org.das2.qds.util

Class Reduction

java.lang.Object

org.das2.qds.util.Reduction

public class Reduction
extends java.lang.Object

Reduction is set of static methods for reducing data, or averaging data to make smaller datasets.

Constructor Summary

Constructors

Constructor and Description
Reduction()

Method Summary

Modifier and Type	Method and Description
static QDataSet	hexbin(QDataSet ds, QDataSet z) reduce the buckshot scatter data by laying it out on a 2-D hexgrid and accumulating the hits to each cell.
static QDataSet	histogram2D(QDataSet ds, QDataSet xxx, QDataSet yyy) reduce the buckshot scatter data by laying it out on a 2-D grid and accumulating the hits to each cell.
static QDataSet	lastPointAt2D(QDataSet ds, QDataSet xx, QDataSet yy, QDataSet xxx, QDataSet yyy) return the value of the last data point at each location.
static QDataSet	reduce2D(QDataSet ds, QDataSet xLimit, QDataSet yLimit) produce a simpler version of the dataset by averaging adjacent data.
static QDataSet	reducex(QDataSet ds, QDataSet xLimit) produce a simpler version of the dataset by averaging data adjacent in X.
static QDataSet	reducex(QDataSet ds, QDataSet xLimit, boolean xregular) produce a simpler version of the dataset by averaging data adjacent in X.

Methods inherited from class java.lang.Object

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

Constructor Detail

Reduction

public Reduction()

Method Detail

reducex

public static QDataSet reducex(QDataSet ds,
                               QDataSet xLimit)

produce a simpler version of the dataset by averaging data adjacent in X. code taken from org.das2.graph.GraphUtil.reducePath. Adjacent points are averaged together until a point is found that is not in the bin, and then a new bin is started. The bin's lower bounds are integer multiples of xLimit. xLimit is a rank 0 dataset. 2015-06-18: xcadence and bins are now regular. Because of high-resolution magnetometer data, this is extended to support this data type. This will set the DELTA_PLUS and DELTA_MINUS variables to the extremes of each bin. To remove these, use putProperty( QDataSet.DELTA_MINUS, None ) (None in Jython, null for Java) and putProperty( QDataSet.DELTA_PLUS, None ).

Parameters:

ds - rank 1 or rank 2 dataset. Must have DEPEND_0 (presently) and be a qube. If this is null, then the result is null.

xLimit - the size of the bins or null to indicate no limit.

Returns:

the reduced dataset, or null if the input dataset was null.

reducex

public static QDataSet reducex(QDataSet ds,
                               QDataSet xLimit,
                               boolean xregular)

produce a simpler version of the dataset by averaging data adjacent in X. code taken from org.das2.graph.GraphUtil.reducePath. Adjacent points are averaged together until a point is found that is not in the bin, and then a new bin is started. The bin's lower bounds are integer multiples of xLimit. xLimit is a rank 0 dataset. 2015-06-18: xcadence and bins are now regular. Because of high-resolution magnetometer data, this is extended to support this data type. This will set the DELTA_PLUS and DELTA_MINUS variables to the extremes of each bin. To remove these, use putProperty( QDataSet.DELTA_MINUS, None ) (None in Jython, null for Java) and putProperty( QDataSet.DELTA_PLUS, None ).

Parameters:

ds - rank 1 or rank 2 dataset. Must have DEPEND_0 (presently) and be a qube. If this is null, then the result is null.

xLimit - the size of the bins or null to indicate no limit.

xregular - if true, then return xtags with a uniform cadence; if false, return averages of x as well, and don't grid x.

Returns:

the reduced dataset, or null if the input dataset was null.

reduce2D

public static QDataSet reduce2D(QDataSet ds,
                                QDataSet xLimit,
                                QDataSet yLimit)

produce a simpler version of the dataset by averaging adjacent data. code taken from org.das2.graph.GraphUtil.reducePath. Adjacent points are averaged together until a point is found that is not in the bin, and then a new bin is started. The bin's lower bounds are integer multiples of xLimit and yLimit. If yLimit is null, then averaging is done for all points in the x bin, regardless of how close they are in Y. This is similarly true when xLimit is null. xLimit and yLimit are rank 0 datasets, so that they can indicate that binning should be done in log space rather than linear. In this case, a SCALE_TYPE for the dataset should be "log" and its unit should be convertible to Units.logERatio (for example, Units.log10Ratio or Units.percentIncrease). Note when either is log, then averaging is done in the log space.

Parameters:

ds - rank 1 dataset. Must have DEPEND_0 (presently)

xLimit - the size of the bins or null to indicate no limit.

yLimit - the size of the bins or null to indicate no limit.

Returns:

the reduced dataset, rank 1 with DEPEND_0.

hexbin

public static QDataSet hexbin(QDataSet ds,
                              QDataSet z)

reduce the buckshot scatter data by laying it out on a 2-D hexgrid and accumulating the hits to each cell. This has not been thoroughly verified.

Parameters:

ds - rank1 Y(X)

z - null or data to average

Returns:

rank 2 ds containing frequency of occurrence for each bin, with DEPEND_0=xxx and DEPEND_1=yyy.

Throws:

java.lang.IllegalArgumentException - when the units cannot be converted

See Also:

Ops.histogram2d(org.das2.qds.QDataSet, org.das2.qds.QDataSet, int[], org.das2.qds.QDataSet, org.das2.qds.QDataSet), https://cran.r-project.org/web/packages/hexbin/vignettes/hexagon_binning.pdf

lastPointAt2D

public static QDataSet lastPointAt2D(QDataSet ds,
                                     QDataSet xx,
                                     QDataSet yy,
                                     QDataSet xxx,
                                     QDataSet yyy)

return the value of the last data point at each location.

Parameters:

ds - rank1 data

xx - rank 1 X values for each data point

yy - rank 1 Y values for each data point

xxx - rank 1 dataset describes the bins, which must be uniformly linearly spaced, or log spaced. Uses SCALE_TYPE property.

yyy - rank 1 dataset describes the bins, which must be uniformly linearly spaced, or log spaced.

Returns:

rank 2 ds containing the last point at each bin, with DEPEND_0=xxx and DEPEND_1=yyy.

Throws:

java.lang.IllegalArgumentException - when the units cannot be converted

See Also:

Ops.histogram2d(org.das2.qds.QDataSet, org.das2.qds.QDataSet, int[], org.das2.qds.QDataSet, org.das2.qds.QDataSet)

histogram2D

public static QDataSet histogram2D(QDataSet ds,
                                   QDataSet xxx,
                                   QDataSet yyy)

reduce the buckshot scatter data by laying it out on a 2-D grid and accumulating the hits to each cell. Written originally to support SeriesRenderer, to replace the "200000 point limit" warning.

Parameters:

ds - rank1 Y(X)

xxx - rank1 dataset describes the bins, which must be uniformly linearly spaced, or log spaced. Uses SCALE_TYPE property.

yyy - rank1 dataset describes the bins, which must be uniformly linearly spaced, or log spaced.

Returns:

rank 2 ds containing frequency of occurrence for each bin, with DEPEND_0=xxx and DEPEND_1=yyy.

Throws:

java.lang.IllegalArgumentException - when the units cannot be converted

See Also:

Ops.histogram2d(org.das2.qds.QDataSet, org.das2.qds.QDataSet, int[], org.das2.qds.QDataSet, org.das2.qds.QDataSet)