[SPARK-14356] Update spark.sql.execution.debug to work on Datasets

## What changes were proposed in this pull request?

Update DebugQuery to work on Datasets of any type, not just DataFrames.

## How was this patch tested?

Added unit tests, checked in spark-shell.

Author: Matei Zaharia <matei@databricks.com>

Closes #12140 from mateiz/debug-dataset.

sql/core/src/main/scala/org/apache/spark/sql/execution/debug/package.scala

@@ -77,9 +77,9 @@ package object debug {
   }
 
   /**
-   * Augments [[DataFrame]]s with debug methods.
+   * Augments [[Dataset]]s with debug methods.
    */
-  implicit class DebugQuery(query: DataFrame) extends Logging {
+  implicit class DebugQuery(query: Dataset[_]) extends Logging {
     def debug(): Unit = {
       val plan = query.queryExecution.executedPlan
       val visited = new collection.mutable.HashSet[TreeNodeRef]()

sql/core/src/test/scala/org/apache/spark/sql/execution/debug/DebuggingSuite.scala

@@ -19,6 +19,7 @@ package org.apache.spark.sql.execution.debug
 
 import org.apache.spark.SparkFunSuite
 import org.apache.spark.sql.test.SharedSQLContext
+import org.apache.spark.sql.test.SQLTestData.TestData
 
 class DebuggingSuite extends SparkFunSuite with SharedSQLContext {
 
@@ -26,6 +27,11 @@ class DebuggingSuite extends SparkFunSuite with SharedSQLContext {
     testData.debug()
   }
 
+  test("Dataset.debug()") {
+    import testImplicits._
+    testData.as[TestData].debug()
+  }
+
   test("debugCodegen") {
     val res = codegenString(sqlContext.range(10).groupBy("id").count().queryExecution.executedPlan)
     assert(res.contains("Subtree 1 / 2"))