gum::learning::DBCell Class Reference

The class representing the original values of the cells of databases. More...

#include <agrum/tools/database/DBCell.h>

Public Member Functions
Constructors / Destructors
	DBCell ()
	default constructor (ontains a missing value) More...
	DBCell (const float nb)
	constructor for a real number More...
	DBCell (const int nb)
	constructor for an integer number More...
	DBCell (const std::string &str)
	constructor for a string More...
	DBCell (const DBCell &from)
	copy constructor More...
	DBCell (DBCell &&from)
	move constructor More...
	~DBCell ()
	destructor More...
Operators
DBCell &	operator= (const DBCell &from)
	copy operator More...
DBCell &	operator= (DBCell &&from)
	move operator More...
DBCell &	operator= (const float x)
	assignment operator More...
DBCell &	operator= (const int x)
	assignment operator More...
DBCell &	operator= (const std::string &x)
	assignment operator More...
bool	operator== (const DBCell &from) const
	test of equality More...
bool	operator!= (const DBCell &from) const
	test of inequality More...
Accessors / Modifiers
EltType	type () const noexcept
	returns the current type of the DBCell More...
bool	convertType (const EltType newtype)
	try to convert the content of the DBCell into another type More...
float	real () const
	returns the DBcell as a real number More...
void	setReal (const float x)
	sets the content of the DBCell More...
void	setReal (const std::string &elt)
	sets the content of the DBCell from a string More...
int	integer () const
	returns the DBcell as an integer More...
void	setInteger (const int x)
	sets the content of the DBCell More...
void	setInteger (const std::string &elt)
	sets the content of the DBCell from a string More...
const std::string &	string () const
	returns the DBcell as a string More...
int	stringIndex () const
	returns the DBcell as the index of a string in a static bijection More...
void	setString (const std::string &elt)
	sets the content of the DBCell More...
void	setMissingState ()
	sets the DBCell as a missing element More...
bool	isMissing () const
	indicates whether the cell contains a missing value More...

Public Types
enum	EltType : unsigned char { EltType::REAL, EltType::INTEGER, EltType::STRING, EltType::MISSING }
	the set of types possibly taken by the last element read More...

Public Static Accessors / Modifiers
template<template< typename > class ALLOC = std::allocator>
std::string	toString (const std::vector< std::string, ALLOC< std::string > > &missingVals) const
	returns the content of the DBCell as a string, whatever its type More...
static const std::string &	string (const int index)
	strings are stored into a static bijection. Get its ith string More...
template<template< typename > class ALLOC = std::allocator>
static EltType	bestType (const std::string &str, const std::vector< std::string, ALLOC< std::string > > &missingVals)
	returns the best type to store a given element encoded as a string More...
template<template< typename > class ALLOC = std::allocator>
static DBCell	bestDBCell (const std::string &str, const std::vector< std::string, ALLOC< std::string > > &missingVals)
	returns the DBCell with the best type for an element encoded as a string More...
static bool	isInteger (const std::string &str)
	determines whether a string corresponds precisely to an integer More...
static bool	isReal (const std::string &str)
	determine whether a string corresponds precisely to a real number More...
template<template< typename > class ALLOC = std::allocator>
static bool	isMissing (const std::string &str, const std::vector< std::string, ALLOC< std::string > > &missingVals)
	checks whether a string correspond to a missing value More...

Detailed Description

The class representing the original values of the cells of databases.

Learning algorithms read input data, the so-called datasets, and infer their models. To be fast, they should not parse directly the original datasets, which are often databases of strings, but rather they should parse preprocessed data with types more suited for fast analysis. To do so, in aGrUM, learning algorithms parse datasets of DBTranslatedValue instances. However, sometimes, these algorithms make some decisions which, afterwards, they may realize were not so good and should be modified. For instance, when coping with continuous variables, they may use a discretization for fast learning the structure of a hybrid Bayesian network. But, given the structure found, they may try to find a better discretization. With a new discretization, the original dataset should be parsed again and mapped into new DBTranslatedValue instances. Unfortunately, reparsing, say, a CSV file or a SQL database, is time consuming. Sometimes, it is faster to read such datasets once and store them in an appropriate form that can be mapped easily into DBTranslatedValue instances. The RawDatabaseTable class is made precisely for this purpose and can be thought of as a 2-dimensional table, the elements of which are DBCell instances. The latter can actually encode compactly integers, real numbers, strings and even missing values. So they are very well suited to be the type of the cells of the RawDatabaseTable class.

Definition at line 71 of file DBCell.h.

Member Enumeration Documentation

EltType

enum gum::learning::DBCell::EltType : unsigned char

strong

the set of types possibly taken by the last element read

Enumerator
REAL
INTEGER
STRING
MISSING

Definition at line 74 of file DBCell.h.

                         : unsigned char
      {
        REAL,
        INTEGER,
        STRING,
        MISSING
      };

Constructor & Destructor Documentation

DBCell() [1/6]

gum::learning::DBCell::DBCell

(

)

default constructor (ontains a missing value)

DBCell() [2/6]

gum::learning::DBCell::DBCell

(

const float

nb

)

constructor for a real number

DBCell() [3/6]

gum::learning::DBCell::DBCell

(

const int

nb

)

constructor for an integer number

DBCell() [4/6]

gum::learning::DBCell::DBCell

(

const std::string &

str

)

constructor for a string

DBCell() [5/6]

gum::learning::DBCell::DBCell

(

const DBCell &

from

)

copy constructor

DBCell() [6/6]

gum::learning::DBCell::DBCell

(

DBCell &&

from

)

move constructor

~DBCell()

gum::learning::DBCell::~DBCell

(

)

destructor

Member Function Documentation

bestDBCell()

template<template< typename > class ALLOC = std::allocator>

static DBCell gum::learning::DBCell::bestDBCell ( const std::string &  str, const std::vector< std::string, ALLOC< std::string > > &  missingVals  )

static

returns the DBCell with the best type for an element encoded as a string

Parameters

str	the string to convert into a DBCell
missingVals	a vector containing the set of strings that should be interpreted as missing values. Whenever str matches one these strings, the returned DBCell represents a missing value.

bestType()

template<template< typename > class ALLOC = std::allocator>

static EltType gum::learning::DBCell::bestType ( const std::string &  str, const std::vector< std::string, ALLOC< std::string > > &  missingVals  )

static

returns the best type to store a given element encoded as a string

Parameters

str	the string to convert into a DBCell
missingVals	a vector containing the set of strings that should be interpreted as missing values. Whenever str matches one these strings, the returned EltType represents a missing value.

convertType()

bool gum::learning::DBCell::convertType

(

const EltType

newtype

)

try to convert the content of the DBCell into another type

Returns

true if the conversion has been successfully performed.

integer()

int gum::learning::DBCell::integer

(

)

const

returns the DBcell as an integer

Warning

if the cell is not of type INTEGER, the dbcell will not try to convert its content into a float, it will raise a TypeError exception.

Exceptions

TypeError

if the DBCell is not labelled as type INTEGER

isInteger()

static bool gum::learning::DBCell::isInteger ( const std::string &  str )

static

determines whether a string corresponds precisely to an integer

isMissing() [1/2]

bool gum::learning::DBCell::isMissing

(

)

const

indicates whether the cell contains a missing value

isMissing() [2/2]

template<template< typename > class ALLOC = std::allocator>

static bool gum::learning::DBCell::isMissing ( const std::string &  str, const std::vector< std::string, ALLOC< std::string > > &  missingVals  )

static

checks whether a string correspond to a missing value

isReal()

static bool gum::learning::DBCell::isReal ( const std::string &  str )

static

determine whether a string corresponds precisely to a real number

operator!=()

bool gum::learning::DBCell::operator!=

(

const DBCell &

from

)

const

test of inequality

operator=() [1/5]

DBCell& gum::learning::DBCell::operator=

(

const DBCell &

from

)

copy operator

operator=() [2/5]

DBCell& gum::learning::DBCell::operator=

(

DBCell &&

from

)

move operator

operator=() [3/5]

DBCell& gum::learning::DBCell::operator=

(

const float

x

)

assignment operator

operator=() [4/5]

DBCell& gum::learning::DBCell::operator=

(

const int

x

)

assignment operator

operator=() [5/5]

DBCell& gum::learning::DBCell::operator=

(

const std::string &

x

)

assignment operator

operator==()

bool gum::learning::DBCell::operator==

(

const DBCell &

from

)

const

test of equality

real()

float gum::learning::DBCell::real

(

)

const

returns the DBcell as a real number

Warning

if the cell is not of type REAL, the dbcell will not try to convert its content into a real number, it will raise a TypeError exception.

Returns

the content of the DBCell (if this is a real number)

Exceptions

TypeError

if the DBCell is not labelled as type REAL

setInteger() [1/2]

void gum::learning::DBCell::setInteger

(

const int

x

)

sets the content of the DBCell

setInteger() [2/2]

void gum::learning::DBCell::setInteger

(

const std::string &

elt

)

sets the content of the DBCell from a string

Exceptions

TypeError

if the string does not correspond to an integer

setMissingState()

void gum::learning::DBCell::setMissingState

(

)

sets the DBCell as a missing element

setReal() [1/2]

void gum::learning::DBCell::setReal

(

const float

x

)

sets the content of the DBCell

setReal() [2/2]

void gum::learning::DBCell::setReal

(

const std::string &

elt

)

sets the content of the DBCell from a string

Exceptions

TypeError

if the string does not correspond to a real number

setString()

void gum::learning::DBCell::setString

(

const std::string &

elt

)

sets the content of the DBCell

string() [1/2]

const std::string& gum::learning::DBCell::string

(

)

const

returns the DBcell as a string

Warning

if the cell is not of type STRING, the dbcell will not try to convert its content into a string, it will raise a TypeError exception.

Exceptions

TypeError

if the DBCell is not labelled as type STRING

string() [2/2]

static const std::string& gum::learning::DBCell::string ( const int  index )

static

strings are stored into a static bijection. Get its ith string

Exceptions

UndefinedElement

if the index does not correspond to any string

stringIndex()

int gum::learning::DBCell::stringIndex

(

)

const

returns the DBcell as the index of a string in a static bijection

All strings referenced by DBCells are stored in a bijection. This method returns the index of the string referenced by the current DBCell.

Exceptions

TypeError

if the DBCell is not labelled as type STRING

toString()

template<template< typename > class ALLOC = std::allocator>

std::string gum::learning::DBCell::toString

(

const std::vector< std::string, ALLOC< std::string > > &

missingVals

)

const

returns the content of the DBCell as a string, whatever its type

Exceptions

UndefinedElement

is raised if the DBCell corresponds to a missing value but the set of missing values passed in argument is empty.

type()

EltType gum::learning::DBCell::type ( ) const

noexcept

returns the current type of the DBCell

---

The documentation for this class was generated from the following file:

• agrum/tools/database/DBCell.h