csv-parser: Csv::Parser (csv-parser) (original) (raw)

Compile-time and runtime CSV parser written in Modern C++

GitHub

Features

Header-only.
Requires only standard C++ (minimum C++17, with C++23 support providing additional functionality).
Supports reading CSV data from std::string_view at compile-time.
Fully supports RFC 4180, including quoted values, escaped quotes, and newlines in field values.
Liberal in terms of accepting not-quite-standard CSV files, but detects errors when needed.
Supports Excel CSV variations.
Supports reading data as different types (string, double, empty field), with some restrictions when parsing at compile-time.
Modular design allows for easy extension of the library.
Extensively tested using Catch2.

API Reference

Automatically generated API reference describes the public API in detail.

Usage Examples

Runtime Parsing into 2D std::vector

Example:

using namespace std::string_view_literals;

std::string_view data = "abc,def\n5,6"sv;

std::vector<std::vectorCsv::CellReference> cell_refs;

try {

}

std::cerr << "CSV parse error: " << ex.what() << std::endl;

return EXIT_FAILURE;

}

assert(cell_refs.size() == 2);

assert(cell_refs[0].size() == 2);

assert(cell_refs[1].size() == 2);

assert(cell_refs[0][0].getType() == Csv::CellType::String);

assert(cell_refs[1][0].getType() == Csv::CellType::String);

assert(cell_refs[0][1].getType() == Csv::CellType::Double);

assert(cell_refs[1][1].getType() == Csv::CellType::Double);

std::cout << "Column 0, row 0: " << cell_refs[0][0].getCleanString().value() << std::endl;

std::cout << "Column 1, row 0: " << cell_refs[1][0].getCleanString().value() << std::endl;

std::cout << "Column 0, row 1: " << cell_refs[0][1].getDouble().value() << std::endl;

std::cout << "Column 1, row 1: " << cell_refs[1][1].getDouble().value() << std::endl;

Exception thrown on CSV parse error.

Definition csv_error.h:25

The main CSV parser class.

Definition csv_parser.h:70

constexpr void parseTo2DVector(std::string_view data, Vector2D &values) const

Parse CSV string data into a vector of columns.

Definition csv_parser.h:484

Main include file for the library.

Runtime Parsing of Numeric Matrix Into 1D Vector With Row-Major Order

Example:

using namespace std::string_view_literals;

std::string_view data = "11,12,13\n21,22,23"sv;

std::vector matrix_data;

try {

}

std::cerr << "CSV parse error: " << ex.what() << std::endl;

return EXIT_FAILURE;

}

assert(matrix_data.size() == 3 * 2);

std::cout << "Row 0, column 0: " << matrix_data[0] << std::endl;

std::cout << "Row 0, column 1: " << matrix_data[1] << std::endl;

std::cout << "Row 0, column 2: " << matrix_data[2] << std::endl;

std::cout << "Row 1, column 0: " << matrix_data[info.matrixIndex(1, 0)] << std::endl;

std::cout << "Row 1, column 1: " << matrix_data[info.matrixIndex(1, 1)] << std::endl;

std::cout << "Row 1, column 2: " << matrix_data[info.matrixIndex(1, 2)] << std::endl;

return EXIT_SUCCESS;

Matrix information (dimensions, order).

Definition csv_matrix.h:32

constexpr std::size_t getRows() const

Get number of rows in matrix.

Definition csv_matrix.h:64

constexpr std::size_t getColumns() const

Get number of columns in matrix.

Definition csv_matrix.h:78

static constexpr std::size_t matrixIndex(std::size_t row, std::size_t column, std::size_t rows, std::size_t columns, MatrixOrder order)

Get index (offset) in flat-matrix vector.

Definition csv_matrix.h:42

constexpr MatrixInformation parseToVectorRowMajor(std::string_view data, Vector &values) const

Parse CSV string data into a flat matrix in row-major format (A11, A12, A13, A21, ....

Definition csv_parser.h:527

Compile-Time Parsing Into 2D std::array

Currently, parsing at compile-time has some restrictions:

To collapse consecutive double-quotes in strings, a compile-time-allocated buffer has to be used.
By default, only string_views are supported for output (no numeric types).
- If using C++23 or later, integral types are supported as well.
- If compile-time parsing of floating point numbers is needed, fast_float can be plugged in.

One (possibly useful) consequence of compile-time parsing is that a parse error also causes a compilation error.

Example:

using namespace std::string_view_literals;

constexpr std::size_t columns = 2, rows = 2;

constexpr std::string_view data =

R"(abc, "def"

"with ""quote inside",6)";

constexpr auto matrix = parser.parseTo2DArray<rows, columns>(data);

static_assert(matrix[0][0].getOriginalStringView() == "abc"sv);

static_assert(matrix[1][0].getOriginalStringView() == "def"sv);

static_assert(matrix[1][1].getOriginalStringView() == "6"sv);

constexpr auto buffer_size = matrix[0][1].getRequiredBufferSize();

constexpr auto buffer = matrix[0][1].getCleanStringBuffer<buffer_size>();

static_assert(buffer.getStringView() == R"(with "quote inside)"sv);

constexpr auto parseTo2DArray(std::string_view data) const

Parse CSV string to 2D std::array, an array of columns.

Definition csv_parser.h:507

Compile-Time Parsing of Integral Matrix Into 1D Vector

The library can be used to parse CSV data at compile-time into a 1D vector, in row-major or column-major order. Additionally, compile-time parsing of integral types is supported since C++23.

Example:

using namespace std::string_view_literals;

constexpr std::size_t columns = 2, rows = 3;

constexpr std::string_view data =

R"(11, -12

21, 4

60, -10)";

{

constexpr auto matrix = parser.parseToArray<rows, columns>(data, Csv::MatrixOrder::RowMajor);

static_assert(matrix[0].getOriginalStringView() == "11"sv);

static_assert(matrix[2].getOriginalStringView() == "21"sv);

}

#if __cplusplus >= 202300L

{

constexpr auto matrix = parser.parseToArray<rows, columns, std::int64_t>(data, Csv::MatrixOrder::ColumnMajor);

static_assert(matrix[0] == 11);

static_assert(matrix[2] == 60);

}

#endif

constexpr auto parseToArray(std::string_view data, MatrixOrder order) const

Parse CSV string to 1D std::array, stored in row-major or column-major order.

Definition csv_parser.h:635

Compile-Time Parsing of Numeric Matrix Using External Library for Floating Point Numbers

A library like fast_float can be used to parse floating point numbers at compile-time. This is done by creating a custom policy that implements the readNumber and create methods. This approach also allows for improving the performance of parsing floating point numbers compared to the standard library.

Example:

#include "fast_float/fast_float.h"

template

static constexpr std::optional readNumber(std::string_view cell)

{

Number parsed_value = 0;

auto [ptr, ec] = fast_float::from_chars(cell.begin(), cell.end(), parsed_value);

if (ec == std::errc() && ptr == (cell.end())) {

}

return std::nullopt;

}

template

{

return readNumber(cell).value_or(std::numeric_limits::quiet_NaN());

}

};

void parse()

{

constexpr std::string_view data =

R"(11.6, -12.3

2e5, -inf

nan, inf)";

constexpr std::size_t columns = 2, rows = 3;

constexpr auto matrix = parser.parseToArray<rows, columns, double>(data, Csv::MatrixOrder::RowMajor);

static_assert(matrix[0] == 11.6);

static_assert(matrix[3] == -std::numeric_limits::infinity());

}

CellTypeHint

Type hint associated with the cell to determine the type of the cell value.

Definition csv_cell.h:32

This policy controls the behavior of the parser.

Definition csv_policies.h:74

static constexpr std::optional< Number > readNumber(std::string_view cell)

Try to read a numeric value from string data.

Definition csv_policies.h:87

static constexpr CellT create(std::string_view cell, CellTypeHint hint)

Create an object of type CellT from cell contents represented as string_view.

Definition csv_policies.h:96

csv-parser: Csv::Parser (csv-parser) (original) (raw)

Features

API Reference

Usage Examples

Runtime Parsing into 2D std::vector

Example:

Runtime Parsing of Numeric Matrix Into 1D Vector With Row-Major Order

Example:

Compile-Time Parsing Into 2D std::array

Example:

Compile-Time Parsing of Integral Matrix Into 1D Vector

Example:

Compile-Time Parsing of Numeric Matrix Using External Library for Floating Point Numbers

Example:

Copyright