[SPARK-14796][SQL] Add spark.sql.optimizer.minSetSize config option.

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/CatalystConf.scala

@@ -29,6 +29,7 @@ trait CatalystConf {
   def groupByOrdinal: Boolean
 
   def optimizerMaxIterations: Int
+  def optimizerMinSetSize: Int
   def maxCaseBranchesForCodegen: Int
 
   /**
@@ -47,6 +48,7 @@ case class SimpleCatalystConf(
     orderByOrdinal: Boolean = true,
     groupByOrdinal: Boolean = true,
     optimizerMaxIterations: Int = 100,
+    optimizerMinSetSize: Int = 10,
     maxCaseBranchesForCodegen: Int = 20)
   extends CatalystConf {
 }

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala

@@ -87,7 +87,7 @@ abstract class Optimizer(sessionCatalog: SessionCatalog, conf: CatalystConf)
       CombineUnions,
       // Constant folding and strength reduction
       NullPropagation,
-      OptimizeIn,
+      OptimizeIn(conf),
       ConstantFolding,
       LikeSimplification,
       BooleanSimplification,
@@ -682,10 +682,11 @@ object ConstantFolding extends Rule[LogicalPlan] {
  * Replaces [[In (value, seq[Literal])]] with optimized version[[InSet (value, HashSet[Literal])]]
  * which is much faster
  */
-object OptimizeIn extends Rule[LogicalPlan] {
+case class OptimizeIn(conf: CatalystConf) extends Rule[LogicalPlan] {
   def apply(plan: LogicalPlan): LogicalPlan = plan transform {
     case q: LogicalPlan => q transformExpressionsDown {
-      case In(v, list) if !list.exists(!_.isInstanceOf[Literal]) && list.size > 10 =>
+      case In(v, list) if !list.exists(!_.isInstanceOf[Literal]) &&
+          list.size > conf.optimizerMinSetSize =>
         val hSet = list.map(e => e.eval(EmptyRow))
         InSet(v, HashSet() ++ hSet)
     }

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/ConstantFoldingSuite.scala

@@ -17,6 +17,7 @@
 
 package org.apache.spark.sql.catalyst.optimizer
 
+import org.apache.spark.sql.catalyst.SimpleCatalystConf
 import org.apache.spark.sql.catalyst.analysis.{EliminateSubqueryAliases, UnresolvedExtractValue}
 import org.apache.spark.sql.catalyst.dsl.expressions._
 import org.apache.spark.sql.catalyst.dsl.plans._
@@ -33,7 +34,7 @@ class ConstantFoldingSuite extends PlanTest {
       Batch("AnalysisNodes", Once,
         EliminateSubqueryAliases) ::
       Batch("ConstantFolding", Once,
-        OptimizeIn,
+        OptimizeIn(SimpleCatalystConf(true)),
         ConstantFolding,
         BooleanSimplification) :: Nil
   }

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/OptimizeInSuite.scala

@@ -17,11 +17,14 @@
 
 package org.apache.spark.sql.catalyst.optimizer
 
+import scala.collection.immutable.HashSet
+
+import org.apache.spark.sql.catalyst.SimpleCatalystConf
 import org.apache.spark.sql.catalyst.analysis.{EliminateSubqueryAliases, UnresolvedAttribute}
 import org.apache.spark.sql.catalyst.dsl.expressions._
 import org.apache.spark.sql.catalyst.dsl.plans._
 import org.apache.spark.sql.catalyst.expressions._
-import org.apache.spark.sql.catalyst.plans.logical.{LocalRelation, LogicalPlan}
+import org.apache.spark.sql.catalyst.plans.logical.{Filter, LocalRelation, LogicalPlan}
 import org.apache.spark.sql.catalyst.plans.PlanTest
 import org.apache.spark.sql.catalyst.rules.RuleExecutor
 import org.apache.spark.sql.types._
@@ -36,7 +39,7 @@ class OptimizeInSuite extends PlanTest {
         NullPropagation,
         ConstantFolding,
         BooleanSimplification,
-        OptimizeIn) :: Nil
+        OptimizeIn(SimpleCatalystConf(true))) :: Nil
   }
 
   val testRelation = LocalRelation('a.int, 'b.int, 'c.int)
@@ -128,4 +131,21 @@ class OptimizeInSuite extends PlanTest {
     comparePlans(optimized, correctAnswer)
   }
 
+  test("OptimizedIn test: Use configuration.") {
+    val plan =
+      testRelation
+        .where(In(UnresolvedAttribute("a"), Seq(Literal(1), Literal(2), Literal(3))))
+        .analyze
+
+    val notOptimizedPlan = OptimizeIn(SimpleCatalystConf(true))(plan)
+    comparePlans(notOptimizedPlan, plan)
+
+    val optimizedPlan = OptimizeIn(SimpleCatalystConf(true, optimizerMinSetSize = 2))(plan)
+    optimizedPlan match {
+      case Filter(cond, _)
+        if cond.isInstanceOf[InSet] && cond.asInstanceOf[InSet].getHSet().size == 3 =>
+          // pass
+      case _ => fail("Unexpected result for OptimizedIn")
+    }
+  }
 }

sql/core/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala

@@ -54,10 +54,16 @@ object SQLConf {
 
   val OPTIMIZER_MAX_ITERATIONS = SQLConfigBuilder("spark.sql.optimizer.maxIterations")
     .internal()
-    .doc("The max number of iterations the optimizer and analyzer runs")
+    .doc("The max number of iterations the optimizer and analyzer runs.")
     .intConf
     .createWithDefault(100)
 
+  val OPTIMIZER_MIN_SET_SIZE = SQLConfigBuilder("spark.sql.optimizer.minSetSize")
+    .internal()
+    .doc("The minimum threshold of set size to be optimized.")
+    .intConf
+    .createWithDefault(10)
+
   val ALLOW_MULTIPLE_CONTEXTS = SQLConfigBuilder("spark.sql.allowMultipleContexts")
     .doc("When set to true, creating multiple SQLContexts/HiveContexts is allowed. " +
       "When set to false, only one SQLContext/HiveContext is allowed to be created " +
@@ -529,6 +535,8 @@ private[sql] class SQLConf extends Serializable with CatalystConf with Logging {
 
   def optimizerMaxIterations: Int = getConf(OPTIMIZER_MAX_ITERATIONS)
 
+  def optimizerMinSetSize: Int = getConf(OPTIMIZER_MIN_SET_SIZE)
+
   def checkpointLocation: String = getConf(CHECKPOINT_LOCATION)
 
   def filesMaxPartitionBytes: Long = getConf(FILES_MAX_PARTITION_BYTES)