From 71291daa74b9e687d897a2890bdc37e0d9156ba9 Mon Sep 17 00:00:00 2001
From: Robert Muir <rmuir@apache.org>
Date: Wed, 11 Apr 2012 12:16:31 +0000
Subject: [PATCH] LUCENE-3969: when outputting a bigram token, mark posLen=2 to
 note that it spans two tokens

git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/branches/lucene3969@1324727 13f79535-47bb-0310-9956-ffa450edef68
---
 .../apache/lucene/analysis/commongrams/CommonGramsFilter.java  | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/modules/analysis/common/src/java/org/apache/lucene/analysis/commongrams/CommonGramsFilter.java b/modules/analysis/common/src/java/org/apache/lucene/analysis/commongrams/CommonGramsFilter.java
index 8232b88c2bf..9798464f938 100644
--- a/modules/analysis/common/src/java/org/apache/lucene/analysis/commongrams/CommonGramsFilter.java
+++ b/modules/analysis/common/src/java/org/apache/lucene/analysis/commongrams/CommonGramsFilter.java
@@ -16,6 +16,7 @@ import org.apache.lucene.analysis.TokenStream;
 import org.apache.lucene.analysis.tokenattributes.OffsetAttribute;
 import org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute;
 import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
+import org.apache.lucene.analysis.tokenattributes.PositionLengthAttribute;
 import org.apache.lucene.analysis.tokenattributes.TypeAttribute;
 import org.apache.lucene.analysis.util.CharArraySet;
 import org.apache.lucene.util.Version;
@@ -54,6 +55,7 @@ public final class CommonGramsFilter extends TokenFilter {
   private final OffsetAttribute offsetAttribute = addAttribute(OffsetAttribute.class);
   private final TypeAttribute typeAttribute = addAttribute(TypeAttribute.class);
   private final PositionIncrementAttribute posIncAttribute = addAttribute(PositionIncrementAttribute.class);
+  private final PositionLengthAttribute posLenAttribute = addAttribute(PositionLengthAttribute.class);
 
   private int lastStartOffset;
   private boolean lastWasCommon;
@@ -166,6 +168,7 @@ public final class CommonGramsFilter extends TokenFilter {
     buffer.getChars(0, length, termText, 0);
     termAttribute.setLength(length);
     posIncAttribute.setPositionIncrement(0);
+    posLenAttribute.setPositionLength(2); // bigram
     offsetAttribute.setOffset(lastStartOffset, endOffset);
     typeAttribute.setType(GRAM_TYPE);
     buffer.setLength(0);