(hbase) branch branch-2.6 updated: HBASE-29696 Fix incorrect TableInputFormat split boundaries (#8491)

[junegunn]

This is an automated email from the ASF dual-hosted git repository.

junegunn pushed a commit to branch branch-2.6
in repository https://gitbox.apache.org/repos/asf/hbase.git


The following commit(s) were added to refs/heads/branch-2.6 by this push:
     new dd429343105 HBASE-29696 Fix incorrect TableInputFormat split 
boundaries (#8491)
dd429343105 is described below

commit dd429343105e929603c1b2eea1e85bec7e2b2adc
Author: Ma Zhengxuan <49856528+sigma...@users.noreply.github.com>
AuthorDate: Thu Jul 30 13:48:43 2026 +0800

    HBASE-29696 Fix incorrect TableInputFormat split boundaries (#8491)
    
    Signed-off-by: Junegunn Choi <juneg...@apache.org>
    Reviewed-by: Andrew Olson <930946+noslower...@users.noreply.github.com>
---
 .../hbase/mapreduce/TableInputFormatBase.java       |  3 +++
 .../hbase/mapreduce/TestTableInputFormatBase.java   | 21 +++++++++++++++++++++
 2 files changed, 24 insertions(+)

diff --git 
a/hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableInputFormatBase.java
 
b/hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableInputFormatBase.java
index 7caf98a856b..5c1e7c6e33a 100644
--- 
a/hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableInputFormatBase.java
+++ 
b/hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableInputFormatBase.java
@@ -407,6 +407,9 @@ public abstract class TableInputFormatBase extends 
InputFormat<ImmutableBytesWri
 
     // Split Region into n chunks evenly
     byte[][] splitKeys = Bytes.split(startRow, endRow, true, n - 1);
+    // Restore the original boundaries after using synthetic ones to calculate 
the split keys.
+    splitKeys[0] = ts.getStartRow();
+    splitKeys[splitKeys.length - 1] = ts.getEndRow();
     for (int i = 0; i < splitKeys.length - 1; i++) {
       // In the table input format for single table we do not need to
       // store the scan object in table split because it can be memory 
intensive and redundant
diff --git 
a/hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TestTableInputFormatBase.java
 
b/hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TestTableInputFormatBase.java
index 89c8ae1e687..93e8b16ad11 100644
--- 
a/hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TestTableInputFormatBase.java
+++ 
b/hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TestTableInputFormatBase.java
@@ -17,6 +17,7 @@
  */
 package org.apache.hadoop.hbase.mapreduce;
 
+import static org.junit.jupiter.api.Assertions.assertArrayEquals;
 import static org.junit.jupiter.api.Assertions.assertEquals;
 import static org.mockito.ArgumentMatchers.any;
 import static org.mockito.ArgumentMatchers.anyBoolean;
@@ -27,6 +28,7 @@ import java.io.IOException;
 import java.net.Inet6Address;
 import java.net.InetAddress;
 import java.net.UnknownHostException;
+import java.util.List;
 import java.util.Map;
 import java.util.TreeMap;
 import java.util.concurrent.ExecutorService;
@@ -50,6 +52,7 @@ import org.apache.hadoop.hbase.security.User;
 import org.apache.hadoop.hbase.testclassification.SmallTests;
 import org.apache.hadoop.hbase.util.Bytes;
 import org.apache.hadoop.hbase.util.Pair;
+import org.apache.hadoop.mapreduce.InputSplit;
 import org.apache.hadoop.mapreduce.JobContext;
 import org.junit.jupiter.api.Tag;
 import org.junit.jupiter.api.Test;
@@ -60,6 +63,24 @@ import org.mockito.stubbing.Answer;
 @Tag(SmallTests.TAG)
 public class TestTableInputFormatBase {
 
+  @Test
+  public void testCreateNInputSplitsUniformPreservesOriginalBoundaries() 
throws IOException {
+    TableInputFormat inputFormat = new TableInputFormat();
+    for (byte[] startRow : new byte[][] { HConstants.EMPTY_START_ROW, 
Bytes.toBytes("start") }) {
+      TableSplit split =
+        new TableSplit(TableName.valueOf("test"), startRow, 
HConstants.EMPTY_END_ROW, "localhost");
+
+      List<InputSplit> splits = inputFormat.createNInputSplitsUniform(split, 
2);
+
+      assertEquals(2, splits.size());
+      TableSplit first = (TableSplit) splits.get(0);
+      TableSplit last = (TableSplit) splits.get(1);
+      assertArrayEquals(startRow, first.getStartRow());
+      assertArrayEquals(first.getEndRow(), last.getStartRow());
+      assertArrayEquals(HConstants.EMPTY_END_ROW, last.getEndRow());
+    }
+  }
+
   @Test
   public void testReuseRegionSizeCalculator() throws IOException {
     JobContext context = mock(JobContext.class);