Packages

c

org.apache.spark.sql.execution.python

WindowInPandasExec

case class WindowInPandasExec(windowExpression: Seq[NamedExpression], partitionSpec: Seq[Expression], orderSpec: Seq[SortOrder], child: SparkPlan) extends WindowExecBase with Product with Serializable

This class calculates and outputs windowed aggregates over the rows in a single partition.

This is similar to WindowExec. The main difference is that this node does not compute any window aggregation values. Instead, it computes the lower and upper bound for each window (i.e. window bounds) and pass the data and indices to Python worker to do the actual window aggregation.

It currently materializes all data associated with the same partition key and passes them to Python worker. This is not strictly necessary for sliding windows and can be improved (by possibly slicing data into overlapping chunks and stitching them together).

This class groups window expressions by their window boundaries so that window expressions with the same window boundaries can share the same window bounds. The window bounds are prepended to the data passed to the python worker.

For example, if we have: avg(v) over specifiedwindowframe(RowFrame, -5, 5), avg(v) over specifiedwindowframe(RowFrame, UnboundedPreceding, UnboundedFollowing), avg(v) over specifiedwindowframe(RowFrame, -3, 3), max(v) over specifiedwindowframe(RowFrame, -3, 3)

The python input will look like: (lower_bound_w1, upper_bound_w1, lower_bound_w3, upper_bound_w3, v)

where w1 is specifiedwindowframe(RowFrame, -5, 5) w2 is specifiedwindowframe(RowFrame, UnboundedPreceding, UnboundedFollowing) w3 is specifiedwindowframe(RowFrame, -3, 3)

Note that w2 doesn't have bound indices in the python input because it's unbounded window so it's bound indices will always be the same.

Bounded window and Unbounded window are evaluated differently in Python worker: (1) Bounded window takes the window bound indices in addition to the input columns. Unbounded window takes only input columns. (2) Bounded window evaluates the udf once per input row. Unbounded window evaluates the udf once per window partition. This is controlled by Python runner conf "pandas_window_bound_types"

The logic to compute window bounds is delegated to WindowFunctionFrame and shared with WindowExec

Note this doesn't support partial aggregation and all aggregation is computed from the entire window.

Linear Supertypes
WindowExecBase, UnaryExecNode, SparkPlan, Serializable, Serializable, Logging, QueryPlan[SparkPlan], TreeNode[SparkPlan], Product, Equals, AnyRef, Any
Ordering
  1. Alphabetic
  2. By Inheritance
Inherited
  1. WindowInPandasExec
  2. WindowExecBase
  3. UnaryExecNode
  4. SparkPlan
  5. Serializable
  6. Serializable
  7. Logging
  8. QueryPlan
  9. TreeNode
  10. Product
  11. Equals
  12. AnyRef
  13. Any
  1. Hide All
  2. Show All
Visibility
  1. Public
  2. All

Instance Constructors

  1. new WindowInPandasExec(windowExpression: Seq[NamedExpression], partitionSpec: Seq[Expression], orderSpec: Seq[SortOrder], child: SparkPlan)

Value Members

  1. lazy val allAttributes: AttributeSeq
    Definition Classes
    QueryPlan
  2. def apply(number: Int): TreeNode[_]
    Definition Classes
    TreeNode
  3. def argString(maxFields: Int): String
    Definition Classes
    TreeNode
  4. def asCode: String
    Definition Classes
    TreeNode
  5. final lazy val canonicalized: SparkPlan
    Definition Classes
    QueryPlan
    Annotations
    @transient()
  6. val child: SparkPlan
    Definition Classes
    WindowInPandasExec → UnaryExecNode
  7. final def children: Seq[SparkPlan]
    Definition Classes
    UnaryExecNode → TreeNode
  8. def clone(): SparkPlan
    Definition Classes
    TreeNode → AnyRef
  9. def collect[B](pf: PartialFunction[SparkPlan, B]): Seq[B]
    Definition Classes
    TreeNode
  10. def collectFirst[B](pf: PartialFunction[SparkPlan, B]): Option[B]
    Definition Classes
    TreeNode
  11. def collectLeaves(): Seq[SparkPlan]
    Definition Classes
    TreeNode
  12. def collectWithSubqueries[B](f: PartialFunction[SparkPlan, B]): Seq[B]
    Definition Classes
    QueryPlan
  13. def conf: SQLConf
    Definition Classes
    QueryPlan
  14. lazy val containsChild: Set[TreeNode[_]]
    Definition Classes
    TreeNode
  15. final def execute(): RDD[InternalRow]

    Returns the result of this query as an RDD[InternalRow] by delegating to doExecute after preparations.

    Returns the result of this query as an RDD[InternalRow] by delegating to doExecute after preparations.

    Concrete implementations of SparkPlan should override doExecute.

    Definition Classes
    SparkPlan
  16. final def executeBroadcast[T](): Broadcast[T]

    Returns the result of this query as a broadcast variable by delegating to doExecuteBroadcast after preparations.

    Returns the result of this query as a broadcast variable by delegating to doExecuteBroadcast after preparations.

    Concrete implementations of SparkPlan should override doExecuteBroadcast.

    Definition Classes
    SparkPlan
  17. def executeCollect(): Array[InternalRow]

    Runs this query returning the result as an array.

    Runs this query returning the result as an array.

    Definition Classes
    SparkPlan
  18. def executeCollectPublic(): Array[Row]

    Runs this query returning the result as an array, using external Row format.

    Runs this query returning the result as an array, using external Row format.

    Definition Classes
    SparkPlan
  19. final def executeColumnar(): RDD[ColumnarBatch]

    Returns the result of this query as an RDD[ColumnarBatch] by delegating to doColumnarExecute after preparations.

    Returns the result of this query as an RDD[ColumnarBatch] by delegating to doColumnarExecute after preparations.

    Concrete implementations of SparkPlan should override doColumnarExecute if supportsColumnar returns true.

    Definition Classes
    SparkPlan
  20. def executeTail(n: Int): Array[InternalRow]

    Runs this query returning the last n rows as an array.

    Runs this query returning the last n rows as an array.

    This is modeled after RDD.take but never runs any job locally on the driver.

    Definition Classes
    SparkPlan
  21. def executeTake(n: Int): Array[InternalRow]

    Runs this query returning the first n rows as an array.

    Runs this query returning the first n rows as an array.

    This is modeled after RDD.take but never runs any job locally on the driver.

    Definition Classes
    SparkPlan
  22. def executeToIterator(): Iterator[InternalRow]

    Runs this query returning the result as an iterator of InternalRow.

    Runs this query returning the result as an iterator of InternalRow.

    Definition Classes
    SparkPlan
    Note

    Triggers multiple jobs (one for each partition).

  23. final def expressions: Seq[Expression]
    Definition Classes
    QueryPlan
  24. def fastEquals(other: TreeNode[_]): Boolean
    Definition Classes
    TreeNode
  25. def find(f: (SparkPlan) ⇒ Boolean): Option[SparkPlan]
    Definition Classes
    TreeNode
  26. def flatMap[A](f: (SparkPlan) ⇒ TraversableOnce[A]): Seq[A]
    Definition Classes
    TreeNode
  27. def foreach(f: (SparkPlan) ⇒ Unit): Unit
    Definition Classes
    TreeNode
  28. def foreachUp(f: (SparkPlan) ⇒ Unit): Unit
    Definition Classes
    TreeNode
  29. def generateTreeString(depth: Int, lastChildren: Seq[Boolean], append: (String) ⇒ Unit, verbose: Boolean, prefix: String, addSuffix: Boolean, maxFields: Int, printNodeId: Boolean): Unit
    Definition Classes
    TreeNode
  30. def getTagValue[T](tag: TreeNodeTag[T]): Option[T]
    Definition Classes
    TreeNode
  31. def hashCode(): Int
    Definition Classes
    TreeNode → AnyRef → Any
  32. val id: Int
    Definition Classes
    SparkPlan
  33. def innerChildren: Seq[QueryPlan[_]]
    Definition Classes
    QueryPlan → TreeNode
  34. def inputSet: AttributeSet
    Definition Classes
    QueryPlan
  35. def logicalLink: Option[LogicalPlan]

    returns

    The logical plan this plan is linked to.

    Definition Classes
    SparkPlan
  36. def longMetric(name: String): SQLMetric

    returns

    SQLMetric for the name.

    Definition Classes
    SparkPlan
  37. def makeCopy(newArgs: Array[AnyRef]): SparkPlan

    Overridden make copy also propagates sqlContext to copied plan.

    Overridden make copy also propagates sqlContext to copied plan.

    Definition Classes
    SparkPlan → TreeNode
  38. def map[A](f: (SparkPlan) ⇒ A): Seq[A]
    Definition Classes
    TreeNode
  39. def mapChildren(f: (SparkPlan) ⇒ SparkPlan): SparkPlan
    Definition Classes
    TreeNode
  40. def mapExpressions(f: (Expression) ⇒ Expression): WindowInPandasExec.this.type
    Definition Classes
    QueryPlan
  41. def metrics: Map[String, SQLMetric]

    returns

    All metrics containing metrics of this SparkPlan.

    Definition Classes
    SparkPlan
  42. final def missingInput: AttributeSet
    Definition Classes
    QueryPlan
  43. def nodeName: String
    Definition Classes
    TreeNode
  44. def numberedTreeString: String
    Definition Classes
    TreeNode
  45. val orderSpec: Seq[SortOrder]
  46. val origin: Origin
    Definition Classes
    TreeNode
  47. def output: Seq[Attribute]
    Definition Classes
    WindowInPandasExec → QueryPlan
  48. def outputOrdering: Seq[SortOrder]

    Specifies how data is ordered in each partition.

    Specifies how data is ordered in each partition.

    Definition Classes
    WindowInPandasExec → SparkPlan
  49. def outputPartitioning: Partitioning

    Specifies how data is partitioned across different nodes in the cluster.

    Specifies how data is partitioned across different nodes in the cluster.

    Definition Classes
    WindowInPandasExec → SparkPlan
  50. lazy val outputSet: AttributeSet
    Definition Classes
    QueryPlan
    Annotations
    @transient()
  51. def p(number: Int): SparkPlan
    Definition Classes
    TreeNode
  52. val partitionSpec: Seq[Expression]
  53. final def prepare(): Unit

    Prepares this SparkPlan for execution.

    Prepares this SparkPlan for execution. It's idempotent.

    Definition Classes
    SparkPlan
  54. def prettyJson: String
    Definition Classes
    TreeNode
  55. def printSchema(): Unit
    Definition Classes
    QueryPlan
  56. def producedAttributes: AttributeSet
    Definition Classes
    QueryPlan
  57. lazy val references: AttributeSet
    Definition Classes
    QueryPlan
    Annotations
    @transient()
  58. def requiredChildDistribution: Seq[Distribution]

    Specifies the data distribution requirements of all the children for this operator.

    Specifies the data distribution requirements of all the children for this operator. By default it's UnspecifiedDistribution for each child, which means each child can have any distribution.

    If an operator overwrites this method, and specifies distribution requirements(excluding UnspecifiedDistribution and BroadcastDistribution) for more than one child, Spark guarantees that the outputs of these children will have same number of partitions, so that the operator can safely zip partitions of these children's result RDDs. Some operators can leverage this guarantee to satisfy some interesting requirement, e.g., non-broadcast joins can specify HashClusteredDistribution(a,b) for its left child, and specify HashClusteredDistribution(c,d) for its right child, then it's guaranteed that left and right child are co-partitioned by a,b/c,d, which means tuples of same value are in the partitions of same index, e.g., (a=1,b=2) and (c=1,d=2) are both in the second partition of left and right child.

    Definition Classes
    WindowInPandasExec → SparkPlan
  59. def requiredChildOrdering: Seq[Seq[SortOrder]]

    Specifies sort order for each partition requirements on the input data for this operator.

    Specifies sort order for each partition requirements on the input data for this operator.

    Definition Classes
    WindowInPandasExec → SparkPlan
  60. def resetMetrics(): Unit

    Resets all the metrics.

    Resets all the metrics.

    Definition Classes
    SparkPlan
  61. final def sameResult(other: SparkPlan): Boolean
    Definition Classes
    QueryPlan
  62. lazy val schema: StructType
    Definition Classes
    QueryPlan
  63. def schemaString: String
    Definition Classes
    QueryPlan
  64. final def semanticHash(): Int
    Definition Classes
    QueryPlan
  65. def setLogicalLink(logicalPlan: LogicalPlan): Unit

    Set logical plan link recursively if unset.

    Set logical plan link recursively if unset.

    Definition Classes
    SparkPlan
  66. def setTagValue[T](tag: TreeNodeTag[T], value: T): Unit
    Definition Classes
    TreeNode
  67. def simpleString(maxFields: Int): String
    Definition Classes
    QueryPlan → TreeNode
  68. def simpleStringWithNodeId(): String
    Definition Classes
    QueryPlan → TreeNode
  69. final val sqlContext: SQLContext

    A handle to the SQL Context that was used to create this plan.

    A handle to the SQL Context that was used to create this plan. Since many operators need access to the sqlContext for RDD operations or configuration this field is automatically populated by the query planning infrastructure.

    Definition Classes
    SparkPlan
  70. def subqueries: Seq[SparkPlan]
    Definition Classes
    QueryPlan
  71. def subqueriesAll: Seq[SparkPlan]
    Definition Classes
    QueryPlan
  72. def supportsColumnar: Boolean

    Return true if this stage of the plan supports columnar execution.

    Return true if this stage of the plan supports columnar execution.

    Definition Classes
    SparkPlan
  73. def toJSON: String
    Definition Classes
    TreeNode
  74. def toString(): String
    Definition Classes
    TreeNode → AnyRef → Any
  75. def transform(rule: PartialFunction[SparkPlan, SparkPlan]): SparkPlan
    Definition Classes
    TreeNode
  76. def transformAllExpressions(rule: PartialFunction[Expression, Expression]): WindowInPandasExec.this.type
    Definition Classes
    QueryPlan
  77. def transformDown(rule: PartialFunction[SparkPlan, SparkPlan]): SparkPlan
    Definition Classes
    TreeNode
  78. def transformExpressions(rule: PartialFunction[Expression, Expression]): WindowInPandasExec.this.type
    Definition Classes
    QueryPlan
  79. def transformExpressionsDown(rule: PartialFunction[Expression, Expression]): WindowInPandasExec.this.type
    Definition Classes
    QueryPlan
  80. def transformExpressionsUp(rule: PartialFunction[Expression, Expression]): WindowInPandasExec.this.type
    Definition Classes
    QueryPlan
  81. def transformUp(rule: PartialFunction[SparkPlan, SparkPlan]): SparkPlan
    Definition Classes
    TreeNode
  82. def treeString(append: (String) ⇒ Unit, verbose: Boolean, addSuffix: Boolean, maxFields: Int, printOperatorId: Boolean): Unit
    Definition Classes
    TreeNode
  83. final def treeString(verbose: Boolean, addSuffix: Boolean, maxFields: Int, printOperatorId: Boolean): String
    Definition Classes
    TreeNode
  84. final def treeString: String
    Definition Classes
    TreeNode
  85. def unsetTagValue[T](tag: TreeNodeTag[T]): Unit
    Definition Classes
    TreeNode
  86. def vectorTypes: Option[Seq[String]]

    The exact java types of the columns that are output in columnar processing mode.

    The exact java types of the columns that are output in columnar processing mode. This is a performance optimization for code generation and is optional.

    Definition Classes
    SparkPlan
  87. def verboseString(maxFields: Int): String
    Definition Classes
    QueryPlan → TreeNode
  88. def verboseStringWithOperatorId(): String
    Definition Classes
    UnaryExecNode → QueryPlan
  89. def verboseStringWithSuffix(maxFields: Int): String
    Definition Classes
    TreeNode
  90. val windowExpression: Seq[NamedExpression]
  91. def withNewChildren(newChildren: Seq[SparkPlan]): SparkPlan
    Definition Classes
    TreeNode