Package dev.frostlake.functions.aggregate

Class RegrAccumulator

java.lang.Object

dev.frostlake.functions.aggregate.RegrAccumulator

All Implemented Interfaces:

AggregateFunction.Accumulator

public class RegrAccumulator
extends Object
implements AggregateFunction.Accumulator

Shared accumulator for the REGR_* linear-regression family. All nine functions are derived from the same running sums over the non-null (y, x) pairs — note Snowflake's argument order REGR_xxx(y, x), with the dependent variable y first and the independent variable x second. getResult() returns the member selected by RegrKind.

With n pairs and centered sums Sxx = Σx² − (Σx)²/n, Syy, Sxy: COUNT = n; AVGX = Σx/n; AVGY = Σy/n; SXX = Sxx; SYY = Syy; SXY = Sxy; SLOPE = Sxy/Sxx; INTERCEPT = AVGY − SLOPE·AVGX; R2 = NULL if Sxx=0, else 1 if Syy=0, else Sxy²/(Sxx·Syy) (= CORR²).

Constructor Summary

Constructors

Constructor	Description
RegrAccumulator(RegrKind kind)

Method Summary

Modifier and Type	Method	Description
void	accumulate(double y, double x)	Accumulate one non-null pair.
void	accumulate(Object value)	Folds one row's argument value into the running state.
Object	getResult()	The aggregate's result over everything accumulated so far.
RegrKind	kind()	The member this accumulator computes, which decides which side of a pair is read at all.
void	merge(AggregateFunction.Accumulator other)	Folds another accumulator's state into this one, combining two partial aggregations of the same group.
void	reset()	Clears the state back to the empty-group starting point so the accumulator can be reused.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

RegrAccumulator

public RegrAccumulator(RegrKind kind)

Method Details

kind

public RegrKind kind()

The member this accumulator computes, which decides which side of a pair is read at all.

accumulate

public void accumulate(Object value)

Description copied from interface: AggregateFunction.Accumulator

Folds one row's argument value into the running state.

Specified by:

accumulate in interface AggregateFunction.Accumulator

Parameters:

value - the aggregate argument's value for one row; SQL NULL arrives as null

accumulate

public void accumulate(double y,
 double x)

Accumulate one non-null pair. Snowflake order: y is the dependent (first) arg, x the second.

getResult

public Object getResult()

Description copied from interface: AggregateFunction.Accumulator

The aggregate's result over everything accumulated so far.

Specified by:

getResult in interface AggregateFunction.Accumulator

Returns:

the group's result, or null where the aggregate defines an empty group as NULL

reset

public void reset()

Description copied from interface: AggregateFunction.Accumulator

Clears the state back to the empty-group starting point so the accumulator can be reused.

Specified by:

reset in interface AggregateFunction.Accumulator

merge

public void merge(AggregateFunction.Accumulator other)

Description copied from interface: AggregateFunction.Accumulator

Folds another accumulator's state into this one, combining two partial aggregations of the same group.

Specified by:

merge in interface AggregateFunction.Accumulator

Parameters:

other - a partial state of the same aggregate implementation