| public DeleteDocuments ( Lucene.Net.Index.Term term ) : int | ||
| term | Lucene.Net.Index.Term | |
| return | int | |
/// <summary> Deletes all documents containing <code>term</code>.
/// This is useful if one uses a document field to hold a unique ID string for
/// the document. Then to delete such a document, one merely constructs a
/// term with the appropriate field and the unique ID string as its text and
/// passes it to this method. Returns the number of documents deleted.
/// </summary>
/// <returns> the number of documents deleted
/// </returns>
/// <seealso cref="IndexReader#DeleteDocuments(Term)">
/// </seealso>
/// <throws> IllegalStateException if the index is closed </throws>
public virtual int DeleteDocuments(Term term)
{
lock (directory)
{
AssureOpen();
CreateIndexReader();
return(indexReader.DeleteDocuments(term));
}
}
/* 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 TestMergeDocCount0()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.UNLIMITED, null);
writer.SetMergePolicy(new LogDocMergePolicy(writer));
writer.SetMaxBufferedDocs(10);
writer.MergeFactor = 100;
for (int i = 0; i < 250; i++)
{
AddDoc(writer);
CheckInvariants(writer);
}
writer.Close();
IndexReader reader = IndexReader.Open(dir, false, null);
reader.DeleteDocuments(new Term("content", "aaa"), null);
reader.Close();
writer = new IndexWriter(dir, new WhitespaceAnalyzer(), false, IndexWriter.MaxFieldLength.UNLIMITED, null);
writer.SetMergePolicy(new LogDocMergePolicy(writer));
writer.SetMaxBufferedDocs(10);
writer.MergeFactor = 5;
// merge factor is changed, so check invariants after all adds
for (int i = 0; i < 10; i++)
{
AddDoc(writer);
}
writer.Commit(null);
((ConcurrentMergeScheduler)writer.MergeScheduler).Sync();
writer.Commit(null);
CheckInvariants(writer);
Assert.AreEqual(10, writer.MaxDoc());
writer.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 TestMergeDocCount0()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, true, new WhitespaceAnalyzer(), true);
writer.SetMergePolicy(new LogDocMergePolicy(writer));
writer.SetMaxBufferedDocs(10);
writer.SetMergeFactor(100);
for (int i = 0; i < 250; i++)
{
AddDoc(writer);
CheckInvariants(writer);
}
writer.Close();
IndexReader reader = IndexReader.Open(dir);
reader.DeleteDocuments(new Term("content", "aaa"));
reader.Close();
writer = new IndexWriter(dir, true, new WhitespaceAnalyzer(), false);
writer.SetMergePolicy(new LogDocMergePolicy(writer));
writer.SetMaxBufferedDocs(10);
writer.SetMergeFactor(5);
// merge factor is changed, so check invariants after all adds
for (int i = 0; i < 10; i++)
{
AddDoc(writer);
}
CheckInvariants(writer);
Assert.AreEqual(10, writer.DocCount());
writer.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);
}
}
/* 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));
}
}
// Apply buffered delete terms to this reader.
private void ApplyDeletes(System.Collections.Hashtable deleteTerms, IndexReader reader)
{
System.Collections.IEnumerator iter = new System.Collections.Hashtable(deleteTerms).GetEnumerator();
while (iter.MoveNext())
{
System.Collections.DictionaryEntry entry = (System.Collections.DictionaryEntry) iter.Current;
reader.DeleteDocuments((Term) entry.Key);
}
}
public virtual void TestEmptyIndex()
{
// creating two directories for indices
Directory indexStoreA = new MockRAMDirectory();
Directory indexStoreB = new MockRAMDirectory();
// creating a document to store
Document lDoc = new Document();
lDoc.Add(new Field("fulltext", "Once upon a time.....", Field.Store.YES, Field.Index.ANALYZED));
lDoc.Add(new Field("id", "doc1", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating a document to store
Document lDoc2 = new Document();
lDoc2.Add(new Field("fulltext", "in a galaxy far far away.....", Field.Store.YES, Field.Index.ANALYZED));
lDoc2.Add(new Field("id", "doc2", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc2.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating a document to store
Document lDoc3 = new Document();
lDoc3.Add(new Field("fulltext", "a bizarre bug manifested itself....", Field.Store.YES, Field.Index.ANALYZED));
lDoc3.Add(new Field("id", "doc3", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc3.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating an index writer for the first index
IndexWriter writerA = new IndexWriter(indexStoreA, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), true, IndexWriter.MaxFieldLength.LIMITED, null);
// creating an index writer for the second index, but writing nothing
IndexWriter writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), true, IndexWriter.MaxFieldLength.LIMITED, null);
//--------------------------------------------------------------------
// scenario 1
//--------------------------------------------------------------------
// writing the documents to the first index
writerA.AddDocument(lDoc, null);
writerA.AddDocument(lDoc2, null);
writerA.AddDocument(lDoc3, null);
writerA.Optimize(null);
writerA.Close();
// closing the second index
writerB.Close();
// creating the query
QueryParser parser = new QueryParser(Util.Version.LUCENE_CURRENT, "fulltext", new StandardAnalyzer(Util.Version.LUCENE_CURRENT));
Query query = parser.Parse("handle:1");
// building the searchables
Searcher[] searchers = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers[0] = new IndexSearcher(indexStoreB, true, null);
searchers[1] = new IndexSearcher(indexStoreA, true, null);
// creating the multiSearcher
Searcher mSearcher = GetMultiSearcherInstance(searchers);
// performing the search
ScoreDoc[] hits = mSearcher.Search(query, null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits.Length);
// iterating over the hit documents
for (int i = 0; i < hits.Length; i++)
{
mSearcher.Doc(hits[i].Doc, null);
}
mSearcher.Close();
//--------------------------------------------------------------------
// scenario 2
//--------------------------------------------------------------------
// adding one document to the empty index
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), false, IndexWriter.MaxFieldLength.LIMITED, null);
writerB.AddDocument(lDoc, null);
writerB.Optimize(null);
writerB.Close();
// building the searchables
Searcher[] searchers2 = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers2[0] = new IndexSearcher(indexStoreB, true, null);
searchers2[1] = new IndexSearcher(indexStoreA, true, null);
// creating the mulitSearcher
MultiSearcher mSearcher2 = GetMultiSearcherInstance(searchers2);
// performing the same search
ScoreDoc[] hits2 = mSearcher2.Search(query, null, 1000, null).ScoreDocs;
Assert.AreEqual(4, hits2.Length);
// iterating over the hit documents
for (int i = 0; i < hits2.Length; i++)
{
// no exception should happen at this point
mSearcher2.Doc(hits2[i].Doc, null);
}
// test the subSearcher() method:
Query subSearcherQuery = parser.Parse("id:doc1");
hits2 = mSearcher2.Search(subSearcherQuery, null, 1000, null).ScoreDocs;
Assert.AreEqual(2, hits2.Length);
Assert.AreEqual(0, mSearcher2.SubSearcher(hits2[0].Doc)); // hit from searchers2[0]
Assert.AreEqual(1, mSearcher2.SubSearcher(hits2[1].Doc)); // hit from searchers2[1]
subSearcherQuery = parser.Parse("id:doc2");
hits2 = mSearcher2.Search(subSearcherQuery, null, 1000, null).ScoreDocs;
Assert.AreEqual(1, hits2.Length);
Assert.AreEqual(1, mSearcher2.SubSearcher(hits2[0].Doc)); // hit from searchers2[1]
mSearcher2.Close();
//--------------------------------------------------------------------
// scenario 3
//--------------------------------------------------------------------
// deleting the document just added, this will cause a different exception to take place
Term term = new Term("id", "doc1");
IndexReader readerB = IndexReader.Open(indexStoreB, false, null);
readerB.DeleteDocuments(term, null);
readerB.Close();
// optimizing the index with the writer
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), false, IndexWriter.MaxFieldLength.LIMITED, null);
writerB.Optimize(null);
writerB.Close();
// building the searchables
Searcher[] searchers3 = new Searcher[2];
searchers3[0] = new IndexSearcher(indexStoreB, true, null);
searchers3[1] = new IndexSearcher(indexStoreA, true, null);
// creating the mulitSearcher
Searcher mSearcher3 = GetMultiSearcherInstance(searchers3);
// performing the same search
ScoreDoc[] hits3 = mSearcher3.Search(query, null, 1000, null).ScoreDocs;
Assert.AreEqual(3, hits3.Length);
// iterating over the hit documents
for (int i = 0; i < hits3.Length; i++)
{
mSearcher3.Doc(hits3[i].Doc, null);
}
mSearcher3.Close();
indexStoreA.Close();
indexStoreB.Close();
}