[SPARK-10190] Fix NPE in CatalystTypeConverters Decimal toScala converter

This adds a missing null check to the Decimal `toScala` converter in `CatalystTypeConverters`, fixing an NPE.

Author: Josh Rosen <joshrosen@databricks.com>

Closes #8401 from JoshRosen/SPARK-10190.

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/CatalystTypeConverters.scala

@@ -329,7 +329,10 @@ object CatalystTypeConverters {
         null
       }
     }
-    override def toScala(catalystValue: Decimal): JavaBigDecimal = catalystValue.toJavaBigDecimal
+    override def toScala(catalystValue: Decimal): JavaBigDecimal = {
+      if (catalystValue == null) null
+      else catalystValue.toJavaBigDecimal
+    }
     override def toScalaImpl(row: InternalRow, column: Int): JavaBigDecimal =
       row.getDecimal(column, dataType.precision, dataType.scale).toJavaBigDecimal
   }

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/CatalystTypeConvertersSuite.scala

@@ -32,7 +32,9 @@ class CatalystTypeConvertersSuite extends SparkFunSuite {
     IntegerType,
     LongType,
     FloatType,
-    DoubleType)
+    DoubleType,
+    DecimalType.SYSTEM_DEFAULT,
+    DecimalType.USER_DEFAULT)
 
   test("null handling in rows") {
     val schema = StructType(simpleTypes.map(t => StructField(t.getClass.getName, t)))