[SPARK-12077][SQL] change the default plan for single distinct

Davies Liu · yhuai · commit a5743affcf73 · 2015-12-01T20:17:44.000-08:00
Use try to match the behavior for single distinct aggregation with Spark 1.5, but that's not scalable, we should be robust by default, have a flag to address performance regression for low cardinality aggregation. cc yhuai nongli Author: Davies Liu <davies@databricks.com> Closes apache#10075 from davies/agg_15. (cherry picked from commit 96691fe) Signed-off-by: Yin Huai <yhuai@databricks.com>
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/SQLConf.scala b/sql/core/src/main/scala/org/apache/spark/sql/SQLConf.scala
@@ -451,7 +451,7 @@ private[spark] object SQLConf {
 
   val SPECIALIZE_SINGLE_DISTINCT_AGG_PLANNING =
     booleanConf("spark.sql.specializeSingleDistinctAggPlanning",
-      defaultValue = Some(true),
+      defaultValue = Some(false),
       isPublic = false,
       doc = "When true, if a query only has a single distinct column and it has " +
         "grouping expressions, we will use our planner rule to handle this distinct " +
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/PlannerSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/PlannerSuite.scala
@@ -44,10 +44,10 @@ class PlannerSuite extends SharedSQLContext {
         fail(s"Could query play aggregation query $query. Is it an aggregation query?"))
     val aggregations = planned.collect { case n if n.nodeName contains "Aggregate" => n }
 
-    // For the new aggregation code path, there will be three aggregate operator for
+    // For the new aggregation code path, there will be four aggregate operator for
     // distinct aggregations.
     assert(
-      aggregations.size == 2 || aggregations.size == 3,
+      aggregations.size == 2 || aggregations.size == 4,
       s"The plan of query $query does not have partial aggregations.")
   }
 
