Developer Guide and Reference

  • 2021.1
  • 12/04/2020
  • Public Content
Contents

CSV data source

Class
csv::data_source
is an API for accessing the data source represented as a csv file. CSV data source is used with
read
operation to extract data in text format from the given input file, process it using provided parameters (such as delimiter and read options), transform it into numerical representation, and store it as an in-memory dm dataset of a chosen type.
Supported type of in-memory object for
read
operation with CSV data source is
oneapi::dal::table
.
CSV data source requires input file name to be set in the constructor, while the other parameters of the constructor such as delimiter and read options rely on default values.

Usage example

using namespace oneapi; const auto data_source = dal::csv::data_source("data.csv", ','); const auto table = dal::read<dal::table>(data_source);

Programming Interface

All types and functions in this section are declared in the
oneapi::dal::csv
namespace and be available via inclusion of the
oneapi/dal/io/csv.hpp
header file.
enum class read_options : std::uint64_t { none = 0, parse_header = 1 << 0 }; constexpr char default_delimiter = ','; constexpr read_options default_read_options = read_options::none; class data_source { public: data_source(const char *file_name, char delimiter = default_delimiter, read_options opts = default_read_options); data_source(const std::string &file_name, char delimiter = default_delimiter, read_options opts = default_read_options); std::string get_file_name() const; char get_delimiter() const; read_options get_read_options() const; };
class
data_source
data_source
(
const
char *
file_name
, char
delimiter
= default_delimiter, read_options
opts
= default_read_options)
Creates a new instance of a CSV data source with the given , and read options flag.
data_source
(
const
std::string &
file_name
, char
delimiter
= default_delimiter, read_options
opts
= default_read_options)
Creates a new instance of a CSV data source with the given , and read options flag.
std::string
file_name
= ""
A string that contains the name of the file with the dataset to read.
Getter


std::string get_filename() const

char
delimiter
= default_delimiter
A character that represents the delimiter between separate features in the input file.
Getter


char get_delimter() const

read_options
options
= default_read_options
Value that stores read options to be applied during reading of the input file. Enabled
parse_header
option indicates that the first line in the input file is processed as a header record with features names.
Getter


read_options get_read_options() const

Reading
oneapi::dal::read<Object>(...)

Args
template <typename Object> class read_args { public: read_args(); };
template<typename
Object
>
class
read_args
read_args
()
Creates args for the read operation with the default attribute values.
Operation
oneapi::dal::table
is the only supported value of the
Object
template parameter for operation with CSV data source.
template<typename
Object
, typename
DataSource
> Object
read
(
const
DataSource &
ds
)
Template Parameters
  • Object
    – oneDAL object type that is produced as a result of reading from the data source.
  • DataSource
    – CSV data source .

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.