directory)dir, false, null);
IndexSearcher is_Renamed = new IndexSearcher(ir);
ScoreDoc[] hits;
// assert with norms scoring turned off
hits = is_Renamed.Search(new MatchAllDocsQuery(), null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits.Length);
Assert.AreEqual(ir.Document(hits[0].Doc, null).Get("key", null), "one");
Assert.AreEqual(ir.Document(hits[1].Doc, null).Get("key", null), "two");
Assert.AreEqual(ir.Document(hits[2].Doc, null).Get("key", null), "three four");
// assert with norms scoring turned on
MatchAllDocsQuery normsQuery = new MatchAllDocsQuery("key");
hits = is_Renamed.Search(normsQuery, null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits.Length);
Assert.AreEqual(ir.Document(hits[0].Doc, null).Get("key", null), "three four");
Assert.AreEqual(ir.Document(hits[1].Doc, null).Get("key", null), "two");
Assert.AreEqual(ir.Document(hits[2].Doc, null).Get("key", null), "one");
// change norm & retest
ir.SetNorm(0, "key", 400f, null);
normsQuery = new MatchAllDocsQuery("key");
hits = is_Renamed.Search(normsQuery, null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits.Length);
Assert.AreEqual(ir.Document(hits[0].Doc, null).Get("key", null), "one");
Assert.AreEqual(ir.Document(hits[1].Doc, null).Get("key", null), "three four");
Assert.AreEqual(ir.Document(hits[2].Doc, null).Get("key", null), "two");
// some artificial queries to trigger the use of skipTo():
BooleanQuery bq = new BooleanQuery();
bq.Add(new MatchAllDocsQuery(), Occur.MUST);
bq.Add(new MatchAllDocsQuery(), Occur.MUST);
hits = is_Renamed.Search(bq, null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits.Length);
bq = new BooleanQuery();
bq.Add(new MatchAllDocsQuery(), Occur.MUST);
bq.Add(new TermQuery(new Term("key", "three")), Occur.MUST);
hits = is_Renamed.Search(bq, null, 1000, null).ScoreDocs;
Assert.AreEqual(1, hits.Length);
// delete a document:
is_Renamed.IndexReader.DeleteDocument(0, null);
hits = is_Renamed.Search(new MatchAllDocsQuery(), null, 1000, null).ScoreDocs;
Assert.AreEqual(2, hits.Length);
// test parsable toString()
QueryParser qp = new QueryParser(Util.Version.LUCENE_CURRENT, "key", analyzer);
hits = is_Renamed.Search(qp.Parse(new MatchAllDocsQuery().ToString()), null, 1000, null).ScoreDocs;
Assert.AreEqual(2, hits.Length);
// test parsable toString() with non default boost
Query maq = new MatchAllDocsQuery();
maq.Boost = 2.3f;
Query pq = qp.Parse(maq.ToString());
hits = is_Renamed.Search(pq, null, 1000, null).ScoreDocs;
Assert.AreEqual(2, hits.Length);
is_Renamed.Close();
ir.Close();
dir.Close();
}
