rolling

Documentation

Index

• type ColAggregation
  • func NewColAggregation(inputName string, needInclusiveWindow bool, typ bow.Type, ...) ColAggregation
• type ColAggregationConstruct
• type ColAggregationFunc
• type ColInterpolation
  • func NewColInterpolation(colName string, inputTypes []bow.Type, fn ColInterpolationFunc) ColInterpolation
• type ColInterpolationFunc
• type Options
• type Rolling
  • func IntervalRolling(b bow.Bow, colName string, interval int64, options Options) (Rolling, error)
• type Window
  • func (w Window) UnsetInclusive() Window

Types

type ColAggregation

type ColAggregation interface {
	// InputName returns the name of the input column.
	InputName() string
	// InputIndex returns the index of the input column.
	InputIndex() int
	// SetInputIndex sets the index of the input column.
	SetInputIndex(int)

	// OutputName returns the name of the output column.
	OutputName() string
	// RenameOutput returns a copy of the ColAggregation with a new output column name.
	RenameOutput(string) ColAggregation
	// NeedInclusiveWindow returns true if the ColAggregation needs to have inclusive windows.
	NeedInclusiveWindow() bool

	// Type returns the return type of the ColAggregation.
	Type() bow.Type
	// GetReturnType returns the return type of the ColAggregation depending on an input and an iterator type.
	GetReturnType(inputType, iteratorType bow.Type) bow.Type

	// Func returns the ColAggregationFunc of the ColAggregation.
	Func() ColAggregationFunc

	// Transformations returns the transformation functions of the ColAggregation.
	Transformations() []transformation.Func
	// SetTransformations returns a copy of the ColAggregation with new transformations functions.
	SetTransformations(...transformation.Func) ColAggregation
}

ColAggregation is a set of methods to aggregate and transform a Window.

func NewColAggregation

func NewColAggregation(inputName string, needInclusiveWindow bool, typ bow.Type, fn ColAggregationFunc) ColAggregation

NewColAggregation returns a new ColAggregation.

type ColAggregationConstruct

type ColAggregationConstruct func(colName string) ColAggregation

type ColAggregationFunc

type ColAggregationFunc func(colIndex int, w Window) (interface{}, error)

type ColInterpolation

type ColInterpolation struct {
	// contains filtered or unexported fields
}

ColInterpolation is used to interpolate a column.

func NewColInterpolation

func NewColInterpolation(colName string, inputTypes []bow.Type, fn ColInterpolationFunc) ColInterpolation

NewColInterpolation returns a new ColInterpolation.

type ColInterpolationFunc

type ColInterpolationFunc func(colIndex int, window Window, fullBow, prevRow bow.Bow) (interface{}, error)

ColInterpolationFunc is a function that take a column index, a Window, the full bow.Bow and the previous row, and provides a value at the start of the Window.

type Options

type Options struct {
	Offset    int64
	Inclusive bool
	PrevRow   bow.Bow
}

Options sets options for IntervalRolling: - Offset: interval to move the window start, can be negative. - Inclusive: sets if the window needs to be inclusive; i.e., includes the last point. - PrevRow: extra point before the window to enable better interpolation.

type Rolling

type Rolling interface {
	// Aggregate aggregates each column by using a ColAggregation.
	Aggregate(...ColAggregation) Rolling
	// Interpolate fills each window by interpolating its start if missing.
	Interpolate(...ColInterpolation) Rolling

	// NumWindows returns the total number of windows in the Bow.
	NumWindows() (int, error)
	// HasNext returns true if the next call to Next() will return a new Window.
	HasNext() bool
	// Next returns the next Window, along with its index.
	Next() (windowIndex int, window *Window, err error)

	// Bow returns the Bow from the Rolling.
	Bow() (bow.Bow, error)
}

Rolling enables processing a Bow via windows. Use Interpolate() and/or Aggregate() to transform windows. Use Next() to iterate over windows. Use Bow() to get the processed Bow.

func IntervalRolling

func IntervalRolling(b bow.Bow, colName string, interval int64, options Options) (Rolling, error)

IntervalRolling returns a new interval-based Rolling with: - b: Bow to process in windows - colName: column on which the interval is based on - interval: numeric value independent of any unit, length of the windows All windows except the last one may be empty.

type Window

type Window struct {
	Bow              bow.Bow
	FirstIndex       int
	IntervalColIndex int
	FirstValue       int64
	LastValue        int64
	IsInclusive      bool
}

Window represents an interval-based window of data with: Bow: data FirstIndex: index (across all windows) of first row in this window (-1 if none) IntervalColIndex: index of the interval column FirstValue: Window first value LastValue: Window last value IsInclusive: Window is inclusive, i.e. includes the last point at the end of the interval

func (Window) UnsetInclusive

func (w Window) UnsetInclusive() Window

UnsetInclusive returns a copy of the Window with the IsInclusive parameter set to false and with the last row sliced off. Returns the unchanged Window if the IsInclusive parameter is not set.