[SPARK-20668][SQL] Modify ScalaUDF to handle nullability.

## What changes were proposed in this pull request?

When registering Scala UDF, we can know if the udf will return nullable value or not. `ScalaUDF` and related classes should handle the nullability.

## How was this patch tested?

Existing tests.

Author: Takuya UESHIN <ueshin@databricks.com>

Closes apache#17911 from ueshin/issues/SPARK-20668.

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala

@@ -2128,7 +2128,7 @@ class Analyzer(
 
       case p => p transformExpressionsUp {
 
-        case udf @ ScalaUDF(func, _, inputs, _, _) =>
+        case udf @ ScalaUDF(func, _, inputs, _, _, _) =>
           val parameterTypes = ScalaReflection.getParameterTypes(func)
           assert(parameterTypes.length == inputs.length)
 

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/ScalaUDF.scala

@@ -36,17 +36,17 @@ import org.apache.spark.sql.types.DataType
  *                    better to use Option of Seq[DataType] so we can use "None" as the case for no
  *                    type coercion. However, that would require more refactoring of the codebase.
  * @param udfName   The user-specified name of this UDF.
+ * @param nullable  True if the UDF can return null value.
  */
 case class ScalaUDF(
     function: AnyRef,
     dataType: DataType,
     children: Seq[Expression],
     inputTypes: Seq[DataType] = Nil,
-    udfName: Option[String] = None)
+    udfName: Option[String] = None,
+    nullable: Boolean = true)
   extends Expression with ImplicitCastInputTypes with NonSQLExpression {
 
-  override def nullable: Boolean = true
-
   override def toString: String =
     s"${udfName.map(name => s"UDF:$name").getOrElse("UDF")}(${children.mkString(", ")})"