Packages

package statistics

Common statistics utilities.

Linear Supertypes
AnyRef, Any
Ordering
  1. Alphabetic
  2. By Inheritance
Inherited
  1. statistics
  2. AnyRef
  3. Any
  1. Hide All
  2. Show All
Visibility
  1. Public
  2. All

Type Members

  1. case class Buckets[T](bucketDelimiters: Array[T], itemsNumInBuckets: Array[Int])(implicit evidence$1: Ordering[T], evidence$2: Numeric[T]) extends Product with Serializable

    Groups incoming values by user defined buckets.

    Groups incoming values by user defined buckets. The bucket must be of type that supports ordering. User is not supposed to create the objects of this class directly, only by means of factory method.

    T

    the type of element in the bucket. Defines the type of delimiters of your buckets and the type of value that this accumulable will accumulate

    bucketDelimiters

    the delimiters of buckets as Array of Numeric. For example, given (1, 10, 100), four buckets will be created (-inf:1), [1:10), [10:100), [100:+inf)

    itemsNumInBuckets

    Array that contains the number of elements in bucket for each bucket

  2. class BucketsAccumulator[T] extends AccumulatorV2[T, Buckets[T]]

    Adapter for Spark 2.0 accumulatorV2.

    Adapter for Spark 2.0 accumulatorV2.

    T

    type of the value to accumulate, must be Numeric

  3. class CatalogStats extends AnyRef

    Accumulates the catalog metrics using Spark.

    Accumulates the catalog metrics using Spark.

    Note

    A case class is not useful because there may be multiple accumulators with different versions.

  4. case class PayloadStats(calls: LongAccumulator, kBytes: LongAccumulator, kBytesBuckets: BucketsAccumulator[Int], time: LongAccumulator) extends Product with Serializable

    Payload statistics used by both Retriever and Uploader.

  5. case class PublisherStats(compiled: LongAccumulator, compiledSize: LongAccumulator, added: LongAccumulator, changed: LongAccumulator, deleted: LongAccumulator) extends Product with Serializable

    Spark accumulators for publishing statistics.

    Spark accumulators for publishing statistics.

    compiled

    The total number of output partitions compiled, but not necessarily published.

    added

    The number of output partitions added.

    changed

    The number of output partitions changed.

    deleted

    The number of output partitions deleted.

  6. class Statistics extends ContextLogging

    Container for all the statistics accumulators created.

    Container for all the statistics accumulators created. All methods can be called in driver only.

Value Members

  1. val PayloadBucketLimits: Array[Int]
  2. def roundKBytes(bytes: Long): Int
  3. object Buckets extends Serializable
  4. object BucketsAccumulator extends Serializable
  5. object CatalogStats
  6. object PayloadStats extends Serializable
  7. object PublisherStats extends Serializable
  8. object RetrieveStats

    Retriever statistics factory.

  9. object UploadStats

    Uploader statistics factory.

Inherited from AnyRef

Inherited from Any

Ungrouped