C# IndexWriter.SetMergeFactor方法代码示例

        public void ApplyToWriter(IndexWriter writer)
        {
            try
            {
                if (MergeFactor != null)
                {
                    writer.SetMergeFactor((int) MergeFactor);
                }

                if (MaxMergeDocs != null)
                {
                    writer.SetMaxMergeDocs((int) MaxMergeDocs);
                }

                if (MaxBufferedDocs != null)
                {
                    writer.SetMaxBufferedDocs((int) MaxBufferedDocs);
                }

                if (RamBufferSizeMb != null)
                {
                    writer.SetRAMBufferSizeMB((int) RamBufferSizeMb);
                }

                if (TermIndexInterval != null)
                {
                    writer.SetTermIndexInterval((int) TermIndexInterval);
                }
            }
            catch (ArgumentOutOfRangeException)
            {
                // TODO: Log it
            }
        }

		public virtual void  TestSorting()
		{
			Directory directory = new MockRAMDirectory();
			IndexWriter writer = new IndexWriter(directory, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
			writer.SetMaxBufferedDocs(2);
			writer.SetMergeFactor(1000);
			writer.AddDocument(Adoc(new System.String[]{"id", "a", "title", "ipod", "str_s", "a"}));
			writer.AddDocument(Adoc(new System.String[]{"id", "b", "title", "ipod ipod", "str_s", "b"}));
			writer.AddDocument(Adoc(new System.String[]{"id", "c", "title", "ipod ipod ipod", "str_s", "c"}));
			writer.AddDocument(Adoc(new System.String[]{"id", "x", "title", "boosted", "str_s", "x"}));
			writer.AddDocument(Adoc(new System.String[]{"id", "y", "title", "boosted boosted", "str_s", "y"}));
			writer.AddDocument(Adoc(new System.String[]{"id", "z", "title", "boosted boosted boosted", "str_s", "z"}));
			
			IndexReader r = writer.GetReader();
			writer.Close();
			
			IndexSearcher searcher = new IndexSearcher(r);
			
			RunTest(searcher, true);
			RunTest(searcher, false);
			
			searcher.Close();
			r.Close();
			directory.Close();
		}

		public virtual void  TestLucene()
		{
			
			int num = 100;
			
			Directory indexA = new MockRAMDirectory();
			Directory indexB = new MockRAMDirectory();
			
			FillIndex(indexA, 0, num);
			bool fail = VerifyIndex(indexA, 0);
			if (fail)
			{
				Assert.Fail("Index a is invalid");
			}
			
			FillIndex(indexB, num, num);
			fail = VerifyIndex(indexB, num);
			if (fail)
			{
				Assert.Fail("Index b is invalid");
			}
			
			Directory merged = new MockRAMDirectory();
			
			IndexWriter writer = new IndexWriter(merged, new StandardAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
			writer.SetMergeFactor(2);
			
			writer.AddIndexes(new Directory[]{indexA, indexB});
			writer.Close();
			
			fail = VerifyIndex(merged, 0);
			merged.Close();
			
			Assert.IsFalse(fail, "The merged index is invalid");
		}

        public static IndexWriter GetAzureIndexWriter(this LuceneIndexer indexer)
        {
            indexer.EnsureIndex(false);
            var writer = new IndexWriter(indexer.GetLuceneDirectory(), indexer.IndexingAnalyzer, false, IndexWriter.MaxFieldLength.UNLIMITED);

            writer.SetRAMBufferSizeMB(10.0);
            writer.SetUseCompoundFile(false);
            writer.SetMaxMergeDocs(10000);
            writer.SetMergeFactor(100);
            return writer;
        }

		public virtual void  TestNormalCase()
		{
			Directory dir = new RAMDirectory();
			
			IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
			writer.SetMaxBufferedDocs(10);
			writer.SetMergeFactor(10);
			writer.SetMergePolicy(new LogDocMergePolicy(writer));
			
			for (int i = 0; i < 100; i++)
			{
				AddDoc(writer);
				CheckInvariants(writer);
			}
			
			writer.Close();
		}

		public virtual void  TestIndexing()
		{
			Directory mainDir = new MockRAMDirectory();
			IndexWriter writer = new IndexWriter(mainDir, new WhitespaceAnalyzer(), IndexWriter.MaxFieldLength.LIMITED);
			writer.SetUseCompoundFile(false);
			IndexReader reader = writer.GetReader(); // start pooling readers
			reader.Close();
			writer.SetMergeFactor(2);
			writer.SetMaxBufferedDocs(10);
			RunThread[] indexThreads = new RunThread[4];
			for (int x = 0; x < indexThreads.Length; x++)
			{
				indexThreads[x] = new RunThread(this, x % 2, writer);
				indexThreads[x].Name = "Thread " + x;
				indexThreads[x].Start();
			}
			long startTime = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond);
			long duration = 5 * 1000;
			while (((DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond) - startTime) < duration)
			{
				System.Threading.Thread.Sleep(new System.TimeSpan((System.Int64) 10000 * 100));
			}
			int delCount = 0;
			int addCount = 0;
			for (int x = 0; x < indexThreads.Length; x++)
			{
				indexThreads[x].run_Renamed_Field = false;
				Assert.IsTrue(indexThreads[x].ex == null);
				addCount += indexThreads[x].addCount;
				delCount += indexThreads[x].delCount;
			}
			for (int x = 0; x < indexThreads.Length; x++)
			{
				indexThreads[x].Join();
			}
			//System.out.println("addCount:"+addCount);
			//System.out.println("delCount:"+delCount);
			writer.Close();
			mainDir.Close();
		}

		public virtual void  TestNoOverMerge()
		{
			Directory dir = new RAMDirectory();
			
			IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
			writer.SetMaxBufferedDocs(10);
			writer.SetMergeFactor(10);
			writer.SetMergePolicy(new LogDocMergePolicy(writer));
			
			bool noOverMerge = false;
			for (int i = 0; i < 100; i++)
			{
				AddDoc(writer);
				CheckInvariants(writer);
				if (writer.GetNumBufferedDocuments() + writer.GetSegmentCount() >= 18)
				{
					noOverMerge = true;
				}
			}
			Assert.IsTrue(noOverMerge);
			
			writer.Close();
		}

		public virtual void  TestOptimizeMaxNumSegments()
		{
			
			MockRAMDirectory dir = new MockRAMDirectory();
			
			Document doc = new Document();
			doc.Add(new Field("content", "aaa", Field.Store.YES, Field.Index.TOKENIZED));
			
			for (int numDocs = 38; numDocs < 500; numDocs += 38)
			{
				IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true);
				LogDocMergePolicy ldmp = new LogDocMergePolicy();
				ldmp.SetMinMergeDocs(1);
				writer.SetMergePolicy(ldmp);
				writer.SetMergeFactor(5);
				writer.SetMaxBufferedDocs(2);
				for (int j = 0; j < numDocs; j++)
					writer.AddDocument(doc);
				writer.Close();
				
				SegmentInfos sis = new SegmentInfos();
				sis.Read(dir);
				int segCount = sis.Count;
				
				writer = new IndexWriter(dir, new WhitespaceAnalyzer());
				writer.SetMergePolicy(ldmp);
				writer.SetMergeFactor(5);
				writer.Optimize(3);
				writer.Close();
				
				sis = new SegmentInfos();
				sis.Read(dir);
				int optSegCount = sis.Count;
				
				if (segCount < 3)
					Assert.AreEqual(segCount, optSegCount);
				else
					Assert.AreEqual(3, optSegCount);
			}
		}

		public virtual void  TestOptimizeMaxNumSegments2()
		{
			MockRAMDirectory dir = new MockRAMDirectory();
			
			Document doc = new Document();
			doc.Add(new Field("content", "aaa", Field.Store.YES, Field.Index.TOKENIZED));
			
			IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer(), true);
			LogDocMergePolicy ldmp = new LogDocMergePolicy();
			ldmp.SetMinMergeDocs(1);
			writer.SetMergePolicy(ldmp);
			writer.SetMergeFactor(4);
			writer.SetMaxBufferedDocs(2);
			
			for (int iter = 0; iter < 10; iter++)
			{
				
				for (int i = 0; i < 19; i++)
					writer.AddDocument(doc);
				
				writer.Flush();
				
				SegmentInfos sis = new SegmentInfos();
				((ConcurrentMergeScheduler) writer.GetMergeScheduler()).Sync();
				sis.Read(dir);
				
				int segCount = sis.Count;
				
				writer.Optimize(7);
				
				sis = new SegmentInfos();
				((ConcurrentMergeScheduler) writer.GetMergeScheduler()).Sync();
				sis.Read(dir);
				int optSegCount = sis.Count;
				
				if (segCount < 7)
					Assert.AreEqual(segCount, optSegCount);
				else
					Assert.AreEqual(7, optSegCount);
			}
		}

		// Runs test, with multiple threads, using the specific
		// failure to trigger an IOException
		public virtual void  _testMultipleThreadsFailure(MockRAMDirectory.Failure failure)
		{
			
			int NUM_THREADS = 3;
			
			for (int iter = 0; iter < 5; iter++)
			{
				MockRAMDirectory dir = new MockRAMDirectory();
				IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer());
				ConcurrentMergeScheduler cms = new ConcurrentMergeScheduler();
				// We expect disk full exceptions in the merge threads
				cms.SetSuppressExceptions_ForNUnitTest();
				writer.SetMergeScheduler(cms);
				writer.SetMaxBufferedDocs(2);
				writer.SetMergeFactor(4);
				
				IndexerThread[] threads = new IndexerThread[NUM_THREADS];
				//bool diskFull = false;
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i] = new IndexerThread(this, writer, true);
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i].Start();
				
				try
				{
					System.Threading.Thread.Sleep(new System.TimeSpan((System.Int64) 10000 * 10));
				}
				catch (System.Threading.ThreadInterruptedException)
				{
					SupportClass.ThreadClass.Current().Interrupt();
				}
				
				dir.FailOn(failure);
				failure.SetDoFail();
				
				for (int i = 0; i < NUM_THREADS; i++)
				{
					while (true)
					{
						try
						{
							threads[i].Join();
							break;
						}
						catch (System.Threading.ThreadInterruptedException)
						{
							SupportClass.ThreadClass.Current().Interrupt();
						}
					}
					if (threads[i].IsAlive)
						Assert.Fail("thread seems to be hung");
					else
						Assert.IsTrue(threads[i].error == null, "hit unexpected Throwable");
				}
				
				bool success = false;
				try
				{
					writer.Close(false);
					success = true;
				}
				catch (System.IO.IOException)
				{
				}
				
				if (success)
				{
					IndexReader reader = IndexReader.Open(dir);
					for (int j = 0; j < reader.MaxDoc(); j++)
					{
						if (!reader.IsDeleted(j))
						{
							reader.Document(j);
							reader.GetTermFreqVectors(j);
						}
					}
					reader.Close();
				}
				
				dir.Close();
			}
		}

		public virtual void  TestOptimizeOverMerge()
		{
			Directory dir = new MockRAMDirectory();
			IndexWriter writer = new IndexWriter(dir, false, new StandardAnalyzer());
			writer.SetMaxBufferedDocs(2);
			writer.SetMergeFactor(100);
			writer.SetRAMBufferSizeMB(IndexWriter.DISABLE_AUTO_FLUSH);
			
			Document document = new Document();
			
			document = new Document();
			Field storedField = new Field("stored", "stored", Field.Store.YES, Field.Index.NO);
			document.Add(storedField);
			Field termVectorField = new Field("termVector", "termVector", Field.Store.NO, Field.Index.UN_TOKENIZED, Field.TermVector.WITH_POSITIONS_OFFSETS);
			document.Add(termVectorField);
			for (int i = 0; i < 170; i++)
				writer.AddDocument(document);
			
			writer.Close();
			MyIndexWriter myWriter = new MyIndexWriter(this, dir);
			myWriter.Optimize();
			Assert.AreEqual(10, myWriter.mergeCount);
		}

		public virtual void  TestCloseWithThreads()
		{
			int NUM_THREADS = 3;
			
			for (int iter = 0; iter < 50; iter++)
			{
				MockRAMDirectory dir = new MockRAMDirectory();
				IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer());
				ConcurrentMergeScheduler cms = new ConcurrentMergeScheduler();
				
				writer.SetMergeScheduler(cms);
				writer.SetMaxBufferedDocs(10);
				writer.SetMergeFactor(4);
				
				IndexerThread[] threads = new IndexerThread[NUM_THREADS];
				//bool diskFull = false;
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i] = new IndexerThread(this, writer, false);
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i].Start();
				
				try
				{
					System.Threading.Thread.Sleep(new System.TimeSpan((System.Int64) 10000 * 50));
				}
				catch (System.Threading.ThreadInterruptedException)
				{
					SupportClass.ThreadClass.Current().Interrupt();
				}
				
				writer.Close(false);
				
				// Make sure threads that are adding docs are not hung:
				for (int i = 0; i < NUM_THREADS; i++)
				{
					while (true)
					{
						try
						{
							// Without fix for LUCENE-1130: one of the
							// threads will hang
							threads[i].Join();
							break;
						}
						catch (System.Threading.ThreadInterruptedException)
						{
							SupportClass.ThreadClass.Current().Interrupt();
						}
					}
					if (threads[i].IsAlive)
						Assert.Fail("thread seems to be hung");
				}
				
				// Quick test to make sure index is not corrupt:
				IndexReader reader = IndexReader.Open(dir);
				TermDocs tdocs = reader.TermDocs(new Term("field", "aaa"));
				int count = 0;
				while (tdocs.Next())
				{
					count++;
				}
				Assert.IsTrue(count > 0);
				reader.Close();
				
				dir.Close();
			}
		}

		public virtual void  TestImmediateDiskFullWithThreads()
		{
			
			int NUM_THREADS = 3;
			
			for (int iter = 0; iter < 10; iter++)
			{
				MockRAMDirectory dir = new MockRAMDirectory();
				IndexWriter writer = new IndexWriter(dir, new WhitespaceAnalyzer());
				ConcurrentMergeScheduler cms = new ConcurrentMergeScheduler();
				// We expect disk full exceptions in the merge threads
				cms.SetSuppressExceptions_ForNUnitTest();
				writer.SetMergeScheduler(cms);
				writer.SetMaxBufferedDocs(2);
				writer.SetMergeFactor(4);
				dir.SetMaxSizeInBytes(4 * 1024 + 20 * iter);
				
				IndexerThread[] threads = new IndexerThread[NUM_THREADS];
				//bool diskFull = false;
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i] = new IndexerThread(this, writer, true);
				
				for (int i = 0; i < NUM_THREADS; i++)
					threads[i].Start();
				
				for (int i = 0; i < NUM_THREADS; i++)
				{
					while (true)
					{
						try
						{
							// Without fix for LUCENE-1130: one of the
							// threads will hang
							threads[i].Join();
							break;
						}
						catch (System.Threading.ThreadInterruptedException)
						{
							SupportClass.ThreadClass.Current().Interrupt();
						}
					}
					if (threads[i].IsAlive)
						Assert.Fail("thread seems to be hung");
					else
						Assert.IsTrue(threads[i].error == null, "hit unexpected Throwable");
				}
				
				try
				{
					writer.Close(false);
				}
				catch (System.IO.IOException)
				{
				}
				
				dir.Close();
			}
		}

 public static bool Index(Analyzer analyzer, FileIndexSet fileIndex,IndexerSet indexer, bool create)
 {
     try
     {
         IndexWriter writer = new IndexWriter(fileIndex.Path, analyzer, create);
         writer.SetMaxFieldLength(indexer.MaxFieldLength);
         writer.SetRAMBufferSizeMB(indexer.RamBufferSize);
         writer.SetMergeFactor(indexer.MergeFactor);
         writer.SetMaxBufferedDocs(indexer.MaxBufferedDocs);
         foreach (string dir in fileIndex.BaseDirs)
         {
             IndexDir(writer, dir);
         }
         return true;
     }
     catch (Exception )
     {
         return false;
     }
 }

		private void  CreateIndex(Directory dir)
		{
			IndexWriter iw = new IndexWriter(dir, anlzr, true, IndexWriter.MaxFieldLength.LIMITED);
			iw.SetMaxBufferedDocs(5);
			iw.SetMergeFactor(3);
			iw.SetSimilarity(similarityOne);
			iw.SetUseCompoundFile(true);
			iw.Close();
		}