[SPARK-7686] [SQL] DescribeCommand is assigned wrong output attributes in SparkStrategies

In `SparkStrategies`, `RunnableDescribeCommand` is called with the output attributes of the table being described rather than the attributes for the `describe` command's output.  I discovered this issue because it caused type conversion errors in some UnsafeRow conversion code that I'm writing.

Author: Josh Rosen <[email protected]>

Closes apache#6217 from JoshRosen/SPARK-7686 and squashes the following commits:

953a344 [Josh Rosen] Fix SPARK-7686 with a simple change in SparkStrategies.
a4eec9f [Josh Rosen] Add failing regression test for SPARK-7686

Author: JoshRosen

sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala

@@ -382,10 +382,10 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
       case c: CreateTableUsingAsSelect if !c.temporary =>
         sys.error("Tables created with SQLContext must be TEMPORARY. Use a HiveContext instead.")
 
-      case LogicalDescribeCommand(table, isExtended) =>
+      case describe @ LogicalDescribeCommand(table, isExtended) =>
         val resultPlan = self.sqlContext.executePlan(table).executedPlan
         ExecutedCommand(
-          RunnableDescribeCommand(resultPlan, resultPlan.output, isExtended)) :: Nil
+          RunnableDescribeCommand(resultPlan, describe.output, isExtended)) :: Nil
 
       case _ => Nil
     }

sql/core/src/test/scala/org/apache/spark/sql/sources/DDLTestSuite.scala

@@ -99,4 +99,10 @@ class DDLTestSuite extends DataSourceTest {
         Row("arrayType", "array<string>", ""),
         Row("structType", "struct<f1:string,f2:int>", "")
       ))
+
+  test("SPARK-7686 DescribeCommand should have correct physical plan output attributes") {
+    val attributes = sql("describe ddlPeople").queryExecution.executedPlan.output
+    assert(attributes.map(_.name) === Seq("col_name", "data_type", "comment"))
+    assert(attributes.map(_.dataType).toSet === Set(StringType))
+  }
 }