[SPARK-10197] [SQL] Add null check in wrapperFor (inside HiveInspecto…

…rs). https://issues.apache.org/jira/browse/SPARK-10197 Author: Yin Huai <yhuai@databricks.com> Closes #8407 from yhuai/ORCSPARK-10197.
apache · Aug 25, 2015 · 0e6368f · 0e6368f
1 parent 7bc9a8c
commit 0e6368f
Show file tree

Hide file tree

Showing 2 changed files with 53 additions and 5 deletions.
diff --git a/sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveInspectors.scala b/sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveInspectors.scala
@@ -370,17 +370,36 @@ private[hive] trait HiveInspectors {
   protected def wrapperFor(oi: ObjectInspector, dataType: DataType): Any => Any = oi match {
     case _: JavaHiveVarcharObjectInspector =>
       (o: Any) =>
-        val s = o.asInstanceOf[UTF8String].toString
-        new HiveVarchar(s, s.size)
+        if (o != null) {
+          val s = o.asInstanceOf[UTF8String].toString
+          new HiveVarchar(s, s.size)
+        } else {
+          null
+        }
 
     case _: JavaHiveDecimalObjectInspector =>
-      (o: Any) => HiveDecimal.create(o.asInstanceOf[Decimal].toJavaBigDecimal)
+      (o: Any) =>
+        if (o != null) {
+          HiveDecimal.create(o.asInstanceOf[Decimal].toJavaBigDecimal)
+        } else {
+          null
+        }
 
     case _: JavaDateObjectInspector =>
-      (o: Any) => DateTimeUtils.toJavaDate(o.asInstanceOf[Int])
+      (o: Any) =>
+        if (o != null) {
+          DateTimeUtils.toJavaDate(o.asInstanceOf[Int])
+        } else {
+          null
+        }
 
     case _: JavaTimestampObjectInspector =>
-      (o: Any) => DateTimeUtils.toJavaTimestamp(o.asInstanceOf[Long])
+      (o: Any) =>
+        if (o != null) {
+          DateTimeUtils.toJavaTimestamp(o.asInstanceOf[Long])
+        } else {
+          null
+        }
 
     case soi: StandardStructObjectInspector =>
       val schema = dataType.asInstanceOf[StructType]

diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/hive/orc/OrcSourceSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/hive/orc/OrcSourceSuite.scala
@@ -121,6 +121,35 @@ abstract class OrcSuite extends QueryTest with BeforeAndAfterAll {
       sql("SELECT * FROM normal_orc_as_source"),
       (6 to 10).map(i => Row(i, s"part-$i")))
   }
+
+  test("write null values") {
+    sql("DROP TABLE IF EXISTS orcNullValues")
+
+    val df = sql(
+      """
+        |SELECT
+        |  CAST(null as TINYINT),
+        |  CAST(null as SMALLINT),
+        |  CAST(null as INT),
+        |  CAST(null as BIGINT),
+        |  CAST(null as FLOAT),
+        |  CAST(null as DOUBLE),
+        |  CAST(null as DECIMAL(7,2)),
+        |  CAST(null as TIMESTAMP),
+        |  CAST(null as DATE),
+        |  CAST(null as STRING),
+        |  CAST(null as VARCHAR(10))
+        |FROM orc_temp_table limit 1
+      """.stripMargin)
+
+    df.write.format("orc").saveAsTable("orcNullValues")
+
+    checkAnswer(
+      sql("SELECT * FROM orcNullValues"),
+      Row.fromSeq(Seq.fill(11)(null)))
+
+    sql("DROP TABLE IF EXISTS orcNullValues")
+  }
 }
 
 class OrcSourceSuite extends OrcSuite {