RDataFrame data source class for reading CSV files.

The RCsvDS class implements a CSV file reader for RDataFrame.

A RDataFrame that reads from a CSV file can be constructed using the factory method ROOT::RDF::FromCSV, which accepts five parameters:

Path to the CSV file.
Boolean that specifies whether the first row of the CSV file contains headers or not (optional, default true). If false, header names will be automatically generated as Col0, Col1, ..., ColN.
Delimiter (optional, default ',').
Chunk size (optional, default is -1 to read all) - number of lines to read at a time
Column Types (optional, default is an empty map). A map with column names as keys and their type (expressed as a single character, see below) as values.

The type of columns that do not appear in the map is inferred from the data. The supported types are:

Integer: stored as a 64-bit long long int; can be specified in the column types map with 'L'.
Floating point number: stored with double precision; specified with 'D'.
Boolean: matches the literals true and false; specified with 'O'.
String: stored as an std::string, matches anything that does not fall into any of the previous types; specified with 'T'.

These are some formatting rules expected by the RCsvDS implementation:

All records must have the same number of fields, in the same order.
Any field may be quoted.
"1997","Ford","E350"
Fields with embedded delimiters (e.g. comma) must be quoted.
1997,Ford,E350,"Super, luxurious truck"
Fields with double-quote characters must be quoted, and each of the embedded double-quote characters must be represented by a pair of double-quote characters.
1997,Ford,E350,"Super, ""luxurious"" truck"
Fields with embedded line breaks are not supported, even when quoted.
1997,Ford,E350,"Go get one now

they are going fast"
Spaces are considered part of a field and are not ignored.
1997, Ford , E350

not same as

1997,Ford,E350

but same as

1997, "Ford" , E350
If a header row is provided, it must contain column names for each of the fields.
Year,Make,Model

1997,Ford,E350

2000,Mercury,Cougar

The current implementation of RCsvDS reads the entire CSV file content into memory before RDataFrame starts processing it. Therefore, before creating a CSV RDataFrame, it is important to check both how much memory is available and the size of the CSV file.

RCsvDS can handle empty cells and also allows the usage of the special keywords "NaN" and "nan" to indicate nan values. If the column is of type double, these cells are stored internally as nan. Empty cells and explicit nan-s inside columns of type Long64_t/bool are stored as zeros.

Definition at line 35 of file RCsvDS.hxx.

Classes
struct	ROptions
	Options that control how the CSV file is parsed. More...

Public Member Functions
	RCsvDS (std::string_view fileName, bool readHeaders=true, char delimiter=',', Long64_t linesChunkSize=-1LL, std::unordered_map< std::string, char > &&colTypes={})
	Constructor to create a CSV RDataSource for RDataFrame.

	RCsvDS (std::string_view fileName, const ROptions &options)
	Constructor to create a CSV RDataSource for RDataFrame.

	~RCsvDS ()
	Destructor.

void	Finalize () final
	Convenience method called after concluding an event-loop.

const std::vector< std::string > &	GetColumnNames () const final
	Returns a reference to the collection of the dataset's column names.

std::vector< std::pair< ULong64_t, ULong64_t > >	GetEntryRanges () final
	Return ranges of entries to distribute to tasks.

std::string	GetLabel () final
	Return a string representation of the datasource type.

std::size_t	GetNFiles () const final
	Returns the number of files from which the dataset is constructed.

std::string	GetTypeName (std::string_view colName) const final
	Type of a column as a string, e.g.

bool	HasColumn (std::string_view colName) const final
	Checks if the dataset has a certain column.

bool	SetEntry (unsigned int slot, ULong64_t entry) final
	Advance the "cursors" returned by GetColumnReaders to the selected entry for a particular slot.

void	SetNSlots (unsigned int nSlots) final
	Inform RDataSource of the number of processing slots (i.e.

Public Member Functions inherited from ROOT::RDF::RDataSource
virtual	~RDataSource ()=default

virtual void	FinalizeSlot (unsigned int)
	Convenience method called at the end of the data processing associated to a slot.

template<typename T >
std::vector< T ** >	GetColumnReaders (std::string_view columnName)
	Called at most once per column by RDF.

virtual std::unique_ptr< ROOT::Detail::RDF::RColumnReaderBase >	GetColumnReaders (unsigned int, std::string_view, const std::type_info &)
	If the other GetColumnReaders overload returns an empty vector, this overload will be called instead.

virtual void	Initialize ()
	Convenience method called before starting an event-loop.

virtual void	InitSlot (unsigned int, ULong64_t)
	Convenience method called at the start of the data processing associated to a slot.

Protected Member Functions
std::string	AsString () final

Private Types
using	ColType_t = char

Private Member Functions
void	Construct ()

void	FillHeaders (const std::string &)

void	FillRecord (const std::string &, Record_t &)

void	FreeRecords ()

void	GenerateHeaders (size_t)

std::vector< void * >	GetColumnReadersImpl (std::string_view, const std::type_info &) final
	type-erased vector of pointers to pointers to column values - one per slot

ColType_t	GetType (std::string_view colName) const

void	InferColTypes (std::vector< std::string > &)

void	InferType (const std::string &, unsigned int)

std::vector< std::string >	ParseColumns (const std::string &)

size_t	ParseValue (const std::string &, std::vector< std::string > &, size_t)

bool	Readln (std::string &line)

void	RewindToData ()

void	ValidateColTypes (std::vector< std::string > &) const

Private Attributes
std::vector< std::deque< bool > >	fBoolEvtValues

std::vector< std::vector< void * > >	fColAddresses

std::set< std::string >	fColContainingEmpty

std::unordered_map< std::string, ColType_t >	fColTypes

std::list< ColType_t >	fColTypesList

std::unique_ptr< ROOT::Internal::RRawFile >	fCsvFile

std::int64_t	fDataLineNumber = 0

std::uint64_t	fDataPos = 0

std::vector< std::vector< double > >	fDoubleEvtValues

ULong64_t	fEntryRangesRequested = 0ULL

std::vector< std::string >	fHeaders

std::int64_t	fLineNumber = 0

std::vector< std::vector< Long64_t > >	fLong64EvtValues

std::int64_t	fMaxLineNumber = -1

unsigned int	fNSlots = 0U

ROptions	fOptions

ULong64_t	fProcessedLines = 0ULL

std::vector< Record_t >	fRecords

std::vector< std::vector< std::string > >	fStringEvtValues

Static Private Attributes
static const std::unordered_map< ColType_t, std::string >	fgColTypeMap

static const TRegexp	fgDoubleRegex1

static const TRegexp	fgDoubleRegex2

static const TRegexp	fgDoubleRegex3

static const TRegexp	fgFalseRegex

static const TRegexp	fgIntRegex

static const TRegexp	fgTrueRegex

Additional Inherited Members
Protected Types inherited from ROOT::RDF::RDataSource
using	Record_t = std::vector< void * >

#include <ROOT/RCsvDS.hxx>

Inheritance diagram for ROOT::RDF::RCsvDS:

[legend]

Member Typedef Documentation

◆ ColType_t

using ROOT::RDF::RCsvDS::ColType_t = char

private

Definition at line 64 of file RCsvDS.hxx.

Constructor & Destructor Documentation

◆ RCsvDS() [1/2]

ROOT::RDF::RCsvDS::RCsvDS	(	std::string_view	fileName,
		const ROptions &	options
	)

Constructor to create a CSV RDataSource for RDataFrame.

Parameters

[in]	fileName	Path or URL of the CSV file.
[in]	options	File parsing settings

Definition at line 450 of file RCsvDS.cxx.

◆ RCsvDS() [2/2]

ROOT::RDF::RCsvDS::RCsvDS	(	std::string_view	fileName,
		bool	readHeaders = `true`,
		char	delimiter = `','`,
		Long64_t	linesChunkSize = `-1LL`,
		std::unordered_map< std::string, char > &&	colTypes = `{}`
	)

Constructor to create a CSV RDataSource for RDataFrame.

Parameters

[in]	fileName	Path or URL of the CSV file.
[in]	readHeaders	`true` if the CSV file contains headers as first row, `false` otherwise (default `true`).
[in]	delimiter	Delimiter character (default ',').
[in]	linesChunkSize	bunch of lines to read, use -1 to read all
[in]	colTypes	Allows users to manually specify column types. Accepts an unordered map with keys being column names, values being type specifiers ('O' for boolean, 'D' for double, 'L' for Long64_t, 'T' for std::string)

Definition at line 468 of file RCsvDS.cxx.

◆ ~RCsvDS()

ROOT::RDF::RCsvDS::~RCsvDS ( )

Destructor.

Definition at line 510 of file RCsvDS.cxx.

Member Function Documentation

◆ AsString()

std::string ROOT::RDF::RCsvDS::AsString ( )

finalprotectedvirtual

Reimplemented from ROOT::RDF::RDataSource.

Definition at line 98 of file RCsvDS.cxx.

◆ Construct()

void ROOT::RDF::RCsvDS::Construct ( )

private

Definition at line 385 of file RCsvDS.cxx.

◆ FillHeaders()

void ROOT::RDF::RCsvDS::FillHeaders ( const std::string & line )

private

Definition at line 165 of file RCsvDS.cxx.

◆ FillRecord()

void ROOT::RDF::RCsvDS::FillRecord	(	const std::string &	line,
		Record_t &	record
	)

private

Definition at line 179 of file RCsvDS.cxx.

◆ Finalize()

void ROOT::RDF::RCsvDS::Finalize ( )

finalvirtual

Convenience method called after concluding an event-loop.

See Initialize for more details.

Reimplemented from ROOT::RDF::RDataSource.

Definition at line 515 of file RCsvDS.cxx.

◆ FreeRecords()

void ROOT::RDF::RCsvDS::FreeRecords ( )

private

Definition at line 479 of file RCsvDS.cxx.

◆ GenerateHeaders()

void ROOT::RDF::RCsvDS::GenerateHeaders ( size_t size )

private

Definition at line 222 of file RCsvDS.cxx.

◆ GetColumnNames()

const std::vector< std::string > & ROOT::RDF::RCsvDS::GetColumnNames ( ) const

finalvirtual

Returns a reference to the collection of the dataset's column names.

Implements ROOT::RDF::RDataSource.

Definition at line 523 of file RCsvDS.cxx.

◆ GetColumnReadersImpl()

std::vector< void * > ROOT::RDF::RCsvDS::GetColumnReadersImpl	(	std::string_view	name,
		const std::type_info &
	)

finalprivatevirtual

type-erased vector of pointers to pointers to column values - one per slot

Implements ROOT::RDF::RDataSource.

Definition at line 235 of file RCsvDS.cxx.

◆ GetEntryRanges()

std::vector< std::pair< ULong64_t, ULong64_t > > ROOT::RDF::RCsvDS::GetEntryRanges ( )

finalvirtual

Return ranges of entries to distribute to tasks.

They are required to be contiguous intervals with no entries skipped. Supposing a dataset with nEntries, the intervals must start at 0 and end at nEntries, e.g. [0-5],[5-10] for 10 entries. This function will be invoked repeatedly by RDataFrame as it needs additional entries to process. The same entry range should not be returned more than once. Returning an empty collection of ranges signals to RDataFrame that the processing can stop.

Implements ROOT::RDF::RDataSource.

Definition at line 528 of file RCsvDS.cxx.

◆ GetLabel()

std::string ROOT::RDF::RCsvDS::GetLabel ( )

finalvirtual

Return a string representation of the datasource type.

The returned string will be used by ROOT::RDF::SaveGraph() to represent the datasource in the visualization of the computation graph. Concrete datasources can override the default implementation.

Reimplemented from ROOT::RDF::RDataSource.

Definition at line 655 of file RCsvDS.cxx.

◆ GetNFiles()

std::size_t ROOT::RDF::RCsvDS::GetNFiles ( ) const

inlinefinalvirtual

Returns the number of files from which the dataset is constructed.

Reimplemented from ROOT::RDF::RDataSource.

Definition at line 117 of file RCsvDS.hxx.

◆ GetType()

RCsvDS::ColType_t ROOT::RDF::RCsvDS::GetType ( std::string_view colName ) const

private

Definition at line 586 of file RCsvDS.cxx.

◆ GetTypeName()

std::string ROOT::RDF::RCsvDS::GetTypeName ( std::string_view colName ) const

finalvirtual

Type of a column as a string, e.g.

GetTypeName("x") == "double". Required for jitting e.g. df.Filter("x>0").

Parameters

[in] colName The name of the column

Implements ROOT::RDF::RDataSource.

Definition at line 597 of file RCsvDS.cxx.

◆ HasColumn()

bool ROOT::RDF::RCsvDS::HasColumn ( std::string_view colName ) const

finalvirtual

Checks if the dataset has a certain column.

Parameters

[in] colName The name of the column

Implements ROOT::RDF::RDataSource.

Definition at line 602 of file RCsvDS.cxx.

◆ InferColTypes()

void ROOT::RDF::RCsvDS::InferColTypes ( std::vector< std::string > & columns )

private

Definition at line 286 of file RCsvDS.cxx.

◆ InferType()

void ROOT::RDF::RCsvDS::InferType	(	const std::string &	col,
		unsigned int	idxCol
	)

private

Definition at line 319 of file RCsvDS.cxx.

◆ ParseColumns()

std::vector< std::string > ROOT::RDF::RCsvDS::ParseColumns ( const std::string & line )

private

Definition at line 340 of file RCsvDS.cxx.

◆ ParseValue()

size_t ROOT::RDF::RCsvDS::ParseValue	(	const std::string &	line,
		std::vector< std::string > &	columns,
		size_t	i
	)

private

Definition at line 351 of file RCsvDS.cxx.

◆ Readln()

bool ROOT::RDF::RCsvDS::Readln ( std::string & line )

private

Definition at line 114 of file RCsvDS.cxx.

◆ RewindToData()

void ROOT::RDF::RCsvDS::RewindToData ( )

private

Definition at line 159 of file RCsvDS.cxx.

◆ SetEntry()

bool ROOT::RDF::RCsvDS::SetEntry	(	unsigned int	slot,
		ULong64_t	entry
	)

finalvirtual

Advance the "cursors" returned by GetColumnReaders to the selected entry for a particular slot.

Parameters

[in]	slot	The data processing slot that needs to be considered
[in]	entry	The entry which needs to be pointed to by the reader pointers Slots are adopted to accommodate parallel data processing. Different workers will loop over different ranges and will be labelled by different "slot" values. Returns true if the entry has to be processed, false otherwise.

Implements ROOT::RDF::RDataSource.

Definition at line 607 of file RCsvDS.cxx.

◆ SetNSlots()

void ROOT::RDF::RCsvDS::SetNSlots ( unsigned int nSlots )

finalvirtual

Inform RDataSource of the number of processing slots (i.e.

worker threads) used by the associated RDataFrame. Slots numbers are used to simplify parallel execution: RDataFrame guarantees that different threads will always pass different slot values when calling methods concurrently.

Implements ROOT::RDF::RDataSource.

Definition at line 638 of file RCsvDS.cxx.

◆ ValidateColTypes()

void ROOT::RDF::RCsvDS::ValidateColTypes ( std::vector< std::string > & columns ) const

private

Definition at line 267 of file RCsvDS.cxx.

Member Data Documentation

◆ fBoolEvtValues

std::vector<std::deque<bool> > ROOT::RDF::RCsvDS::fBoolEvtValues

private

Definition at line 90 of file RCsvDS.hxx.

◆ fColAddresses

std::vector<std::vector<void *> > ROOT::RDF::RCsvDS::fColAddresses

private

Definition at line 83 of file RCsvDS.hxx.

◆ fColContainingEmpty

std::set<std::string> ROOT::RDF::RCsvDS::fColContainingEmpty

private

Definition at line 81 of file RCsvDS.hxx.

◆ fColTypes

std::unordered_map<std::string, ColType_t> ROOT::RDF::RCsvDS::fColTypes

private

Definition at line 80 of file RCsvDS.hxx.

◆ fColTypesList

std::list<ColType_t> ROOT::RDF::RCsvDS::fColTypesList

private

Definition at line 82 of file RCsvDS.hxx.

◆ fCsvFile

std::unique_ptr<ROOT::Internal::RRawFile> ROOT::RDF::RCsvDS::fCsvFile

private

Definition at line 76 of file RCsvDS.hxx.

◆ fDataLineNumber

std::int64_t ROOT::RDF::RCsvDS::fDataLineNumber = 0

private

Definition at line 72 of file RCsvDS.hxx.

◆ fDataPos

std::uint64_t ROOT::RDF::RCsvDS::fDataPos = 0

private

Definition at line 71 of file RCsvDS.hxx.

◆ fDoubleEvtValues

std::vector<std::vector<double> > ROOT::RDF::RCsvDS::fDoubleEvtValues

private

Definition at line 85 of file RCsvDS.hxx.

◆ fEntryRangesRequested

ULong64_t ROOT::RDF::RCsvDS::fEntryRangesRequested = 0ULL

private

Definition at line 77 of file RCsvDS.hxx.

◆ fgColTypeMap

const std::unordered_map< RCsvDS::ColType_t, std::string > ROOT::RDF::RCsvDS::fgColTypeMap

staticprivate

Definition at line 65 of file RCsvDS.hxx.