svn commit: r1039314 - in /lucene/dev/trunk/lucene/src: java/org/apache/lucene/index/TermVectorsReader.java java/org/apache/lucene/index/TermVectorsTermsWriter.java test/org/apache/lucene/index/TestIndexWriter.java

[mi...@apache.org]

Author: mikemccand
Date: Fri Nov 26 11:25:42 2010
New Revision: 1039314

URL: http://svn.apache.org/viewvc?rev=1039314&view=rev
Log:
LUCENE-2776: don't create term vectors files if no doc uses term vectors

Modified:
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsReader.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsTermsWriter.java
    lucene/dev/trunk/lucene/src/test/org/apache/lucene/index/TestIndexWriter.java

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsReader.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsReader.java?rev=1039314&r1=1039313&r2=1039314&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsReader.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsReader.java Fri Nov 26 11:25:42 2010
@@ -104,9 +104,10 @@ class TermVectorsReader implements Clone
           assert numTotalDocs >= size + docStoreOffset: "numTotalDocs=" + numTotalDocs + " size=" + size + " docStoreOffset=" + docStoreOffset;
         }
       } else {
-        // TODO: understand why FieldInfos.hasVectors() can
-        // return true yet the term vectors files don't
-        // exist...
+        // If all documents flushed in a segment had hit
+        // non-aborting exceptions, it's possible that
+        // FieldInfos.hasVectors returns true yet the term
+        // vector files don't exist.
         format = 0;
       }
 

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsTermsWriter.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsTermsWriter.java?rev=1039314&r1=1039313&r2=1039314&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsTermsWriter.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/index/TermVectorsTermsWriter.java Fri Nov 26 11:25:42 2010
@@ -50,12 +50,13 @@ final class TermVectorsTermsWriter exten
   @Override
   synchronized void flush(Map<TermsHashConsumerPerThread,Collection<TermsHashConsumerPerField>> threadsAndFields, final SegmentWriteState state) throws IOException {
 
-    if (state.numDocsInStore > 0) {
-      // It's possible that all documents seen in this segment
-      // hit non-aborting exceptions, in which case we will
-      // not have yet init'd the TermVectorsWriter:
-      initTermVectorsWriter();
-    }
+    // NOTE: it's possible that all documents seen in this segment
+    // hit non-aborting exceptions, in which case we will
+    // not have yet init'd the TermVectorsWriter.  This is
+    // actually OK (unlike in the stored fields case)
+    // because, although IieldInfos.hasVectors() will return
+    // true, the TermVectorsReader gracefully handles
+    // non-existence of the term vectors files.
 
     if (tvx != null) {
 

Modified: lucene/dev/trunk/lucene/src/test/org/apache/lucene/index/TestIndexWriter.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/test/org/apache/lucene/index/TestIndexWriter.java?rev=1039314&r1=1039313&r2=1039314&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/test/org/apache/lucene/index/TestIndexWriter.java (original)
+++ lucene/dev/trunk/lucene/src/test/org/apache/lucene/index/TestIndexWriter.java Fri Nov 26 11:25:42 2010
@@ -2961,4 +2961,37 @@ public class TestIndexWriter extends Luc
     w.close();
     dir.close();
   }
+
+  public void testNoUnwantedTVFiles() throws Exception {
+
+    Directory dir = newDirectory();
+    IndexWriter indexWriter = new IndexWriter(dir, newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer()).setRAMBufferSizeMB(0.01));
+    ((LogMergePolicy) indexWriter.getMergePolicy()).setUseCompoundFile(false);
+
+    String BIG="alskjhlaksjghlaksjfhalksvjepgjioefgjnsdfjgefgjhelkgjhqewlrkhgwlekgrhwelkgjhwelkgrhwlkejg";
+    BIG=BIG+BIG+BIG+BIG;
+
+    for (int i=0; i<2; i++) {
+      Document doc = new Document();
+      doc.add(new Field("id", Integer.toString(i)+BIG, Field.Store.YES, Field.Index.NOT_ANALYZED_NO_NORMS));
+      doc.add(new Field("str", Integer.toString(i)+BIG, Field.Store.YES, Field.Index.NOT_ANALYZED));
+      doc.add(new Field("str2", Integer.toString(i)+BIG, Field.Store.YES, Field.Index.ANALYZED));
+      doc.add(new Field("str3", Integer.toString(i)+BIG, Field.Store.YES, Field.Index.ANALYZED_NO_NORMS));
+      indexWriter.addDocument(doc);
+    }
+
+    indexWriter.close();
+
+    _TestUtil.checkIndex(dir);
+
+    assertNoUnreferencedFiles(dir, "no tv files");
+    String[] files = dir.listAll();
+    for(String file : files) {
+      assertTrue(!file.endsWith(IndexFileNames.VECTORS_FIELDS_EXTENSION));
+      assertTrue(!file.endsWith(IndexFileNames.VECTORS_INDEX_EXTENSION));
+      assertTrue(!file.endsWith(IndexFileNames.VECTORS_DOCUMENTS_EXTENSION));
+    }
+
+    dir.close();
+  }
 }