Sample¶

class Sample(*args)¶

Sample of real vectors.

Available constructors:

Sample(size, dim)

Sample(size, point)

Sample(other, first, last)

Parameters

sizeint, $m > 0$ , optional: The sample size. Default creates an empty sample with dimension 1.
dimensionint, $n > 0$ , optional: The real vectors dimension. Default creates an empty sample with dimension 1.
pointPoint or flat (1d) array, list or tuple of floats, optional: The point that will be repeated along the sample. Default creates a sample filled with zeros (null vectors).
otherSample: The sample contains points to copy.
firstint, $0 \leq first < m$: The index of the first point to copy.
lastint, $first < last \leq m$ , optional: The index after the last point to copy.

See also

ImportFromCSVFile, ImportFromTextFile

Examples

Create a Sample

>>> import openturns as ot
>>> import numpy as np
>>> sample = ot.Sample(3, 2)
>>> print(sample)
0 : [ 0 0 ]
1 : [ 0 0 ]
2 : [ 0 0 ]
>>> sample = ot.Sample(3, [1.0, 2.0])
>>> print(sample)
0 : [ 1 2 ]
1 : [ 1 2 ]
2 : [ 1 2 ]

Create a Sample from a (2d) array, list or tuple

>>> import numpy as np
>>> sample = ot.Sample(np.array([(1.0, 2.0), (3.0, 4.0), (5.0, 6.0)]))

and back

>>> z = np.array(sample)

Eventually samples may also be generated from probability distributions or experiments

>>> random_sample = ot.Normal(2).getSample(10)
>>> experiment = ot.LHSExperiment(ot.Normal(2), 10).generate()

Translation: addition or subtraction of a (compatible) sample or a point, or a scalar which is promoted into a point of compatible dimension with equal components

>>> print(sample + sample)
: [  2  4 ]
: [  6  8 ]
: [ 10 12 ]
>>> print(sample - sample)
: [ 0 0 ]
: [ 0 0 ]
: [ 0 0 ]
>>> print(sample - sample[0])
: [ 0 0 ]
: [ 2 2 ]
: [ 4 4 ]
>>> print(sample - sample[0, 0])
: [ 0 1 ]
: [ 2 3 ]
: [ 4 5 ]

Methods

`ImportFromCSVFile`(\*args)	Static method for building a sample from a CSV file.
`ImportFromTextFile`(\*args)	Static method for building a sample from a text file.
`add`(self, \*args)	Append a vector (in-place).
`asPoint`(self)	Accessor to the internal linear storage for 1D sample.
`clear`(self)	Erase all values.
`computeCenteredMoment`(self, k)	Estimate componentwise centered moments.
`computeCovariance`(self)	Estimate the covariance matrix.
`computeEmpiricalCDF`(self, point[, tail])	Estimate the empirical cumulative distribution function (ECDF).
`computeKendallTau`(self)	Estimate the Kendall coefficients matrix.
`computeKurtosis`(self)	Estimate the componentwise kurtosis (4th order centered normalized moment).
`computeLinearCorrelation`(self)	(ditch me?)
`computeMean`(self)	Estimate the mean vector.
`computeMedian`(self)	Estimate the componentwise medians (50%-quantiles).
`computePearsonCorrelation`(self)	Estimate the Pearson correlation matrix.
`computeQuantile`(self, \*args)	Estimate the quantile of the joint distribution underlying the sample.
`computeQuantilePerComponent`(self, \*args)	Estimate the componentwise quantiles.
`computeRange`(self)	Compute the range per component.
`computeRawMoment`(self, k)	Compute the raw (non-centered) moment per component.
`computeSkewness`(self)	Estimate the componentwise skewness (3rd order centered normalized moment).
`computeSpearmanCorrelation`(self)	Estimate the Spearman correlation matrix.
`computeStandardDeviation`(self)	Compute the Cholesky factor of the covariance matrix.
`computeStandardDeviationPerComponent`(self)	Estimate the componentwise standard deviations.
`computeVariance`(self)	Estimate the componentwise variances.
`erase`(self, \*args)	Erase point(s) at or between index(es) (in-place).
`exportToCSVFile`(self, \*args)	Dump the sample to a CSV file.
`find`(self, point)	Get the position of a point in the sample.
`getClassName`(self)	Accessor to the object’s name.
`getDescription`(self)	Accessor to the componentwise description.
`getDimension`(self)	Accessor to the sample’s dimension.
`getId`(self)	Accessor to the object’s id.
`getImplementation`(self)	Accessor to the underlying implementation.
`getMarginal`(self, \*args)	Accessor to sample marginal(s) (column(s)).
`getMax`(self)	Accessor to the componentwise maximum values.
`getMin`(self)	Accessor to the componentwise minimum values.
`getName`(self)	Accessor to the object’s name.
`getSize`(self)	Accessor to the sample size.
`rank`(self, \*args)	Compute the sample (componentwise) ranks.
`select`(self, indices)	Select points in a sample.
`setDescription`(self, description)	Accessor to the componentwise description.
`setName`(self, name)	Accessor to the object’s name.
`sort`(self, \*args)	Sort the sample.
`sortAccordingToAComponent`(self, index)	Sort the sample according to the given component.
`sortAccordingToAComponentInPlace`(self, index)	Sort the sample in place according to the given component.
`sortInPlace`(self)	Sort the sample in place.
`sortUnique`(self)	Sort the sample and remove duplicate points.
`sortUniqueInPlace`(self)	Sort the sample in place and remove duplicate points.
`split`(self, index)	Trunk the sample.
`stack`(self, sample)	Stack (horizontally) the given sample to the current one (in-place).

__init__(self, \*args)¶: Initialize self. See help(type(self)) for accurate signature.

static ImportFromCSVFile(\*args)¶

Static method for building a sample from a CSV file.

Parameters

file_namestr: Path to CSV file.
separatorstr: Separating string. Default uses csv-file-separator from the ResourceMap.

Returns

sampleSample: Sample loaded from the CSV file.

See also

exportToCSVFile

Notes

The file may or may not contain a header line (columns spanned with strings delimited with quotes). If it does contain such a header line, it will be used for setting the sample description using setDescription().

Examples

>>> import openturns as ot

Let’s first create a sample CSV file

>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> sample.exportToCSVFile('sample.csv')

And load it back

>>> loaded_sample = ot.Sample.ImportFromCSVFile('sample.csv')
>>> assert sample == loaded_sample

static ImportFromTextFile(\*args)¶

Static method for building a sample from a text file.

Parameters

file_namestr: Path to text file.
separatorstr: Separating string. Default uses a blank space.
skipped_linesint: Number of lines skipped. Default is 0.

Returns

sampleSample: Sample loaded from the text file.

See also

exportToCSVFile

Notes

The file may or may not contain a header line (columns spanned with strings delimited with quotes). If it does contain such a header line, it will be used for setting the sample description using setDescription(). It can also contain some comments, if a line starts with one of the characters contained in Sample-CommentsMarker from the ResourceMap.

Examples

>>> import openturns as ot

Let’s first create a sample text file

>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> sample.exportToCSVFile('sample.txt', ' ')

And load it back

>>> loaded_sample = ot.Sample.ImportFromTextFile('sample.txt')
>>> assert sample == loaded_sample

add(self, \*args)¶

Append a vector (in-place).

Parameters

pointsequence of float: The point to append.

Examples

>>> import openturns as ot
>>> sample = ot.Sample(3, 2)
>>> sample.add([1.0, 2.0])
>>> print(sample)
0 : [ 0 0 ]
1 : [ 0 0 ]
2 : [ 0 0 ]
3 : [ 1 2 ]

asPoint(self)¶

Accessor to the internal linear storage for 1D sample.

Returns

valuesPoint: Flat internal representation of the sample.

Notes

Available only for 1D sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal().getSample(5)
>>> print(sample)
    [ X0        ]
0 : [  0.608202 ]
1 : [ -1.26617  ]
2 : [ -0.438266 ]
3 : [  1.20548  ]
4 : [ -2.18139  ]
>>> print(sample.asPoint())
[0.608202,-1.26617,-0.438266,1.20548,-2.18139]

clear(self)¶: Erase all values.

computeCenteredMoment(self, k)¶

Estimate componentwise centered moments.

Parameters

kint: The centered moment’s order.

Returns

mPoint: Componentwise centered moment of order $k$ estimated from the sample.

Notes

The centered moment of order $k$ is estimated as follows:

$\vect{\widehat{m}}^{(k)}_0 = \Tr{\left(\frac{1}{m} \sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^k, \quad i = 1, \ldots, n\right)}$

where $\vect{\widehat{\mu}}$ is the estimator of the mean.

These estimators are the natural (possibly biased) estimators. For unbiased estimators use the other dedicated methods such as computeVariance() for the variance.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeCenteredMoment(2))
[0.915126,0.873119]

computeCovariance(self)¶

Estimate the covariance matrix.

Returns

covarianceCovarianceMatrix: Covariance matrix estimated from the sample.

Notes

The covariance matrix is estimated as follows:

$\mat{\widehat{\Sigma}} = \left[\frac{1}{m - 1} \sum_{k=1}^m \left(x_i^{(k)} - \widehat{\mu}_i\right) \left(x_j^{(k)} - \widehat{\mu}_j\right), \quad i, j = 1, \ldots, n\right]$

where $\vect{\widehat{\mu}}$ denotes the estimate of the mean.

This is an unbiased estimator.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeCovariance())
[[ 0.946682  0.0182104 ]
 [ 0.0182104 0.903226  ]]

computeEmpiricalCDF(self, point, tail=False)¶

Estimate the empirical cumulative distribution function (ECDF).

Parameters

xsequence of float: CDF input.
survivalbool, optional: A flag telling whether this should estimate the empirical cumulative distribution function or the empirical survival function. Default is False and estimates the CDF.

Returns

pfloat, $0 \leq p \leq 1$: Empirical CDF or SF value at point x.

Notes

The empirical cumulative distribution function (CDF) is estimated as follows:

$\hat{F}(\vect{x}) = \frac{1}{m} \sum_{j=1}^m \mathbf{1}_{\cap_{i=1}^n x_i^{(j)} \leq x_i}(\vect{x})$

The empirical survival function (SF) is estimated in a similar way:

$\hat{S}(\vect{x}) = \frac{1}{m} \sum_{j=1}^m \mathbf{1}_{\cap_{i=1}^n x_i^{(j)} > x_i}(\vect{x})$

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeEmpiricalCDF(sample[0]))
0.1

computeKendallTau(self)¶

Estimate the Kendall coefficients matrix.

Returns

tauCorrelationMatrix: Kendall coefficients matrix estimated from the sample.

Notes

This uses an external implementation provided under the Boost Software License by David Simcha based on the paper by [knight1966]. It actually switches between two implementations depending on the sample size:

The most basic implementation performing in $O(m^2)$ is used when the sample size is less than SampleImplementation-SmallKendallTau from the ResourceMap.

The other more complex implementation performing in $O(m\log(m))$ is used for larger samples.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeKendallTau())
[[ 1          0.00689655 ]
 [ 0.00689655 1          ]]

computeKurtosis(self)¶

Estimate the componentwise kurtosis (4th order centered normalized moment).

Returns

kurtosisPoint: Componentwise kurtosis estimated from the sample.

Notes

The componentwise kurtosis are estimated as follows:

$\vect{\widehat{\kappa}} = \Tr{\left(\frac{m (m-1) (m+1)}{(m-2) (m-3)} \frac{\sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^4} {\left(\sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^2 \right)^2} - 3 \frac{3 (m-5)}{(m-2) (m-3)}, \quad i = 1, \ldots, n\right)}$

where $\vect{\widehat{\mu}}$ is the estimate of the mean.

This estimator is unbiased.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeKurtosis())
[3.27647,2.40275]

computeLinearCorrelation(self)¶: (ditch me?)

computeMean(self)¶

Estimate the mean vector.

Returns

meanPoint: Mean vector estimated from the sample.

Notes

The mean is estimated as follows:

$\vect{\widehat{\mu}} = \Tr{\left(\frac{1}{m} \sum_{j=1}^m x_i^{(j)}, \quad i = 1, \ldots, n\right)}$

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeMean())
[-0.0512622,0.136653]

computeMedian(self)¶

Estimate the componentwise medians (50%-quantiles).

Returns

medianPoint: Median vector estimated from the sample.

See also

computeQuantilePerComponent

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeMedian())
[0.221141,0.108703]

computePearsonCorrelation(self)¶

Estimate the Pearson correlation matrix.

Returns

rhoCorrelationMatrix: Pearson correlation matrix estimated from the sample.

Notes

The Pearson correlation matrix is estimated as follows:

$\mat{\widehat{\rho}} = \left[\frac{\widehat{\Sigma}_{i,j}} {\widehat{\Sigma}_{i,i} \widehat{\Sigma}_{j,j}}, \quad i,j = 1, \ldots, n\right]$

where $\mat{\widehat{\Sigma}}$ denotes the estimate of the covariance.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computePearsonCorrelation())
[[ 1         0.0196933 ]
 [ 0.0196933 1         ]]

computeQuantile(self, \*args)¶

Estimate the quantile of the joint distribution underlying the sample.

Parameters

pfloat, $0 \leq p \leq 1$ , or sequence of float: Input probability level.

Returns

quantilePoint or Sample: Quantile of the joint distribution at probability level $p$ , estimated from the sample.

Raises

NotImplementedYetErrorIf the dimension is greater than 1.

See also

computeQuantilePerComponent

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(1).getSample(30)
>>> print(sample.computeQuantile(.2))
[-0.947394]

computeQuantilePerComponent(self, \*args)¶

Estimate the componentwise quantiles.

Parameters

pfloat, $0 \leq p \leq 1$ , or sequence of float: Input probability level.

Returns

quantilePoint or Sample: Componentwise quantile at probability level $p$ , estimated from the sample.

Notes

The present implementation interpolates the quantile between the two adjacent empirical quantiles ( $\widehat{x}_i^-$ and $\widehat{x}_i^+$ ):

$\widehat{q}_i = \alpha \widehat{x}_i^- + (1 - \alpha) \widehat{x}_i^+$

where $\alpha = p m - 0.5$ .

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeQuantilePerComponent(0.2))
[-0.696412,-0.767092]

computeRange(self)¶

Compute the range per component.

Returns

rangePoint: Componentwise ranges estimated from the sample.

Notes

The statistical range is defined as the deviation between the maximal and the minimal value of the sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeRange())
[4.02827,3.49949]

computeRawMoment(self, k)¶

Compute the raw (non-centered) moment per component.

Parameters

kint, $k \geq 0$: Componentwise moment’s order.

Returns

momentsPoint: Componentwise moments estimated from the sample.

Notes

The (raw) moment of order $k$ is estimated as follows:

$\vect{\widehat{m}}^{(k)} = \Tr{\left(\frac{1}{m} \sum_{j=1}^m {x_i^{(j)}}^k, \quad i = 1, \ldots, n\right)}$

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeRawMoment(2))
[0.917754,0.891793]

computeSkewness(self)¶

Estimate the componentwise skewness (3rd order centered normalized moment).

Returns

skewnessPoint: Componentwise skewness estimated from the sample.

Notes

The componentwise skewnesses are estimated as follows:

$\vect{\widehat{\delta}} = \Tr{\left(m \frac{\sqrt{m-1}}{m-2} \frac{\sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^3} {\left(\sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^2 \right)^{3/2}}, \quad i = 1, \ldots, n\right)}$

where $\vect{\widehat{\mu}}$ is the estimate of the mean.

This is an unbiased estimator.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeSkewness())
[-0.69393,0.231931]

computeSpearmanCorrelation(self)¶

Estimate the Spearman correlation matrix.

Returns

rhoCorrelationMatrix: Spearman correlation matrix estimated from the sample.

See also

rank, computePearsonCorrelation

Notes

The Spearman correlation matrix is estimated as the Pearson correlation matrix of the ranks sample (i.e. using self.rank().computePearsonCorrelation()).

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeSpearmanCorrelation())
[[  1          -0.00556174 ]
 [ -0.00556174  1          ]]

computeStandardDeviation(self)¶

Compute the Cholesky factor of the covariance matrix.

Estimated from the sample.

Returns

LTriangularMatrix: Lower (left) Cholesky factor of the covariance matrix estimated from the sample.

Raises

RuntimeErrorIf the estimated covariance matrix is not positive definite. In: this case, you might want to estimate the covariance and manually shrink negative eigenvalues.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeStandardDeviation())
[[ 0.972976  0         ]
 [ 0.0187161 0.950198  ]]

computeStandardDeviationPerComponent(self)¶

Estimate the componentwise standard deviations.

Returns

standard_deviationsPoint: Componentwise standard deviation estimated from the sample.

See also

computeVariance

Notes

The componentwise standard deviations are estimated as the square root of the componentwise variances.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeStandardDeviationPerComponent())
[0.972976,0.950382]

computeVariance(self)¶

Estimate the componentwise variances.

Returns

variancesPoint: Componentwise variances estimated from the sample.

Notes

The componentwise variances are estimated as follows:

$\vect{\widehat{\sigma^2}} = \Tr{\left(\frac{1}{m-1} \sum_{j=1}^m \left(x_i^{(j)} - \widehat{\mu}_i\right)^2, \quad i = 1, \ldots, n\right)}$

where $\vect{\widehat{\mu}}$ is the estimate of the mean.

This estimator is unbiased.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.computeVariance())
[0.946682,0.903226]

erase(self, \*args)¶

Erase point(s) at or between index(es) (in-place).

Parameters

fint, $0 \leq f < m$: The index of the first point to erase.
lint, $f < l \leq m$ , optional: The index after the last point to erase. Default uses l = f + 1 and only removes sample[f].
Returns
——-
erased_sampleSample: Erased sample [sample[:i_start:], sample[i_stop::]].

Examples

>>> import openturns as ot
>>> sample = ot.Sample([[i] for i in range(5)])
>>> print(sample)
0 : [ 0 ]
1 : [ 1 ]
2 : [ 2 ]
3 : [ 3 ]
4 : [ 4 ]
>>> sample.erase(1, 3)
>>> print(sample)
0 : [ 0 ]
1 : [ 3 ]
2 : [ 4 ]

exportToCSVFile(self, \*args)¶

Dump the sample to a CSV file.

Parameters

file_namestr: Path to CSV file.
separatorstr: Separating string. Default uses csv-file-separator from the ResourceMap.

See also

ImportFromCSVFile, ImportFromTextFile

Notes

This will create a header line with componentwise descriptions (obtained from getDescription()) between quotes as column names.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> sample.exportToCSVFile('sample.csv', '; ')

find(self, point)¶

Get the position of a point in the sample.

Parameters

pointsequence of float: The wanted point.

Returns

indexint,: Returns $m$ if the point does not belong to the sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(30)
>>> print(sample.find(sample[10]))
10
>>> print(sample.find([0.0, 0.0]))
30

getClassName(self)¶

Accessor to the object’s name.

Returns

class_namestr: The object class name (object.__class__.__name__).

getDescription(self)¶

Accessor to the componentwise description.

Returns

descriptionDescription: Description of the sample’s components.

See also

setDescription

getDimension(self)¶

Accessor to the sample’s dimension.

Returns

nint: The number of components of the points in the sample.

getId(self)¶

Accessor to the object’s id.

Returns

idint: Internal unique identifier.

getImplementation(self)¶

Accessor to the underlying implementation.

Returns

implImplementation: The implementation class.

getMarginal(self, \*args)¶

Accessor to sample marginal(s) (column(s)).

Parameters

indicesint, sequence of int, $0 \leq i < n$ or sequence of str: The identifiers of the wanted marginal(s). When the description contains duplicate labels, the first marginal is picked up.

Returns

sampleSample: A subsample of the present sample with the requested marginal(s).

Notes

The Sample also implements slicing in its __getitem__ method.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(10).getSample(3)
>>> print(sample.getMarginal([1, 4]))
    [ X1        X4        ]
0 : [ -1.26617  -2.18139  ]
1 : [  0.261018 -1.31178  ]
2 : [  0.445785  0.473617 ]

getMax(self)¶

Accessor to the componentwise maximum values.

Returns

maximum_valuesPoint: Componentwise maximum values.

getMin(self)¶

Accessor to the componentwise minimum values.

Returns

minimum_valuesPoint: Componentwise minimum values.

getName(self)¶

Accessor to the object’s name.

Returns

namestr: The name of the object.

getSize(self)¶

Accessor to the sample size.

Returns

mint: The number points in the sample.

rank(self, \*args)¶

Compute the sample (componentwise) ranks.

Parameters

marginal_indexint, $0 \leq i < n$ , optional: The component whose ranks are wanted. Default computes the ranks of all the components.

Returns

ranksSample: The requested ranks.

Notes

The ranks of a 1d sample is a list of indices that sorts the points in the ascending order. Ties (equal points) are averaged.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> print(sample.rank())
    [ X0 X1 ]
0 : [ 2  0  ]
1 : [ 1  2  ]
2 : [ 0  1  ]

select(self, indices)¶

Select points in a sample.

It selects the points at given locations and returns them as a new sample.

Parameters

indicessequence of int, $0 \leq i < m$: The selected indices.

Returns

selected_sampleSample: The selected points as a sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> selected_sample = sample.select([1, 0, 1])
>>> print(selected_sample)
    [ X0        X1        ]
0 : [ -0.438266  1.20548  ]
1 : [  0.608202 -1.26617  ]
2 : [ -0.438266  1.20548  ]

setDescription(self, description)¶

Accessor to the componentwise description.

Parameters

descriptionsequence of str: Description of the sample’s components.

See also

getDescription

setName(self, name)¶

Accessor to the object’s name.

Parameters

namestr: The name of the object.

sort(self, \*args)¶

Sort the sample.

Parameters

marginal_indexint, $0 \leq i < n$ , optional: The component to sort. Default sorts the whole sample.

Returns

sorted_sampleSample: The requested sorted sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> print(sample.sort())
    [ X0        X1        ]
0 : [ -2.18139   0.350042 ]
1 : [ -0.438266  1.20548  ]
2 : [  0.608202 -1.26617  ]

sortAccordingToAComponent(self, index)¶

Sort the sample according to the given component.

Parameters

marginal_indexint, $0 \leq i < n$: The component to use for sorting the sample.

Returns

sorted_sampleSample: The sample sorted according to the given component.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> print(sample.sortAccordingToAComponent(0))
    [ X0        X1        ]
0 : [ -2.18139   0.350042 ]
1 : [ -0.438266  1.20548  ]
2 : [  0.608202 -1.26617  ]

sortAccordingToAComponentInPlace(self, index)¶

Sort the sample in place according to the given component.

Parameters

marginal_indexint, $0 \leq i < n$: The component to use for sorting the sample.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> sample.sortAccordingToAComponentInPlace(0)
>>> print(sample)
    [ X0        X1        ]
0 : [ -2.18139   0.350042 ]
1 : [ -0.438266  1.20548  ]
2 : [  0.608202 -1.26617  ]

sortInPlace(self)¶

Sort the sample in place.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> sample.sortInPlace()
>>> print(sample)
    [ X0        X1        ]
0 : [ -2.18139   0.350042 ]
1 : [ -0.438266  1.20548  ]
2 : [  0.608202 -1.26617  ]

sortUnique(self)¶

Sort the sample and remove duplicate points.

Returns

unique_sampleSample: The requested sorted sample with duplicate points removed.

Examples

>>> import openturns as ot
>>> sample = ot.Sample([[3, 0, 3], [1, 1, 0], [0, 2, 2], [1, 1, 0]])
>>> print(sample)
0 : [ 3 0 3 ]
1 : [ 1 1 0 ]
2 : [ 0 2 2 ]
3 : [ 1 1 0 ]
>>> print(sample.sortUnique())
0 : [ 0 2 2 ]
1 : [ 1 1 0 ]
2 : [ 3 0 3 ]

sortUniqueInPlace(self)¶

Sort the sample in place and remove duplicate points.

Examples

>>> import openturns as ot
>>> sample = ot.Sample([[3, 0, 3], [1, 1, 0], [0, 2, 2], [1, 1, 0]])
>>> print(sample)
0 : [ 3 0 3 ]
1 : [ 1 1 0 ]
2 : [ 0 2 2 ]
3 : [ 1 1 0 ]
>>> sample.sortUniqueInPlace()
>>> print(sample)
0 : [ 0 2 2 ]
1 : [ 1 1 0 ]
2 : [ 3 0 3 ]

split(self, index)¶

Trunk the sample.

It splits the sample before the index passed as argument and returns the remainder as new sample.

Parameters

indexint, $0 \leq i < m$: The truncation index.

Returns

remainder_sampleSample: The remainder sample (everyting that comes after the truncation index).

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> remainder_sample = sample.split(1)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
>>> print(remainder_sample)
    [ X0        X1        ]
0 : [ -0.438266  1.20548  ]
1 : [ -2.18139   0.350042 ]

stack(self, sample)¶

Stack (horizontally) the given sample to the current one (in-place).

Parameters

sampleSample: Sample to stack with compatible size.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> sample = ot.Normal(2).getSample(3)
>>> print(sample)
    [ X0        X1        ]
0 : [  0.608202 -1.26617  ]
1 : [ -0.438266  1.20548  ]
2 : [ -2.18139   0.350042 ]
>>> another_sample = ot.Normal(2).getSample(3)
>>> print(another_sample)
    [ X0        X1        ]
0 : [ -0.355007  1.43725  ]
1 : [  0.810668  0.793156 ]
2 : [ -0.470526  0.261018 ]
>>> sample.stack(another_sample)
>>> print(sample)
    [ X0        X1        X0        X1        ]
0 : [  0.608202 -1.26617  -0.355007  1.43725  ]
1 : [ -0.438266  1.20548   0.810668  0.793156 ]
2 : [ -2.18139   0.350042 -0.470526  0.261018 ]

OpenTURNS

An Open source initiative for the Treatment of Uncertainties, Risks'N Statistics

Previous topic

Next topic

This Page

Sample¶