sharpetronics/QGIS
1
0
Fork 0
mirror of https://github.com/qgis/QGIS.git synced 2025-04-14 00:07:35 -04:00
Code Issues Packages Projects Releases Wiki Activity

New class QgsDateTimeStatisticalSummary

Browse Source 
Calculates summary statistics on lists of datetime values, such
as count, count distinct, min, max and range
This commit is contained in:
Nyall Dawson 2016-05-10 11:11:33 +10:00
parent 0d165e5621
commit b053ad8769
7 changed files with 528 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
python/core/core.sip
View File

@ -39,6 +39,7 @@
%Include qgsdataitemproviderregistry.sip
%Include qgsdataprovider.sip
%Include qgsdatasourceuri.sip
%Include qgsdatetimestatisticalsummary.sip
%Include qgsdatumtransformstore.sip
%Include qgsdbfilterproxymodel.sip
%Include qgsdistancearea.sip

104
python/core/qgsdatetimestatisticalsummary.sip Normal file
View File

@ -0,0 +1,104 @@
/** \ingroup core
* \class QgsDateTimeStatisticalSummary
* \brief Calculator for summary statistics and aggregates for a list of datetimes.
*
* Statistics are calculated by calling @link calculate @endlink and passing a list of datetimes. The
* individual statistics can then be retrieved using the associated methods. Note that not all statistics
* are calculated by default. Statistics which require slower computations are only calculated by
* specifying the statistic in the constructor or via @link setStatistics @endlink.
*
* \note Added in version 2.16
*/
class QgsDateTimeStatisticalSummary
{
%TypeHeaderCode
#include <qgsdatetimestatisticalsummary.h>
%End
public:
//! Enumeration of flags that specify statistics to be calculated
enum Statistic
{
Count, //!< Count
CountDistinct, //!< Number of distinct datetime values
CountMissing, //!< Number of missing (null) values
Min, //!< Minimum (earliest) datetime value
Max, //!< Maximum (latest) datetime value
Range, //!< Interval between earliest and latest datetime value
All, //! All statistics
};
typedef QFlags<QgsDateTimeStatisticalSummary::Statistic> Statistics;
/** Constructor for QgsDateTimeStatisticalSummary
* @param stats flags for statistics to calculate
*/
QgsDateTimeStatisticalSummary( const QgsDateTimeStatisticalSummary::Statistics& stats = All );
/** Returns flags which specify which statistics will be calculated. Some statistics
* are always calculated (eg count).
* @see setStatistics
*/
Statistics statistics() const;
/** Sets flags which specify which statistics will be calculated. Some statistics
* are always calculated (eg count).
* @param stats flags for statistics to calculate
* @see statistics
*/
void setStatistics( const Statistics& stats );
/** Resets the calculated values
*/
void reset();
/** Calculates summary statistics for a list of variants. Any non-string variants will be
* ignored.
* @param values list of variants
*/
void calculate( const QVariantList& values );
/** Returns the value of a specified statistic
* @param stat statistic to return
* @returns calculated value of statistic
*/
QVariant statistic( Statistic stat ) const;
/** Returns the calculated count of values.
*/
int count() const;
/** Returns the number of distinct datetime values.
*/
int countDistinct() const;
/** Returns the set of distinct datetime values.
*/
QSet< QDateTime > distinctValues() const;
/** Returns the number of missing (null) datetime values.
*/
int countMissing() const;
/** Returns the minimum (earliest) non-null datetime value.
*/
QDateTime min() const;
/** Returns the maximum (latest) non-null datetime value.
*/
QDateTime max() const;
/** Returns the range (interval between earliest and latest non-null datetime values).
*/
QgsInterval range() const;
/** Returns the friendly display name for a statistic
* @param statistic statistic to return name for
*/
static QString displayName( Statistic statistic );
};
QFlags<QgsDateTimeStatisticalSummary::Statistic> operator|(QgsDateTimeStatisticalSummary::Statistic f1, QFlags<QgsDateTimeStatisticalSummary::Statistic> f2);

2
src/core/CMakeLists.txt
View File

@ -100,6 +100,7 @@ SET(QGIS_CORE_SRCS
qgsdataitemprovider.cpp
qgsdataitemproviderregistry.cpp
qgsdatasourceuri.cpp
qgsdatetimestatisticalsummary.cpp
qgsdatumtransformstore.cpp
qgsdbfilterproxymodel.cpp
qgsdiagramrendererv2.cpp
@ -617,6 +618,7 @@ SET(QGIS_CORE_HDRS
qgsdataitemprovider.h
qgsdataitemproviderregistry.h
qgsdatasourceuri.h
qgsdatetimestatisticalsummary.h
qgsdatumtransformstore.h
qgsdbfilterproxymodel.h
qgsdiagramrendererv2.h

144
src/core/qgsdatetimestatisticalsummary.cpp Normal file
View File

@ -0,0 +1,144 @@
/***************************************************************************
qgsdatetimestatisticalsummary.cpp
---------------------------------
Date : May 2016
Copyright : (C) 2016 by Nyall Dawson
Email : nyall dot dawson at gmail dot com
***************************************************************************
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of the GNU General Public License as published by *
* the Free Software Foundation; either version 2 of the License, or *
* (at your option) any later version. *
* *
***************************************************************************/
#include "qgsdatetimestatisticalsummary.h"
#include <QString>
#include <QDateTime>
#include <QStringList>
#include <QObject>
#include <QVariant>
#include <QVariantList>
#include "limits.h"
/***************************************************************************
* This class is considered CRITICAL and any change MUST be accompanied with
* full unit tests in test_qgsdatetimestatisticalsummary.py.
* See details in QEP #17
****************************************************************************/
QgsDateTimeStatisticalSummary::QgsDateTimeStatisticalSummary( const QgsDateTimeStatisticalSummary::Statistics& stats )
: mStatistics( stats )
{
reset();
}
void QgsDateTimeStatisticalSummary::reset()
{
mCount = 0;
mValues.clear();
mCountMissing = 0;
mMin = QDateTime();
mMax = QDateTime();
}
void QgsDateTimeStatisticalSummary::calculate( const QVariantList& values )
{
reset();
Q_FOREACH ( const QVariant& variant, values )
{
if ( variant.type() == QVariant::DateTime )
{
testDateTime( variant.toDateTime() );
}
else if ( variant.type() == QVariant::Date )
{
QDate date = variant.toDate();
testDateTime( date.isValid() ? QDateTime( date, QTime( 0, 0, 0 ) )
: QDateTime() );
}
// QTime?
}
}
void QgsDateTimeStatisticalSummary::testDateTime( const QDateTime& dateTime )
{
mCount++;
if ( !dateTime.isValid() )
mCountMissing++;
if ( mStatistics & CountDistinct )
{
mValues << dateTime;
}
if ( mStatistics & Min || mStatistics & Range )
{
if ( mMin.isValid() && dateTime.isValid() )
{
mMin = qMin( mMin, dateTime );
}
else if ( !mMin.isValid() && dateTime.isValid() )
{
mMin = dateTime;
}
}
if ( mStatistics & Max || mStatistics & Range )
{
if ( mMax.isValid() && dateTime.isValid() )
{
mMax = qMax( mMax, dateTime );
}
else if ( !mMax.isValid() && dateTime.isValid() )
{
mMax = dateTime;
}
}
}
QVariant QgsDateTimeStatisticalSummary::statistic( QgsDateTimeStatisticalSummary::Statistic stat ) const
{
switch ( stat )
{
case Count:
return mCount;
case CountDistinct:
return mValues.count();
case CountMissing:
return mCountMissing;
case Min:
return mMin;
case Max:
return mMax;
case Range:
return QVariant::fromValue( mMax - mMin );
case All:
return 0;
}
return 0;
}
QString QgsDateTimeStatisticalSummary::displayName( QgsDateTimeStatisticalSummary::Statistic statistic )
{
switch ( statistic )
{
case Count:
return QObject::tr( "Count" );
case CountDistinct:
return QObject::tr( "Count (distinct)" );
case CountMissing:
return QObject::tr( "Count (missing)" );
case Min:
return QObject::tr( "Minimum (earliest)" );
case Max:
return QObject::tr( "Maximum (latest)" );
case Range:
return QObject::tr( "Range (interval)" );
case All:
return QString();
}
return QString();
}

142
src/core/qgsdatetimestatisticalsummary.h Normal file
View File

@ -0,0 +1,142 @@
/***************************************************************************
qgsdatetimestatisticalsummary.h
-------------------------------
Date : May 2016
Copyright : (C) 2016 by Nyall Dawson
Email : nyall dot dawson at gmail dot com
***************************************************************************
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of the GNU General Public License as published by *
* the Free Software Foundation; either version 2 of the License, or *
* (at your option) any later version. *
* *
***************************************************************************/
#ifndef QGSDATETIMESTATISTICALSUMMARY_H
#define QGSDATETIMESTATISTICALSUMMARY_H
#include "qgis.h"
#include "qgsinterval.h"
#include <QSet>
#include <QDateTime>
#include <QVariantList>
/***************************************************************************
* This class is considered CRITICAL and any change MUST be accompanied with
* full unit tests in test_qgsdatetimestatisticalsummary.py.
* See details in QEP #17
****************************************************************************/
/** \ingroup core
* \class QgsDateTimeStatisticalSummary
* \brief Calculator for summary statistics and aggregates for a list of datetimes.
*
* Statistics are calculated by calling @link calculate @endlink and passing a list of datetimes. The
* individual statistics can then be retrieved using the associated methods. Note that not all statistics
* are calculated by default. Statistics which require slower computations are only calculated by
* specifying the statistic in the constructor or via @link setStatistics @endlink.
*
* \note Added in version 2.16
*/
class CORE_EXPORT QgsDateTimeStatisticalSummary
{
public:
//! Enumeration of flags that specify statistics to be calculated
enum Statistic
{
Count = 1, //!< Count
CountDistinct = 2, //!< Number of distinct datetime values
CountMissing = 4, //!< Number of missing (null) values
Min = 8, //!< Minimum (earliest) datetime value
Max = 16, //!< Maximum (latest) datetime value
Range = 32, //!< Interval between earliest and latest datetime value
All = Count | CountDistinct | CountMissing | Min | Max | Range, //! All statistics
};
Q_DECLARE_FLAGS( Statistics, Statistic )
/** Constructor for QgsDateTimeStatisticalSummary
* @param stats flags for statistics to calculate
*/
QgsDateTimeStatisticalSummary( const QgsDateTimeStatisticalSummary::Statistics& stats = All );
/** Returns flags which specify which statistics will be calculated. Some statistics
* are always calculated (eg count).
* @see setStatistics
*/
Statistics statistics() const { return mStatistics; }
/** Sets flags which specify which statistics will be calculated. Some statistics
* are always calculated (eg count).
* @param stats flags for statistics to calculate
* @see statistics
*/
void setStatistics( const Statistics& stats ) { mStatistics = stats; }
/** Resets the calculated values
*/
void reset();
/** Calculates summary statistics for a list of variants. Any non-string variants will be
* ignored.
* @param values list of variants
*/
void calculate( const QVariantList& values );
/** Returns the value of a specified statistic
* @param stat statistic to return
* @returns calculated value of statistic
*/
QVariant statistic( Statistic stat ) const;
/** Returns the calculated count of values.
*/
int count() const { return mCount; }
/** Returns the number of distinct datetime values.
*/
int countDistinct() const { return mValues.count(); }
/** Returns the set of distinct datetime values.
*/
QSet< QDateTime > distinctValues() const { return mValues; }
/** Returns the number of missing (null) datetime values.
*/
int countMissing() const { return mCountMissing; }
/** Returns the minimum (earliest) non-null datetime value.
*/
QDateTime min() const { return mMin; }
/** Returns the maximum (latest) non-null datetime value.
*/
QDateTime max() const { return mMax; }
/** Returns the range (interval between earliest and latest non-null datetime values).
*/
QgsInterval range() const { return mMax - mMin; }
/** Returns the friendly display name for a statistic
* @param statistic statistic to return name for
*/
static QString displayName( Statistic statistic );
private:
Statistics mStatistics;
int mCount;
QSet< QDateTime > mValues;
int mCountMissing;
QDateTime mMin;
QDateTime mMax;
void testDateTime( const QDateTime& dateTime );
};
Q_DECLARE_OPERATORS_FOR_FLAGS( QgsDateTimeStatisticalSummary::Statistics )
#endif // QGSDATETIMESTATISTICALSUMMARY_H

1
tests/src/python/CMakeLists.txt
View File

@ -28,6 +28,7 @@ ADD_PYTHON_TEST(PyQgsComposerPolyline test_qgscomposerpolyline.py)
ADD_PYTHON_TEST(PyQgsComposition test_qgscomposition.py)
ADD_PYTHON_TEST(PyQgsConditionalStyle test_qgsconditionalstyle.py)
ADD_PYTHON_TEST(PyQgsCoordinateTransform test_qgscoordinatetransform.py)
ADD_PYTHON_TEST(PyQgsDateTimeStatisticalSummary test_qgsdatetimestatisticalsummary.py)
ADD_PYTHON_TEST(PyQgsDelimitedTextProvider test_qgsdelimitedtextprovider.py)
ADD_PYTHON_TEST(PyQgsDistanceArea test_qgsdistancearea.py)
ADD_PYTHON_TEST(PyQgsEditWidgets test_qgseditwidgets.py)

134
tests/src/python/test_qgsdatetimestatisticalsummary.py Normal file
View File

@ -0,0 +1,134 @@
# -*- coding: utf-8 -*-
"""QGIS Unit tests for QgsDateTimeStatisticalSummary.
.. note:: This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
"""
__author__ = 'Nyall Dawson'
__date__ = '07/05/2016'
__copyright__ = 'Copyright 2016, The QGIS Project'
# This will get replaced with a git SHA1 when you do a git archive
__revision__ = '$Format:%H$'
import qgis # NOQA
from qgis.core import (QgsDateTimeStatisticalSummary,
QgsInterval
)
from qgis.PyQt.QtCore import QDateTime, QDate, QTime
from qgis.testing import unittest
class PyQgsDateTimeStatisticalSummary(unittest.TestCase):
def testStats(self):
s = QgsDateTimeStatisticalSummary()
self.assertEqual(s.statistics(), QgsDateTimeStatisticalSummary.All)
s.calculate([QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2011, 1, 5), QTime(15, 3, 1)),
QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)),
QDateTime(),
QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)),
QDateTime(),
QDateTime(QDate(2011, 1, 5), QTime(11, 10, 54))])
self.assertEqual(s.count(), 9)
self.assertEqual(s.countDistinct(), 6)
self.assertEqual(set(s.distinctValues()),
set([QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2011, 1, 5), QTime(15, 3, 1)),
QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)),
QDateTime(),
QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)),
QDateTime(QDate(2011, 1, 5), QTime(11, 10, 54))]))
self.assertEqual(s.countMissing(), 2)
self.assertEqual(s.min(), QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)))
self.assertEqual(s.max(), QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)))
self.assertEqual(s.range(), QgsInterval(693871147))
def testIndividualStats(self):
# tests calculation of statistics one at a time, to make sure statistic calculations are not
# dependent on each other
tests = [{'stat': QgsDateTimeStatisticalSummary.Count, 'expected': 9},
{'stat': QgsDateTimeStatisticalSummary.CountDistinct, 'expected': 6},
{'stat': QgsDateTimeStatisticalSummary.CountMissing, 'expected': 2},
{'stat': QgsDateTimeStatisticalSummary.Min, 'expected': QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54))},
{'stat': QgsDateTimeStatisticalSummary.Max, 'expected': QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1))},
{'stat': QgsDateTimeStatisticalSummary.Range, 'expected': QgsInterval(693871147)},
]
s = QgsDateTimeStatisticalSummary()
for t in tests:
# test constructor
s2 = QgsDateTimeStatisticalSummary(t['stat'])
self.assertEqual(s2.statistics(), t['stat'])
s.setStatistics(t['stat'])
self.assertEqual(s.statistics(), t['stat'])
s.calculate([QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2011, 1, 5), QTime(15, 3, 1)),
QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)),
QDateTime(),
QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)),
QDateTime(),
QDateTime(QDate(2011, 1, 5), QTime(11, 10, 54))])
self.assertEqual(s.statistic(t['stat']), t['expected'])
# display name
self.assertTrue(len(QgsDateTimeStatisticalSummary.displayName(t['stat'])) > 0)
def testVariantStats(self):
""" test with non-datetime values """
s = QgsDateTimeStatisticalSummary()
self.assertEqual(s.statistics(), QgsDateTimeStatisticalSummary.All)
s.calculate([QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
'asdasd',
QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
34,
QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)),
QDateTime(),
QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)),
QDateTime(),
QDateTime(QDate(2011, 1, 5), QTime(11, 10, 54))])
self.assertEqual(s.count(), 7)
self.assertEqual(set(s.distinctValues()), set([QDateTime(QDate(2015, 3, 4), QTime(11, 10, 54)),
QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)),
QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)),
QDateTime(QDate(2011, 1, 5), QTime(11, 10, 54)),
QDateTime()]))
self.assertEqual(s.countMissing(), 2)
self.assertEqual(s.min(), QDateTime(QDate(1998, 1, 2), QTime(1, 10, 54)))
self.assertEqual(s.max(), QDateTime(QDate(2019, 12, 28), QTime(23, 10, 1)))
self.assertEqual(s.range(), QgsInterval(693871147))
def testDates(self):
""" test with date values """
s = QgsDateTimeStatisticalSummary()
self.assertEqual(s.statistics(), QgsDateTimeStatisticalSummary.All)
s.calculate([QDate(2015, 3, 4),
QDate(2015, 3, 4),
QDate(2019, 12, 28),
QDate(),
QDate(1998, 1, 2),
QDate(),
QDate(2011, 1, 5)])
self.assertEqual(s.count(), 7)
self.assertEqual(set(s.distinctValues()), set([
QDateTime(QDate(2015, 3, 4), QTime()),
QDateTime(QDate(2019, 12, 28), QTime()),
QDateTime(QDate(1998, 1, 2), QTime()),
QDateTime(),
QDateTime(QDate(2011, 1, 5), QTime())]))
self.assertEqual(s.countMissing(), 2)
self.assertEqual(s.min(), QDateTime(QDate(1998, 1, 2), QTime()))
self.assertEqual(s.max(), QDateTime(QDate(2019, 12, 28), QTime()))
self.assertEqual(s.range(), QgsInterval(693792000))
if __name__ == '__main__':
unittest.main()