sort merge join for spark sql

adrian-wang · adrian-wang · commit 880d8e980535 · 2015-04-02T21:57:17.000-07:00
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/physical/partitioning.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/physical/partitioning.scala
@@ -75,6 +75,21 @@ case class OrderedDistribution(ordering: Seq[SortOrder]) extends Distribution {
   def clustering: Set[Expression] = ordering.map(_.child).toSet
 }
 
+/**
+ * Represents data where tuples have been ordered according to the `clustering`
+ * [[Expression Expressions]].  This is a strictly stronger guarantee than
+ * [[ClusteredDistribution]] as this will ensure that tuples in a single partition are sorted
+ * by the expressions.
+ */
+case class ClusteredOrderedDistribution(clustering: Seq[Expression])
+  extends Distribution {
+  require(
+    clustering != Nil,
+    "The clustering expressions of a ClusteredOrderedDistribution should not be Nil. " +
+      "An AllTuples should be used to represent a distribution that only has " +
+      "a single partition.")
+}
+
 sealed trait Partitioning {
   /** Returns the number of partitions that the data is split across */
   val numPartitions: Int
@@ -162,6 +177,40 @@ case class HashPartitioning(expressions: Seq[Expression], numPartitions: Int)
     throw new TreeNodeException(this, s"No function to evaluate expression. type: ${this.nodeName}")
 }
 
+/**
+ * Represents a partitioning where rows are split up across partitions based on the hash
+ * of `expressions`.  All rows where `expressions` evaluate to the same values are guaranteed to be
+ * in the same partition. And rows within the same partition are sorted by the expressions.
+ */
+case class HashSortedPartitioning(expressions: Seq[Expression], numPartitions: Int)
+  extends Expression
+  with Partitioning {
+
+  override def children = expressions
+  override def nullable = false
+  override def dataType = IntegerType
+
+  private[this] lazy val clusteringSet = expressions.toSet
+
+  override def satisfies(required: Distribution): Boolean = required match {
+    case UnspecifiedDistribution => true
+    case ClusteredOrderedDistribution(requiredClustering) =>
+      clusteringSet.subsetOf(requiredClustering.toSet)
+    case ClusteredDistribution(requiredClustering) =>
+      clusteringSet.subsetOf(requiredClustering.toSet)
+    case _ => false
+  }
+
+  override def compatibleWith(other: Partitioning) = other match {
+    case BroadcastPartitioning => true
+    case h: HashSortedPartitioning if h == this => true
+    case _ => false
+  }
+
+  override def eval(input: Row = null): EvaluatedType =
+    throw new TreeNodeException(this, s"No function to evaluate expression. type: ${this.nodeName}")
+}
+
 /**
  * Represents a partitioning where rows are split across partitions based on some total ordering of
  * the expressions specified in `ordering`.  When data is partitioned in this manner the following
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
@@ -27,6 +27,7 @@ private[spark] object SQLConf {
   val COLUMN_BATCH_SIZE = "spark.sql.inMemoryColumnarStorage.batchSize"
   val IN_MEMORY_PARTITION_PRUNING = "spark.sql.inMemoryColumnarStorage.partitionPruning"
   val AUTO_BROADCASTJOIN_THRESHOLD = "spark.sql.autoBroadcastJoinThreshold"
+  val AUTO_SORTMERGEJOIN = "spark.sql.autoSortMergeJoin"
   val DEFAULT_SIZE_IN_BYTES = "spark.sql.defaultSizeInBytes"
   val SHUFFLE_PARTITIONS = "spark.sql.shuffle.partitions"
   val CODEGEN_ENABLED = "spark.sql.codegen"
@@ -143,6 +144,12 @@ private[sql] class SQLConf extends Serializable {
   private[spark] def autoBroadcastJoinThreshold: Int =
     getConf(AUTO_BROADCASTJOIN_THRESHOLD, (10 * 1024 * 1024).toString).toInt
 
+  /**
+   * By default it will choose sort merge join.
+   */
+  private[spark] def autoSortMergeJoin: Boolean =
+    getConf(AUTO_SORTMERGEJOIN, true.toString).toBoolean
+
   /**
    * The default size in bytes to assign to a logical operator's estimation statistics.  By default,
    * it is set to a larger value than `autoBroadcastJoinThreshold`, hence any logical operator
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/Exchange.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/Exchange.scala
@@ -19,12 +19,11 @@ package org.apache.spark.sql.execution
 
 import org.apache.spark.annotation.DeveloperApi
 import org.apache.spark.shuffle.sort.SortShuffleManager
-import org.apache.spark.sql.catalyst.expressions
 import org.apache.spark.{SparkEnv, HashPartitioner, RangePartitioner, SparkConf}
 import org.apache.spark.rdd.{RDD, ShuffledRDD}
 import org.apache.spark.sql.{SQLContext, Row}
 import org.apache.spark.sql.catalyst.errors.attachTree
-import org.apache.spark.sql.catalyst.expressions.{Attribute, RowOrdering}
+import org.apache.spark.sql.catalyst.expressions._
 import org.apache.spark.sql.catalyst.plans.physical._
 import org.apache.spark.sql.catalyst.rules.Rule
 import org.apache.spark.util.MutablePair
@@ -73,6 +72,26 @@ case class Exchange(newPartitioning: Partitioning, child: SparkPlan) extends Una
         shuffled.setSerializer(new SparkSqlSerializer(new SparkConf(false)))
         shuffled.map(_._2)
 
+      case HashSortedPartitioning(expressions, numPartitions) =>
+        val rdd = if (sortBasedShuffleOn && numPartitions > bypassMergeThreshold) {
+          child.execute().mapPartitions { iter =>
+            val hashExpressions = newMutableProjection(expressions, child.output)()
+            iter.map(r => (hashExpressions(r).copy(), r.copy()))
+          }
+        } else {
+          child.execute().mapPartitions { iter =>
+            val hashExpressions = newMutableProjection(expressions, child.output)()
+            val mutablePair = new MutablePair[Row, Row]()
+            iter.map(r => mutablePair.update(hashExpressions(r), r))
+          }
+        }
+        val sortingExpressions = expressions.map(s => new SortOrder(s, Ascending))
+        implicit val ordering = new RowOrdering(sortingExpressions, child.output)
+        val part = new HashPartitioner(numPartitions)
+        val shuffled = new ShuffledRDD[Row, Row, Row](rdd, part).setKeyOrdering(ordering)
+        shuffled.setSerializer(new SparkSqlSerializer(new SparkConf(false)))
+        shuffled.map(_._2)
+
       case RangePartitioning(sortingExpressions, numPartitions) =>
         val rdd = if (sortBasedShuffleOn) {
           child.execute().mapPartitions { iter => iter.map(row => (row.copy(), null))}
@@ -173,6 +192,8 @@ private[sql] case class AddExchange(sqlContext: SQLContext) extends Rule[SparkPl
             addExchangeIfNecessary(SinglePartition, child)
           case (ClusteredDistribution(clustering), child) =>
             addExchangeIfNecessary(HashPartitioning(clustering, numPartitions), child)
+          case (ClusteredOrderedDistribution(clustering), child) =>
+            addExchangeIfNecessary(HashSortedPartitioning(clustering, numPartitions), child)
           case (OrderedDistribution(ordering), child) =>
             addExchangeIfNecessary(RangePartitioning(ordering, numPartitions), child)
           case (UnspecifiedDistribution, child) => child
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala
@@ -90,6 +90,13 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
            left.statistics.sizeInBytes <= sqlContext.conf.autoBroadcastJoinThreshold =>
           makeBroadcastHashJoin(leftKeys, rightKeys, left, right, condition, joins.BuildLeft)
 
+      // for now let's support inner join first, then add outer join
+      case ExtractEquiJoinKeys(Inner, leftKeys, rightKeys, condition, left, right)
+        if sqlContext.conf.autoSortMergeJoin =>
+        val mergeJoin =
+          joins.SortMergeJoin(leftKeys, rightKeys, Inner, planLater(left), planLater(right))
+        condition.map(Filter(_, mergeJoin)).getOrElse(mergeJoin) :: Nil
+
       case ExtractEquiJoinKeys(Inner, leftKeys, rightKeys, condition, left, right) =>
         val buildSide =
           if (right.statistics.sizeInBytes <= left.statistics.sizeInBytes) {
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/joins/SortMergeJoin.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/joins/SortMergeJoin.scala
@@ -0,0 +1,145 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.sql.execution.joins
+
+import org.apache.spark.annotation.DeveloperApi
+import org.apache.spark.sql.Row
+import org.apache.spark.sql.catalyst.expressions._
+import org.apache.spark.sql.catalyst.plans._
+import org.apache.spark.sql.catalyst.plans.physical.{ClusteredOrderedDistribution, Partitioning}
+import org.apache.spark.sql.execution.{BinaryNode, SparkPlan}
+import org.apache.spark.util.collection.CompactBuffer
+
+/**
+ * :: DeveloperApi ::
+ * Performs an sort merge join of two child relations.
+ */
+@DeveloperApi
+case class SortMergeJoin(
+    leftKeys: Seq[Expression],
+    rightKeys: Seq[Expression],
+    joinType: JoinType,
+    left: SparkPlan,
+    right: SparkPlan) extends BinaryNode {
+
+  override def output: Seq[Attribute] = left.output ++ right.output
+
+  override def outputPartitioning: Partitioning = left.outputPartitioning
+
+  override def requiredChildDistribution: Seq[ClusteredOrderedDistribution] =
+    ClusteredOrderedDistribution(leftKeys) :: ClusteredOrderedDistribution(rightKeys) :: Nil
+
+  private val orders: Seq[SortOrder] = leftKeys.map(s => SortOrder(s, Ascending))
+  private val ordering: RowOrdering = new RowOrdering(orders, left.output)
+
+  @transient protected lazy val leftKeyGenerator = newProjection(leftKeys, left.output)
+  @transient protected lazy val rightKeyGenerator = newProjection(rightKeys, right.output)
+
+  override def execute() = {
+    val leftResults = left.execute().map(_.copy())
+    val rightResults = right.execute().map(_.copy())
+
+    leftResults.zipPartitions(rightResults) { (leftIter, rightIter) =>
+      new Iterator[Row] {
+        // Mutable per row objects.
+        private[this] val joinRow = new JoinedRow5
+        private[this] var leftElement: Row = _
+        private[this] var rightElement: Row = _
+        private[this] var leftKey: Row = _
+        private[this] var rightKey: Row = _
+        private[this] var read: Boolean = false
+        private[this] var currentlMatches: CompactBuffer[Row] = _
+        private[this] var currentrMatches: CompactBuffer[Row] = _
+        private[this] var currentlPosition: Int = -1
+        private[this] var currentrPosition: Int = -1
+
+        override final def hasNext: Boolean =
+          (currentlPosition != -1 && currentlPosition < currentlMatches.size) ||
+            (leftIter.hasNext && rightIter.hasNext && nextMatchingPair)
+
+        override final def next(): Row = {
+          val joinedRow =
+            joinRow(currentlMatches(currentlPosition), currentrMatches(currentrPosition))
+          currentrPosition += 1
+          if (currentrPosition >= currentrMatches.size) {
+            currentlPosition += 1
+            currentrPosition = 0
+          }
+          joinedRow
+        }
+
+        /**
+         * Searches the left/right iterator for the next rows that matches.
+         *
+         * @return true if the search is successful, and false if the left/right iterator runs out
+         *         of tuples.
+         */
+        private def nextMatchingPair(): Boolean = {
+          currentlPosition = -1
+          currentlMatches = null
+          if (rightElement == null) {
+            rightElement = rightIter.next()
+            rightKey = rightKeyGenerator(rightElement)
+          }
+          while (currentlMatches == null && leftIter.hasNext) {
+            if (!read) {
+              leftElement = leftIter.next()
+              leftKey = leftKeyGenerator(leftElement)
+            }
+            while (ordering.compare(leftKey, rightKey) > 0 && rightIter.hasNext) {
+              rightElement = rightIter.next()
+              rightKey = rightKeyGenerator(rightElement)
+            }
+            currentrMatches = new CompactBuffer[Row]()
+            while (ordering.compare(leftKey, rightKey) == 0 && rightIter.hasNext) {
+              currentrMatches += rightElement
+              rightElement = rightIter.next()
+              rightKey = rightKeyGenerator(rightElement)
+            }
+            if (ordering.compare(leftKey, rightKey) == 0) {
+              currentrMatches += rightElement
+            }
+            if (currentrMatches.size > 0) {
+              // there exists rows match in right table, should search left table
+              currentlMatches = new CompactBuffer[Row]()
+              val leftMatch = leftKey.copy()
+              while (ordering.compare(leftKey, leftMatch) == 0 && leftIter.hasNext) {
+                currentlMatches += leftElement
+                leftElement = leftIter.next()
+                leftKey = leftKeyGenerator(leftElement)
+              }
+              if (ordering.compare(leftKey, leftMatch) == 0) {
+                currentlMatches += leftElement
+              } else {
+                read = true
+              }
+            }
+          }
+
+          if (currentlMatches == null) {
+            false
+          } else {
+            currentlPosition = 0
+            currentrPosition = 0
+            true
+          }
+        }
+      }
+    }
+  }
+}
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/JoinSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/JoinSuite.scala
@@ -51,6 +51,7 @@ class JoinSuite extends QueryTest with BeforeAndAfterEach {
       case j: CartesianProduct => j
       case j: BroadcastNestedLoopJoin => j
       case j: BroadcastLeftSemiJoinHash => j
+      case j: SortMergeJoin => j
     }
 
     assert(operators.size === 1)
@@ -75,9 +76,9 @@ class JoinSuite extends QueryTest with BeforeAndAfterEach {
       ("SELECT * FROM testData FULL OUTER JOIN testData2 WHERE key = 2", classOf[CartesianProduct]),
       ("SELECT * FROM testData JOIN testData2 WHERE key > a", classOf[CartesianProduct]),
       ("SELECT * FROM testData FULL OUTER JOIN testData2 WHERE key > a", classOf[CartesianProduct]),
-      ("SELECT * FROM testData JOIN testData2 ON key = a", classOf[ShuffledHashJoin]),
-      ("SELECT * FROM testData JOIN testData2 ON key = a and key = 2", classOf[ShuffledHashJoin]),
-      ("SELECT * FROM testData JOIN testData2 ON key = a where key = 2", classOf[ShuffledHashJoin]),
+      ("SELECT * FROM testData JOIN testData2 ON key = a", classOf[SortMergeJoin]),
+      ("SELECT * FROM testData JOIN testData2 ON key = a and key = 2", classOf[SortMergeJoin]),
+      ("SELECT * FROM testData JOIN testData2 ON key = a where key = 2", classOf[SortMergeJoin]),
       ("SELECT * FROM testData LEFT JOIN testData2 ON key = a", classOf[HashOuterJoin]),
       ("SELECT * FROM testData RIGHT JOIN testData2 ON key = a where key = 2",
         classOf[HashOuterJoin]),
