Hive: Return null if Hive inspects a null struct record (#4283)

Fix #4282
apache · Jul 6, 2022 · 3959e2f · 3959e2f
1 parent 56c1993
commit 3959e2f
Show file tree

Hide file tree

Showing 3 changed files with 50 additions and 0 deletions.
diff --git a/...n/java/org/apache/iceberg/mr/hive/serde/objectinspector/IcebergRecordObjectInspector.java b/...n/java/org/apache/iceberg/mr/hive/serde/objectinspector/IcebergRecordObjectInspector.java
@@ -72,11 +72,19 @@ public StructField getStructFieldRef(String name) {
 
   @Override
   public Object getStructFieldData(Object o, StructField structField) {
+    if (o == null) {
+      return null;
+    }
+
     return ((Record) o).get(((IcebergRecordStructField) structField).position());
   }
 
   @Override
   public List<Object> getStructFieldsDataAsList(Object o) {
+    if (o == null) {
+      return null;
+    }
+
     Record record = (Record) o;
     return structFields
             .stream()

diff --git a/mr/src/test/java/org/apache/iceberg/mr/hive/TestHiveIcebergStorageHandlerWithEngine.java b/mr/src/test/java/org/apache/iceberg/mr/hive/TestHiveIcebergStorageHandlerWithEngine.java
@@ -41,6 +41,7 @@
 import org.apache.iceberg.mr.InputFormatConfig;
 import org.apache.iceberg.mr.TestHelper;
 import org.apache.iceberg.relocated.com.google.common.collect.ImmutableList;
+import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;
 import org.apache.iceberg.relocated.com.google.common.collect.Lists;
 import org.apache.iceberg.types.Type;
 import org.apache.iceberg.types.Types;
@@ -707,6 +708,34 @@ public void testMultiTableInsert() throws IOException {
     HiveIcebergTestUtils.validateData(target2, target2Records, 1);
   }
 
+  /**
+  * Fix vectorized parquet
+  * <a href="https://github.com/apache/iceberg/issues/4403">issue-4403</a>.
+  **/
+  @Test
+  public void testStructMapWithNull() throws IOException {
+    Assume.assumeTrue("Vectorized parquet throw class cast exception see : issue 4403",
+        !("PARQUET".equals(fileFormat.name()) && isVectorized));
+    Schema schema = new Schema(required(1, "id", Types.LongType.get()),
+        required(2, "mapofstructs", Types.MapType.ofRequired(3, 4, Types.StringType.get(),
+            Types.StructType.of(required(5, "something", Types.StringType.get()),
+                required(6, "someone", Types.StringType.get()),
+                required(7, "somewhere", Types.StringType.get())
+            ))
+        )
+    );
+
+    List<Record> records = TestHelper.RecordsBuilder.newInstance(schema)
+        .add(0L, ImmutableMap.of())
+        .build();
+
+    testTables.createTable(shell, "mapwithnull", schema, fileFormat, records);
+
+    List<Object[]> results = shell.executeStatement("select mapofstructs['context'].someone FROM mapwithnull");
+    Assert.assertEquals(1, results.size());
+    Assert.assertNull(results.get(0)[0]);
+  }
+
   @Test
   public void testWriteWithDefaultWriteFormat() {
     Assume.assumeTrue("Testing the default file format is enough for a single scenario.",

diff --git a/...va/org/apache/iceberg/mr/hive/serde/objectinspector/TestIcebergRecordObjectInspector.java b/...va/org/apache/iceberg/mr/hive/serde/objectinspector/TestIcebergRecordObjectInspector.java
@@ -61,4 +61,17 @@ public void testIcebergRecordObjectInspector() {
     Assert.assertEquals(innerRecord.get(0), innerSoi.getStructFieldData(innerData, stringField));
   }
 
+  @Test
+  public void testIcebergRecordObjectInspectorWithRowNull() {
+    Schema schema = new Schema(
+        required(1, "integer_field", Types.IntegerType.get()),
+        required(2, "struct_field", Types.StructType.of(
+            Types.NestedField.required(3, "string_field", Types.StringType.get())))
+    );
+    StructObjectInspector soi = (StructObjectInspector) IcebergObjectInspector.create(schema);
+    Assert.assertNull(soi.getStructFieldsDataAsList(null));
+    StructField integerField = soi.getStructFieldRef("integer_field");
+    Assert.assertNull(soi.getStructFieldData(null, integerField));
+  }
+
 }