From efef11e2dc4dec550e37678b4a4117f213139407 Mon Sep 17 00:00:00 2001
From: Michael Stack <stack@apache.org>
Date: Fri, 13 Nov 2009 06:39:58 +0000
Subject: [PATCH] HBASE-1841 If multiple of same key in an hfile and they span
 blocks, may miss the earlier keys on a lookup

git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@835757 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                   |  3 +++
 .../apache/hadoop/hbase/io/hfile/HFile.java   | 19 ++++++++++++++-----
 2 files changed, 17 insertions(+), 5 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 0a0d788be37..86f6fc5886a 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -107,6 +107,9 @@ Release 0.21.0 - Unreleased
                Java 5
    HBASE-1967  [Transactional] client.TestTransactions.testPutPutScan fails
                sometimes -- Temporary fix
+   HBASE-1841  If multiple of same key in an hfile and they span blocks, may
+               miss the earlier keys on a lookup
+               (Schubert Zhang via Stack)
 
   IMPROVEMENTS
    HBASE-1760  Cleanup TODOs in HTable
diff --git a/src/java/org/apache/hadoop/hbase/io/hfile/HFile.java b/src/java/org/apache/hadoop/hbase/io/hfile/HFile.java
index e212cf745f4..ee7b03ae460 100644
--- a/src/java/org/apache/hadoop/hbase/io/hfile/HFile.java
+++ b/src/java/org/apache/hadoop/hbase/io/hfile/HFile.java
@@ -475,9 +475,11 @@ public class HFile {
     public void append(final byte [] key, final int koffset, final int klength,
         final byte [] value, final int voffset, final int vlength)
     throws IOException {
-      checkKey(key, koffset, klength);
+      boolean dupKey = checkKey(key, koffset, klength);
       checkValue(value, voffset, vlength);
-      checkBlockBoundary();
+      if (!dupKey) {
+        checkBlockBoundary();
+      }
       // Write length of key and value and then actual key and value bytes.
       this.out.writeInt(klength);
       this.keylength += klength;
@@ -499,10 +501,13 @@ public class HFile {
 
     /*
      * @param key Key to check.
+     * @return the flag of duplicate Key or not
      * @throws IOException
      */
-    private void checkKey(final byte [] key, final int offset, final int length)
+    private boolean checkKey(final byte [] key, final int offset, final int length)
     throws IOException {
+      boolean dupKey = false;
+      
       if (key == null || length <= 0) {
         throw new IOException("Key cannot be null or empty");
       }
@@ -511,14 +516,18 @@ public class HFile {
           MAXIMUM_KEY_LENGTH);
       }
       if (this.lastKeyBuffer != null) {
-        if (this.comparator.compare(this.lastKeyBuffer, this.lastKeyOffset,
-            this.lastKeyLength, key, offset, length) > 0) {
+        int keyComp = this.comparator.compare(this.lastKeyBuffer, this.lastKeyOffset,
+            this.lastKeyLength, key, offset, length);        
+        if (keyComp > 0) {
           throw new IOException("Added a key not lexically larger than" +
             " previous key=" + Bytes.toString(key, offset, length) +
             ", lastkey=" + Bytes.toString(this.lastKeyBuffer, this.lastKeyOffset,
                 this.lastKeyLength));
+        } else if (keyComp == 0) {
+          dupKey = true;
         }
       }
+      return dupKey;
     }
 
     private void checkValue(final byte [] value, final int offset,