apacheGH-40038: [Java] The offset buffer of empty array with variable…

…-size layout should not be empty
viirya · Feb 11, 2024 · e57be3b · e57be3b
1 parent d989191
commit e57be3b
Show file tree

Hide file tree

Showing 6 changed files with 30 additions and 5 deletions.
diff --git a/java/vector/src/main/java/org/apache/arrow/vector/BaseLargeVariableWidthVector.java b/java/vector/src/main/java/org/apache/arrow/vector/BaseLargeVariableWidthVector.java
@@ -73,7 +73,8 @@ public BaseLargeVariableWidthVector(Field field, final BufferAllocator allocator
     lastValueCapacity = INITIAL_VALUE_ALLOCATION - 1;
     valueCount = 0;
     lastSet = -1;
-    offsetBuffer = allocator.getEmpty();
+    // According to Arrow spec, the offsets buffer contains length + 1 elements
+    allocateOffsetBuffer(OFFSET_WIDTH);
     validityBuffer = allocator.getEmpty();
     valueBuffer = allocator.getEmpty();
   }

diff --git a/java/vector/src/main/java/org/apache/arrow/vector/BaseVariableWidthVector.java b/java/vector/src/main/java/org/apache/arrow/vector/BaseVariableWidthVector.java
@@ -74,7 +74,8 @@ public BaseVariableWidthVector(Field field, final BufferAllocator allocator) {
     lastValueCapacity = INITIAL_VALUE_ALLOCATION - 1;
     valueCount = 0;
     lastSet = -1;
-    offsetBuffer = allocator.getEmpty();
+    // According to Arrow spec, the offsets buffer contains length + 1 elements
+    allocateOffsetBuffer(OFFSET_WIDTH);
     validityBuffer = allocator.getEmpty();
     valueBuffer = allocator.getEmpty();
   }

diff --git a/java/vector/src/main/java/org/apache/arrow/vector/complex/BaseRepeatedValueVector.java b/java/vector/src/main/java/org/apache/arrow/vector/complex/BaseRepeatedValueVector.java
@@ -68,7 +68,8 @@ protected BaseRepeatedValueVector(String name, BufferAllocator allocator, CallBa
   protected BaseRepeatedValueVector(String name, BufferAllocator allocator, FieldVector vector, CallBack callBack) {
     super(allocator);
     this.name = name;
-    this.offsetBuffer = allocator.getEmpty();
+    // According to Arrow spec, the offsets buffer contains length + 1 elements
+    allocateOffsetBuffer(OFFSET_WIDTH);
     this.vector = Preconditions.checkNotNull(vector, "data vector cannot be null");
     this.repeatedCallBack = callBack;
     this.valueCount = 0;

diff --git a/java/vector/src/test/java/org/apache/arrow/vector/TestLargeVarCharVector.java b/java/vector/src/test/java/org/apache/arrow/vector/TestLargeVarCharVector.java
@@ -68,6 +68,13 @@ public void shutdown() {
     allocator.close();
   }
 
+  @Test
+  public void testEmptyLargeVarCharVector() {
+    LargeVarCharVector vector = new LargeVarCharVector("empty", allocator);
+    assertEquals(8, vector.getOffsetBuffer().capacity());
+    vector.close();
+  }
+
   @Test
   public void testTransfer() {
     try (BufferAllocator childAllocator1 = allocator.newChildAllocator("child1", 1000000, 1000000);

diff --git a/java/vector/src/test/java/org/apache/arrow/vector/TestValueVector.java b/java/vector/src/test/java/org/apache/arrow/vector/TestValueVector.java
@@ -315,6 +315,13 @@ public void testFixedType2() {
     }
   }
 
+  @Test
+  public void testEmptyVarCharVector() {
+    VarCharVector vector = new VarCharVector(EMPTY_SCHEMA_PATH, allocator);
+    assertEquals(4, vector.getOffsetBuffer().capacity());
+    vector.close();
+  }
+
   @Test /* VarCharVector */
   public void testSizeOfValueBuffer() {
     try (final VarCharVector vector = new VarCharVector(EMPTY_SCHEMA_PATH, allocator)) {
@@ -3216,7 +3223,7 @@ public void testEmptyBufBehavior() {
       assertEquals(1, vector.getOffsetBuffer().refCnt());
       assertEquals(0, vector.getDataBuffer().capacity());
       assertEquals(0, vector.getValidityBuffer().capacity());
-      assertEquals(0, vector.getOffsetBuffer().capacity());
+      assertEquals(4, vector.getOffsetBuffer().capacity());
 
       vector.allocateNew(valueCount);
       assertEquals(1, vector.getDataBuffer().refCnt());
@@ -3239,7 +3246,7 @@ public void testEmptyBufBehavior() {
       assertEquals(1, vector.getValidityBuffer().refCnt());
       assertEquals(1, vector.getOffsetBuffer().refCnt());
       assertEquals(0, vector.getValidityBuffer().capacity());
-      assertEquals(0, vector.getOffsetBuffer().capacity());
+      assertEquals(4, vector.getOffsetBuffer().capacity());
 
       vector.setValueCount(valueCount);
       vector.allocateNewSafe();

diff --git a/java/vector/src/test/java/org/apache/arrow/vector/TestVarCharListVector.java b/java/vector/src/test/java/org/apache/arrow/vector/TestVarCharListVector.java
@@ -44,6 +44,14 @@ public void terminate() throws Exception {
     allocator.close();
   }
 
+  @Test
+  public void testEmptyVarCharList() {
+    ListVector vector = new ListVector("VarList", allocator,
+            FieldType.nullable(Types.MinorType.VARCHAR.getType()), null);
+    Assert.assertEquals(4, vector.getOffsetBuffer().capacity());
+    vector.close();
+  }
+
   @Test
   public void testVarCharListWithNulls() {
     byte[] bytes = "a".getBytes(StandardCharsets.UTF_8);