Restore "null" as default for RDD name

Instead, RDDInfo's name field will default to the RDD class name
if the RDD's name is not defined. This avoids changing RDD#toString
behavior.

Author: Andrew Or

core/src/main/scala/org/apache/spark/SparkContext.scala

@@ -898,7 +898,7 @@ class SparkContext(config: SparkConf) extends Logging with ExecutorAllocationCli
       inputFormatClass,
       keyClass,
       valueClass,
-      minPartitions).setName(s"HadoopRDD[$path]")
+      minPartitions).setName(path)
   }
 
   /**

core/src/main/scala/org/apache/spark/rdd/RDD.scala

@@ -140,7 +140,7 @@ abstract class RDD[T: ClassTag](
   val id: Int = sc.newRddId()
 
   /** A friendly name for this RDD */
-  @transient var name: String = Utils.getFormattedClassName(this)
+  @transient var name: String = null
 
   /** Assign a name to this RDD */
   def setName(_name: String): this.type = {

core/src/main/scala/org/apache/spark/storage/RDDInfo.scala

@@ -54,7 +54,7 @@ class RDDInfo(
 
 private[spark] object RDDInfo {
   def fromRdd(rdd: RDD[_]): RDDInfo = {
-    val rddName = Option(rdd.name).getOrElse(rdd.id.toString)
+    val rddName = Option(rdd.name).getOrElse(Utils.getFormattedClassName(rdd))
     val parentIds = rdd.dependencies.map(_.rdd.id)
     new RDDInfo(rdd.id, rddName, rdd.partitions.length, rdd.getStorageLevel, parentIds, rdd.scope)
   }

python/pyspark/sql/dataframe.py

@@ -244,8 +244,8 @@ def explain(self, extended=False):
         :param extended: boolean, default ``False``. If ``False``, prints only the physical plan.
 
         >>> df.explain()
-        PhysicalRDD [age#0,name#1], MapPartitionsRDD MapPartitionsRDD[7] at applySchemaToPythonRDD\
-          at NativeMethodAccessorImpl.java:...
+        PhysicalRDD [age#0,name#1], MapPartitionsRDD[...] at applySchemaToPythonRDD at\
+          NativeMethodAccessorImpl.java:...
 
         >>> df.explain(True)
         == Parsed Logical Plan ==