標籤:spark sql catalyst sql hive spark
前幾篇文章介紹了Spark SQL的Catalyst的SqlParser,和Analyzer,本來打算直接寫Optimizer的,但是發現忘記介紹TreeNode這個Catalyst的核心概念,介紹這個可以更好的理解Optimizer是如何對Analyzed Logical Plan進行最佳化的產生Optimized Logical Plan,本文就將TreeNode基本架構進行解釋。
一、TreeNode類型 TreeNode Library是Catalyst的核心類庫,文法樹的構建都是由一個個TreeNode組成。TreeNode本身是一個BaseType <: TreeNode[BaseType] 的類型,並且實現了Product這個trait,這樣可以存放異構的元素了。
TreeNode有三種形態:
BinaryNode、
UnaryNode、
Leaf Node.
在Catalyst裡,這些Node都是繼承自Logical Plan,可以說每一個TreeNode節點就是一個Logical Plan。除了Expression(是直接繼承自TreeNode)
主要繼承關係類圖如下:
1、BinaryNode
二元節點,即有左右孩子的二叉節點
[[TreeNode]] that has two children, [[left]] and [[right]].trait BinaryNode[BaseType <: TreeNode[BaseType]] { def left: BaseType def right: BaseType def children = Seq(left, right)}abstract class BinaryNode extends LogicalPlan with trees.BinaryNode[LogicalPlan] { self: Product =>} 節點定義比較簡單,左孩子,右孩子都是BaseType。 children是一個Seq(left, right)
下面列出主要繼承二元節點的類,可以當查詢手冊用 :)
這裡提示下平常常用的二元節點:Join和Union
2、UnaryNode
一元節點,即只有一個孩子節點
A [[TreeNode]] with a single [[child]].trait UnaryNode[BaseType <: TreeNode[BaseType]] { def child: BaseType def children = child :: Nil}abstract class UnaryNode extends LogicalPlan with trees.UnaryNode[LogicalPlan] { self: Product =>}下面列出主要繼承一元節點的類,可以當查詢手冊用 :)
常用的二元節點有,Project,Subquery,Filter,Limit ...等
3、Leaf Node
葉子節點,沒有孩子節點的節點。
A [[TreeNode]] with no children.trait LeafNode[BaseType <: TreeNode[BaseType]] { def children = Nil}abstract class LeafNode extends LogicalPlan with trees.LeafNode[LogicalPlan] { self: Product => // Leaf nodes by definition cannot reference any input attributes. override def references = Set.empty}下面列出主要繼承葉子節點的類,可以當查詢手冊用 :)
提示常用的葉子節點: Command類系列,一些Funtion函數,以及Unresolved Relation...etc.
二、TreeNode 核心方法 簡單介紹一個TreeNode這個類的屬性和方法
currentId
一顆樹裡的TreeNode有個唯一的id,類型是java.util.concurrent.atomic.AtomicLong原子類型。
private val currentId = new java.util.concurrent.atomic.AtomicLong protected def nextId() = currentId.getAndIncrement()
sameInstance
判斷2個執行個體是否是同一個的時候,只需要判斷TreeNode的id。
def sameInstance(other: TreeNode[_]): Boolean = { this.id == other.id }
fastEquals,更常用的一個快捷的判定方法,沒有重寫Object.Equals,這樣防止scala編譯器產生case class equals 方法
def fastEquals(other: TreeNode[_]): Boolean = { sameInstance(other) || this == other } map,flatMap,collect都是遞迴的對子節點進行應用PartialFunction,其它方法還有很多,篇幅有限這裡不一一描述了。
2.1、核心方法 transform 方法 transform該方法接受一個PartialFunction,就是就是前一篇文章Analyzer裡提到的Batch裡面的Rule。
是會將Rule迭代應用到該節點的所有子節點,最後返回這個節點的副本(一個和當前節點不同的節點,後面會介紹,其實就是利用反射來返回一個修改後的節點)。
如果rule沒有對一個節點進行PartialFunction的操作,就返回這個節點本身。
來看一個例子:
object GlobalAggregates extends Rule[LogicalPlan] { def apply(plan: LogicalPlan): LogicalPlan = plan transform { //apply方法這裡調用了logical plan(TreeNode) 的transform方法來應用一個PartialFunction。 case Project(projectList, child) if containsAggregates(projectList) => Aggregate(Nil, projectList, child) } def containsAggregates(exprs: Seq[Expression]): Boolean = { exprs.foreach(_.foreach { case agg: AggregateExpression => return true case _ => }) false } } 這個方法真正的調用是transformChildrenDown,這裡提到了用先序遍曆來對子節點進行遞迴的Rule應用。
如果在對當前節點應用rule成功,修改後的節點afterRule,來對其children節點進行rule的應用。
transformDown方法:
/** * Returns a copy of this node where `rule` has been recursively applied to it and all of its * children (pre-order). When `rule` does not apply to a given node it is left unchanged. * @param rule the function used to transform this nodes children */ def transformDown(rule: PartialFunction[BaseType, BaseType]): BaseType = { val afterRule = rule.applyOrElse(this, identity[BaseType]) // Check if unchanged and then possibly return old copy to avoid gc churn. if (this fastEquals afterRule) { transformChildrenDown(rule) //修改前節點this.transformChildrenDown(rule) } else { afterRule.transformChildrenDown(rule) //修改後節點進行transformChildrenDown } } 最重要的方法transformChildrenDown:
對children節點進行遞迴的調用PartialFunction,利用最終返回的newArgs來產生一個新的節點,這裡調用了makeCopy()來產生節點。
transformChildrenDown方法:
/** * Returns a copy of this node where `rule` has been recursively applied to all the children of * this node. When `rule` does not apply to a given node it is left unchanged. * @param rule the function used to transform this nodes children */ def transformChildrenDown(rule: PartialFunction[BaseType, BaseType]): this.type = { var changed = false val newArgs = productIterator.map { case arg: TreeNode[_] if children contains arg => val newChild = arg.asInstanceOf[BaseType].transformDown(rule) //遞迴子節點應用rule if (!(newChild fastEquals arg)) { changed = true newChild } else { arg } case Some(arg: TreeNode[_]) if children contains arg => val newChild = arg.asInstanceOf[BaseType].transformDown(rule) if (!(newChild fastEquals arg)) { changed = true Some(newChild) } else { Some(arg) } case m: Map[_,_] => m case args: Traversable[_] => args.map { case arg: TreeNode[_] if children contains arg => val newChild = arg.asInstanceOf[BaseType].transformDown(rule) if (!(newChild fastEquals arg)) { changed = true newChild } else { arg } case other => other } case nonChild: AnyRef => nonChild case null => null }.toArray if (changed) makeCopy(newArgs) else this //根據作用結果返回的newArgs數組,反射產生新的節點副本。 }
makeCopy方法,反射產生節點副本
/** * Creates a copy of this type of tree node after a transformation. * Must be overridden by child classes that have constructor arguments * that are not present in the productIterator. * @param newArgs the new product arguments. */ def makeCopy(newArgs: Array[AnyRef]): this.type = attachTree(this, "makeCopy") { try { val defaultCtor = getClass.getConstructors.head //反射擷取預設建構函式的第一個 if (otherCopyArgs.isEmpty) { defaultCtor.newInstance(newArgs: _*).asInstanceOf[this.type] //反射產生當前節點類型的節點 } else { defaultCtor.newInstance((newArgs ++ otherCopyArgs).toArray: _*).asInstanceOf[this.type] //如果還有其它參數,++ } } catch { case e: java.lang.IllegalArgumentException => throw new TreeNodeException( this, s"Failed to copy node. Is otherCopyArgs specified correctly for $nodeName? " + s"Exception message: ${e.getMessage}.") } }
三、TreeNode執行個體 現在準備從一段sql來出發,畫一下這個spark sql的整體樹的transformation。 SELECT * FROM (SELECT * FROM src) a join (select * from src)b on a.key=b.key 首先,我們先執行一下,在控制台裡看一下產生的計劃:
<span style="font-size:12px;">sbt/sbt hive/consoleUsing /usr/java/default as default JAVA_HOME.Note, this will be overridden by -java-home if it is set.[info] Loading project definition from /app/hadoop/shengli/spark/project/project[info] Loading project definition from /app/hadoop/shengli/spark/project[info] Set current project to root (in build file:/app/hadoop/shengli/spark/)[info] Starting scala interpreter...[info] import org.apache.spark.sql.catalyst.analysis._import org.apache.spark.sql.catalyst.dsl._import org.apache.spark.sql.catalyst.errors._import org.apache.spark.sql.catalyst.expressions._import org.apache.spark.sql.catalyst.plans.logical._import org.apache.spark.sql.catalyst.rules._import org.apache.spark.sql.catalyst.types._import org.apache.spark.sql.catalyst.util._import org.apache.spark.sql.executionimport org.apache.spark.sql.hive._import org.apache.spark.sql.hive.test.TestHive._import org.apache.spark.sql.parquet.ParquetTestData scala> val query = sql("SELECT * FROM (SELECT * FROM src) a join (select * from src)b on a.key=b.key")</span>3.1、UnResolve Logical Plan 第一步產生UnResolve Logical Plan 如下:
scala> query.queryExecution.logicalres0: org.apache.spark.sql.catalyst.plans.logical.LogicalPlan = Project [*] Join Inner, Some(('a.key = 'b.key)) Subquery a Project [*] UnresolvedRelation None, src, None Subquery b Project [*] UnresolvedRelation None, src, None 如果畫成樹是這樣的,僅個人理解: 我將一開始介紹的三種Node分別用綠色UnaryNode,紅色Binary Node 和 藍色 LeafNode 來表示。
3.2、Analyzed Logical Plan Analyzer會將允用Batch的Rules來對Unresolved Logical Plan Tree 進行rule應用,這裡用來EliminateAnalysisOperators將Subquery給消除掉,Batch("Resolution將Atrribute和Relation給Resolve了,Analyzed Logical Plan Tree如:
3.3、Optimized Plan 我把Catalyst裡的Optimizer戲稱為Spark SQL的最佳化大師,因為整個Spark SQL的最佳化都是在這裡進行的,後面會有文章來講解Optimizer。 在這裡,最佳化的不明顯,因為SQL本身不複雜
scala> query.queryExecution.optimizedPlanres3: org.apache.spark.sql.catalyst.plans.logical.LogicalPlan = Project [key#0,value#1,key#2,value#3] Join Inner, Some((key#0 = key#2)) MetastoreRelation default, src, None MetastoreRelation default, src, None
產生的樹如:
3.4、executedPlan 最後一步是最終產生的物理執行計畫,裡面涉及到了Hive的TableScan,涉及到了HashJoin操作,還涉及到了Exchange,Exchange涉及到了Shuffle和Partition操作。
scala> query.queryExecution.executedPlanres4: org.apache.spark.sql.execution.SparkPlan = Project [key#0:0,value#1:1,key#2:2,value#3:3] HashJoin [key#0], [key#2], BuildRight Exchange (HashPartitioning [key#0:0], 150) HiveTableScan [key#0,value#1], (MetastoreRelation default, src, None), None Exchange (HashPartitioning [key#2:0], 150) HiveTableScan [key#2,value#3], (MetastoreRelation default, src, None), None
產生的物理執行樹
四、總結: 本文介紹了Spark SQL的Catalyst架構核心TreeNode類庫,繪製了TreeNode繼承關係的類圖,瞭解了TreeNode這個類在Catalyst所起到的作用。文法樹中的Logical Plan均派生自TreeNode,並且Logical Plan派生出TreeNode的三種形態,即Binary Node, Unary Node, Leaft Node。 正式這幾種節點,組成了Spark SQl的Catalyst的文法樹。
TreeNode的transform方法是核心的方法,它接受一個rule,會對當前節點的孩子節點進行遞迴的調用rule,最後會返回一個TreeNode的copy,這種操作就是transformation,貫穿了Spark SQL執行的幾個核心階段,如Analyze,Optimize階段。
最後用一個實際的例子,展示出來Spark SQL的執行樹產生流程。
我目前的理解就是這些,如果分析不到位的地方,請大家多多指正。
——EOF——原創文章,轉載請註明出自:http://blog.csdn.net/oopsoom/article/details/38084079