Packages

c

com.nvidia.spark.rapids

GpuMergeAggregateIterator

class GpuMergeAggregateIterator extends Iterator[ColumnarBatch] with AutoCloseable with Logging

Iterator that takes another columnar batch iterator as input and emits new columnar batches that are aggregated based on the specified grouping and aggregation expressions. This iterator tries to perform a hash-based aggregation but is capable of falling back to a sort-based aggregation which can operate on data that is either larger than can be represented by a cudf column or larger than can fit in GPU memory.

The iterator starts by pulling all batches from the input iterator, performing an initial projection and aggregation on each individual batch via aggregateInputBatches(). The resulting aggregated batches are cached in memory as spillable batches. Once all input batches have been aggregated, tryMergeAggregatedBatches() is called to attempt a merge of the aggregated batches into a single batch. If this is successful then the resulting batch can be returned, otherwise buildSortFallbackIterator is used to sort the aggregated batches by the grouping keys and performs a final merge aggregation pass on the sorted batches.

Linear Supertypes
Logging, AutoCloseable, Iterator[ColumnarBatch], TraversableOnce[ColumnarBatch], GenTraversableOnce[ColumnarBatch], AnyRef, Any
Ordering
  1. Alphabetic
  2. By Inheritance
Inherited
  1. GpuMergeAggregateIterator
  2. Logging
  3. AutoCloseable
  4. Iterator
  5. TraversableOnce
  6. GenTraversableOnce
  7. AnyRef
  8. Any
  1. Hide All
  2. Show All
Visibility
  1. Public
  2. All

Instance Constructors

  1. new GpuMergeAggregateIterator(firstPassIter: Iterator[SpillableColumnarBatch], inputAttributes: Seq[Attribute], groupingExpressions: Seq[NamedExpression], aggregateExpressions: Seq[GpuAggregateExpression], aggregateAttributes: Seq[Attribute], resultExpressions: Seq[NamedExpression], modeInfo: AggregateModeInfo, metrics: GpuHashAggregateMetrics, configuredTargetBatchSize: Long, conf: SQLConf, allowNonFullyAggregatedOutput: Boolean, skipAggPassReductionRatio: Double, localInputRowsCount: LocalGpuMetric)

    firstPassIter

    iterator that has done a first aggregation pass over the input data.

    inputAttributes

    input attributes to identify the input columns from the input batches

    groupingExpressions

    expressions used for producing the grouping keys

    aggregateExpressions

    GPU aggregate expressions used to produce the aggregations

    aggregateAttributes

    attribute references to each aggregate expression

    resultExpressions

    output expression for the aggregation

    modeInfo

    identifies which aggregation modes are being used

    metrics

    metrics that will be updated during aggregation

    configuredTargetBatchSize

    user-specified value for the targeted input batch size

    allowNonFullyAggregatedOutput

    if allowed to skip third pass Agg

    skipAggPassReductionRatio

    skip if the ratio of rows after a pass is bigger than this value

    localInputRowsCount

    metric to track the number of input rows processed locally

Type Members

  1. class GroupedIterator[B >: A] extends AbstractIterator[Seq[B]] with Iterator[Seq[B]]
    Definition Classes
    Iterator

Value Members

  1. final def !=(arg0: Any): Boolean
    Definition Classes
    AnyRef → Any
  2. final def ##(): Int
    Definition Classes
    AnyRef → Any
  3. def ++[B >: ColumnarBatch](that: ⇒ GenTraversableOnce[B]): Iterator[B]
    Definition Classes
    Iterator
  4. final def ==(arg0: Any): Boolean
    Definition Classes
    AnyRef → Any
  5. def addString(b: StringBuilder): StringBuilder
    Definition Classes
    TraversableOnce
  6. def addString(b: StringBuilder, sep: String): StringBuilder
    Definition Classes
    TraversableOnce
  7. def addString(b: StringBuilder, start: String, sep: String, end: String): StringBuilder
    Definition Classes
    TraversableOnce
  8. def aggregate[B](z: ⇒ B)(seqop: (B, ColumnarBatch) ⇒ B, combop: (B, B) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  9. final def asInstanceOf[T0]: T0
    Definition Classes
    Any
  10. def buffered: BufferedIterator[ColumnarBatch]
    Definition Classes
    Iterator
  11. def clone(): AnyRef
    Attributes
    protected[lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( ... ) @native()
  12. def close(): Unit
    Definition Classes
    GpuMergeAggregateIterator → AutoCloseable
  13. def collect[B](pf: PartialFunction[ColumnarBatch, B]): Iterator[B]
    Definition Classes
    Iterator
    Annotations
    @migration
    Migration

    (Changed in version 2.8.0) collect has changed. The previous behavior can be reproduced with toSeq.

  14. def collectFirst[B](pf: PartialFunction[ColumnarBatch, B]): Option[B]
    Definition Classes
    TraversableOnce
  15. def contains(elem: Any): Boolean
    Definition Classes
    Iterator
  16. def copyToArray[B >: ColumnarBatch](xs: Array[B], start: Int, len: Int): Unit
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  17. def copyToArray[B >: ColumnarBatch](xs: Array[B]): Unit
    Definition Classes
    TraversableOnce → GenTraversableOnce
  18. def copyToArray[B >: ColumnarBatch](xs: Array[B], start: Int): Unit
    Definition Classes
    TraversableOnce → GenTraversableOnce
  19. def copyToBuffer[B >: ColumnarBatch](dest: Buffer[B]): Unit
    Definition Classes
    TraversableOnce
  20. def corresponds[B](that: GenTraversableOnce[B])(p: (ColumnarBatch, B) ⇒ Boolean): Boolean
    Definition Classes
    Iterator
  21. def count(p: (ColumnarBatch) ⇒ Boolean): Int
    Definition Classes
    TraversableOnce → GenTraversableOnce
  22. def drop(n: Int): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  23. def dropWhile(p: (ColumnarBatch) ⇒ Boolean): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  24. def duplicate: (Iterator[ColumnarBatch], Iterator[ColumnarBatch])
    Definition Classes
    Iterator
  25. final def eq(arg0: AnyRef): Boolean
    Definition Classes
    AnyRef
  26. def equals(arg0: Any): Boolean
    Definition Classes
    AnyRef → Any
  27. def exists(p: (ColumnarBatch) ⇒ Boolean): Boolean
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  28. def filter(p: (ColumnarBatch) ⇒ Boolean): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  29. def filterNot(p: (ColumnarBatch) ⇒ Boolean): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  30. def finalize(): Unit
    Attributes
    protected[lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( classOf[java.lang.Throwable] )
  31. def find(p: (ColumnarBatch) ⇒ Boolean): Option[ColumnarBatch]
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  32. def flatMap[B](f: (ColumnarBatch) ⇒ GenTraversableOnce[B]): Iterator[B]
    Definition Classes
    Iterator
  33. def fold[A1 >: ColumnarBatch](z: A1)(op: (A1, A1) ⇒ A1): A1
    Definition Classes
    TraversableOnce → GenTraversableOnce
  34. def foldLeft[B](z: B)(op: (B, ColumnarBatch) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  35. def foldRight[B](z: B)(op: (ColumnarBatch, B) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  36. def forall(p: (ColumnarBatch) ⇒ Boolean): Boolean
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  37. def foreach[U](f: (ColumnarBatch) ⇒ U): Unit
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  38. final def getClass(): Class[_]
    Definition Classes
    AnyRef → Any
    Annotations
    @native()
  39. def grouped[B >: ColumnarBatch](size: Int): GroupedIterator[B]
    Definition Classes
    Iterator
  40. def hasDefiniteSize: Boolean
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  41. def hasNext: Boolean
    Definition Classes
    GpuMergeAggregateIterator → Iterator
  42. def hashCode(): Int
    Definition Classes
    AnyRef → Any
    Annotations
    @native()
  43. def indexOf[B >: ColumnarBatch](elem: B, from: Int): Int
    Definition Classes
    Iterator
  44. def indexOf[B >: ColumnarBatch](elem: B): Int
    Definition Classes
    Iterator
  45. def indexWhere(p: (ColumnarBatch) ⇒ Boolean, from: Int): Int
    Definition Classes
    Iterator
  46. def indexWhere(p: (ColumnarBatch) ⇒ Boolean): Int
    Definition Classes
    Iterator
  47. def initializeLogIfNecessary(isInterpreter: Boolean, silent: Boolean): Boolean
    Attributes
    protected
    Definition Classes
    Logging
  48. def initializeLogIfNecessary(isInterpreter: Boolean): Unit
    Attributes
    protected
    Definition Classes
    Logging
  49. def isEmpty: Boolean
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  50. final def isInstanceOf[T0]: Boolean
    Definition Classes
    Any
  51. def isTraceEnabled(): Boolean
    Attributes
    protected
    Definition Classes
    Logging
  52. def isTraversableAgain: Boolean
    Definition Classes
    Iterator → GenTraversableOnce
  53. def length: Int
    Definition Classes
    Iterator
  54. def log: Logger
    Attributes
    protected
    Definition Classes
    Logging
  55. def logDebug(msg: ⇒ String, throwable: Throwable): Unit
    Attributes
    protected
    Definition Classes
    Logging
  56. def logDebug(msg: ⇒ String): Unit
    Attributes
    protected
    Definition Classes
    Logging
  57. def logError(msg: ⇒ String, throwable: Throwable): Unit
    Attributes
    protected
    Definition Classes
    Logging
  58. def logError(msg: ⇒ String): Unit
    Attributes
    protected
    Definition Classes
    Logging
  59. def logInfo(msg: ⇒ String, throwable: Throwable): Unit
    Attributes
    protected
    Definition Classes
    Logging
  60. def logInfo(msg: ⇒ String): Unit
    Attributes
    protected
    Definition Classes
    Logging
  61. def logName: String
    Attributes
    protected
    Definition Classes
    Logging
  62. def logTrace(msg: ⇒ String, throwable: Throwable): Unit
    Attributes
    protected
    Definition Classes
    Logging
  63. def logTrace(msg: ⇒ String): Unit
    Attributes
    protected
    Definition Classes
    Logging
  64. def logWarning(msg: ⇒ String, throwable: Throwable): Unit
    Attributes
    protected
    Definition Classes
    Logging
  65. def logWarning(msg: ⇒ String): Unit
    Attributes
    protected
    Definition Classes
    Logging
  66. def map[B](f: (ColumnarBatch) ⇒ B): Iterator[B]
    Definition Classes
    Iterator
  67. def max[B >: ColumnarBatch](implicit cmp: Ordering[B]): ColumnarBatch
    Definition Classes
    TraversableOnce → GenTraversableOnce
  68. def maxBy[B](f: (ColumnarBatch) ⇒ B)(implicit cmp: Ordering[B]): ColumnarBatch
    Definition Classes
    TraversableOnce → GenTraversableOnce
  69. def min[B >: ColumnarBatch](implicit cmp: Ordering[B]): ColumnarBatch
    Definition Classes
    TraversableOnce → GenTraversableOnce
  70. def minBy[B](f: (ColumnarBatch) ⇒ B)(implicit cmp: Ordering[B]): ColumnarBatch
    Definition Classes
    TraversableOnce → GenTraversableOnce
  71. def mkString: String
    Definition Classes
    TraversableOnce → GenTraversableOnce
  72. def mkString(sep: String): String
    Definition Classes
    TraversableOnce → GenTraversableOnce
  73. def mkString(start: String, sep: String, end: String): String
    Definition Classes
    TraversableOnce → GenTraversableOnce
  74. final def ne(arg0: AnyRef): Boolean
    Definition Classes
    AnyRef
  75. def next(): ColumnarBatch
    Definition Classes
    GpuMergeAggregateIterator → Iterator
  76. def nonEmpty: Boolean
    Definition Classes
    TraversableOnce → GenTraversableOnce
  77. final def notify(): Unit
    Definition Classes
    AnyRef
    Annotations
    @native()
  78. final def notifyAll(): Unit
    Definition Classes
    AnyRef
    Annotations
    @native()
  79. def padTo[A1 >: ColumnarBatch](len: Int, elem: A1): Iterator[A1]
    Definition Classes
    Iterator
  80. def partition(p: (ColumnarBatch) ⇒ Boolean): (Iterator[ColumnarBatch], Iterator[ColumnarBatch])
    Definition Classes
    Iterator
  81. def patch[B >: ColumnarBatch](from: Int, patchElems: Iterator[B], replaced: Int): Iterator[B]
    Definition Classes
    Iterator
  82. def product[B >: ColumnarBatch](implicit num: Numeric[B]): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  83. def reduce[A1 >: ColumnarBatch](op: (A1, A1) ⇒ A1): A1
    Definition Classes
    TraversableOnce → GenTraversableOnce
  84. def reduceLeft[B >: ColumnarBatch](op: (B, ColumnarBatch) ⇒ B): B
    Definition Classes
    TraversableOnce
  85. def reduceLeftOption[B >: ColumnarBatch](op: (B, ColumnarBatch) ⇒ B): Option[B]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  86. def reduceOption[A1 >: ColumnarBatch](op: (A1, A1) ⇒ A1): Option[A1]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  87. def reduceRight[B >: ColumnarBatch](op: (ColumnarBatch, B) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  88. def reduceRightOption[B >: ColumnarBatch](op: (ColumnarBatch, B) ⇒ B): Option[B]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  89. def reversed: List[ColumnarBatch]
    Attributes
    protected[this]
    Definition Classes
    TraversableOnce
  90. def sameElements(that: Iterator[_]): Boolean
    Definition Classes
    Iterator
  91. def scanLeft[B](z: B)(op: (B, ColumnarBatch) ⇒ B): Iterator[B]
    Definition Classes
    Iterator
  92. def scanRight[B](z: B)(op: (ColumnarBatch, B) ⇒ B): Iterator[B]
    Definition Classes
    Iterator
  93. def seq: Iterator[ColumnarBatch]
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  94. def size: Int
    Definition Classes
    TraversableOnce → GenTraversableOnce
  95. def sizeHintIfCheap: Int
    Attributes
    protected[collection]
    Definition Classes
    GenTraversableOnce
  96. def slice(from: Int, until: Int): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  97. def sliceIterator(from: Int, until: Int): Iterator[ColumnarBatch]
    Attributes
    protected
    Definition Classes
    Iterator
  98. def sliding[B >: ColumnarBatch](size: Int, step: Int): GroupedIterator[B]
    Definition Classes
    Iterator
  99. def span(p: (ColumnarBatch) ⇒ Boolean): (Iterator[ColumnarBatch], Iterator[ColumnarBatch])
    Definition Classes
    Iterator
  100. def sum[B >: ColumnarBatch](implicit num: Numeric[B]): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
  101. final def synchronized[T0](arg0: ⇒ T0): T0
    Definition Classes
    AnyRef
  102. def take(n: Int): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  103. def takeWhile(p: (ColumnarBatch) ⇒ Boolean): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  104. def to[Col[_]](implicit cbf: CanBuildFrom[Nothing, ColumnarBatch, Col[ColumnarBatch]]): Col[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  105. def toArray[B >: ColumnarBatch](implicit arg0: ClassTag[B]): Array[B]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  106. def toBuffer[B >: ColumnarBatch]: Buffer[B]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  107. def toIndexedSeq: IndexedSeq[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  108. def toIterable: Iterable[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  109. def toIterator: Iterator[ColumnarBatch]
    Definition Classes
    Iterator → GenTraversableOnce
  110. def toList: List[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  111. def toMap[T, U](implicit ev: <:<[ColumnarBatch, (T, U)]): Map[T, U]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  112. def toSeq: Seq[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  113. def toSet[B >: ColumnarBatch]: Set[B]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  114. def toStream: Stream[ColumnarBatch]
    Definition Classes
    Iterator → GenTraversableOnce
  115. def toString(): String
    Definition Classes
    Iterator → AnyRef → Any
  116. def toTraversable: Traversable[ColumnarBatch]
    Definition Classes
    Iterator → TraversableOnce → GenTraversableOnce
  117. def toVector: Vector[ColumnarBatch]
    Definition Classes
    TraversableOnce → GenTraversableOnce
  118. final def wait(): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )
  119. final def wait(arg0: Long, arg1: Int): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )
  120. final def wait(arg0: Long): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws( ... ) @native()
  121. def withFilter(p: (ColumnarBatch) ⇒ Boolean): Iterator[ColumnarBatch]
    Definition Classes
    Iterator
  122. def zip[B](that: Iterator[B]): Iterator[(ColumnarBatch, B)]
    Definition Classes
    Iterator
  123. def zipAll[B, A1 >: ColumnarBatch, B1 >: B](that: Iterator[B], thisElem: A1, thatElem: B1): Iterator[(A1, B1)]
    Definition Classes
    Iterator
  124. def zipWithIndex: Iterator[(ColumnarBatch, Int)]
    Definition Classes
    Iterator

Deprecated Value Members

  1. def /:[B](z: B)(op: (B, ColumnarBatch) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
    Annotations
    @deprecated
    Deprecated

    (Since version 2.12.10) Use foldLeft instead of /:

  2. def :\[B](z: B)(op: (ColumnarBatch, B) ⇒ B): B
    Definition Classes
    TraversableOnce → GenTraversableOnce
    Annotations
    @deprecated
    Deprecated

    (Since version 2.12.10) Use foldRight instead of :\

Inherited from Logging

Inherited from AutoCloseable

Inherited from Iterator[ColumnarBatch]

Inherited from TraversableOnce[ColumnarBatch]

Inherited from GenTraversableOnce[ColumnarBatch]

Inherited from AnyRef

Inherited from Any

Ungrouped