[SPARK-15255][SQL] limit the length of name for cached DataFrame

## What changes were proposed in this pull request?

We use the tree string of an SparkPlan as the name of cached DataFrame, that could be very long, cause the browser to be not responsive. This PR will limit the length of the name to 1000 characters.

## How was this patch tested?

Here is how the UI looks right now:

![ui](https://cloud.githubusercontent.com/assets/40902/15163355/d5640f9c-16bc-11e6-8655-809af8a4fed1.png)

Author: Davies Liu <davies@databricks.com>

Closes #13033 from davies/cache_name.

sql/core/src/main/scala/org/apache/spark/sql/execution/columnar/InMemoryTableScanExec.scala

@@ -19,6 +19,8 @@ package org.apache.spark.sql.execution.columnar
 
 import scala.collection.mutable.ArrayBuffer
 
+import org.apache.commons.lang.StringUtils
+
 import org.apache.spark.{Accumulable, Accumulator}
 import org.apache.spark.network.util.JavaUtils
 import org.apache.spark.rdd.RDD
@@ -177,7 +179,9 @@ private[sql] case class InMemoryRelation(
       }
     }.persist(storageLevel)
 
-    cached.setName(tableName.map(n => s"In-memory table $n").getOrElse(child.toString))
+    cached.setName(
+      tableName.map(n => s"In-memory table $n")
+        .getOrElse(StringUtils.abbreviate(child.toString, 1024)))
     _cachedColumnBuffers = cached
   }