water.rapids.ast.prims.mungers

Class AstGroup

java.lang.Object

water.Iced<T>

water.rapids.ast.AstRoot<T>

water.rapids.ast.AstPrimitive

water.rapids.ast.prims.mungers.AstGroup

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, Freezable

public class AstGroup
extends AstPrimitive

GroupBy Group the rows of 'data' by unique combinations of '[group-by-cols]'. Apply function 'fcn' to a Frame for each group, with a single column argument, and a NA-handling flag. Sets of tuples {fun,col,na} are allowed.

'fcn' must be a one of a small set of functions, all reductions, and 'GB' returns a row per unique group, with the first columns being the grouping columns, and the last column(s) the reduction result(s).

Grouping can be performed on both numeric and string columns, but aggregation functions are limited to the ones listed in AstGroup.FCN.

The returned column(s).

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	AstGroup.AGG
static class	AstGroup.FCN
static class	AstGroup.G
static class	AstGroup.GBTaskAggsPerMap
static class	AstGroup.MedianResult
static class	AstGroup.NAHandling

Constructor Summary

Constructors

Constructor and Description
AstGroup()
AstGroup(boolean perNodeAggregates)

Method Summary

Modifier and Type	Method and Description
static AstGroup.AGG[]	aggNRows()
ValFrame	apply(Env env, Env.StackHelp stk, AstRoot[] asts) Primary method to invoke this function, passing all the parameters as the `asts` list.
java.lang.String[]	args() List of argument names.
static Frame	buildOutput(int[] gbCols, int noutCols, Frame fr, java.lang.String[] fcnames, int ngrps, MRTask mrfill)
static AstNumList	check(long dstX, AstRoot ast)
static IcedHashSet<AstGroup.G>	doGroups(Frame fr, int[] gbCols, AstGroup.AGG[] aggs)
static IcedHashSet<AstGroup.G>	doGroups(Frame fr, int[] gbCols, byte[] gbColsTypes, AstGroup.AGG[] aggs, boolean hasMedian, boolean perNodeAggregates)
int	nargs() Number of function's arguments + 1.
ValFrame	performGroupingWithAggregations(Frame fr, int[] gbCols, AstGroup.AGG[] aggs) Performs grouping and aggregations on given data.
java.lang.String	str() String representation of this Ast object in the Rapids language.

Methods inherited from class water.rapids.ast.AstPrimitive

description, example, exec

Methods inherited from class water.rapids.ast.AstRoot

toString

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Constructor Detail

AstGroup

public AstGroup()

AstGroup

public AstGroup(boolean perNodeAggregates)

Method Detail

nargs

public int nargs()

Description copied from class: AstPrimitive

Number of function's arguments + 1. Thus, a binary operator like '+' should be declared with 3 nargs: ["+", lhs, rhs]. For variable-argument functions this method should return -1.

Specified by:

nargs in class AstPrimitive

args

public java.lang.String[] args()

Description copied from class: AstPrimitive

List of argument names. The length of the returned array should be equal to `nargs() - 1` (unless `nargs()` returns -1, in which case this function may return {"..."} or something similar).

Specified by:

args in class AstPrimitive

str

public java.lang.String str()

Description copied from class: AstRoot

String representation of this Ast object in the Rapids language. For AstPrimitives this is the name of the function; for AstParameters this is either the name of the variable, or the value of the numeric constant that the parameter represents. For more complicated constructs such as AstExec or AstFunction this method should return those objects as a Rapids string.

Specified by:

str in class AstRoot

apply

public ValFrame apply(Env env,
                      Env.StackHelp stk,
                      AstRoot[] asts)

Description copied from class: AstPrimitive

Primary method to invoke this function, passing all the parameters as the `asts` list.

Specified by:

apply in class AstPrimitive

Parameters:

env - Current execution environment. Variables are looked up here.

stk - TODO need clarification

asts - List of AstRoot expressions that are arguments to the function. First element in this list is the function itself.

Returns:

value resulting from calling the function with the provided list of arguments.

performGroupingWithAggregations

public ValFrame performGroupingWithAggregations(Frame fr,
                                                int[] gbCols,
                                                AstGroup.AGG[] aggs)

Performs grouping and aggregations on given data.

Parameters:

fr - frame on which grouping should be performed

gbCols - indices of columns (of given frame) to be used for grouping.

aggs - aggregations to be performed on the grouped data.

Returns:

grouped data with aggregations results.

check

public static AstNumList check(long dstX,
                               AstRoot ast)

doGroups

public static IcedHashSet<AstGroup.G> doGroups(Frame fr,
                                               int[] gbCols,
                                               AstGroup.AGG[] aggs)

doGroups

public static IcedHashSet<AstGroup.G> doGroups(Frame fr,
                                               int[] gbCols,
                                               byte[] gbColsTypes,
                                               AstGroup.AGG[] aggs,
                                               boolean hasMedian,
                                               boolean perNodeAggregates)

aggNRows

public static AstGroup.AGG[] aggNRows()

buildOutput

public static Frame buildOutput(int[] gbCols,
                                int noutCols,
                                Frame fr,
                                java.lang.String[] fcnames,
                                int ngrps,
                                MRTask mrfill)