Table of Contents

• CSV Reader
  • Performance Benchmark
  • Reader API
• CSV Writer
  • Writer API
• Compiling Tests
• Generating Single Header
• Contributing
• License

CSV Reader

#include <csv2/reader.hpp>

int main() {
  csv2::Reader<csv2::delimiter<','>, 
               csv2::quote_character<'"'>, 
               csv2::first_row_is_header<true>,
               csv2::trim_policy::trim_whitespace> csv;
               
  if (csv.mmap("foo.csv")) {
    const auto header = csv.header();
    for (const auto row: csv) {
      for (const auto cell: row) {
        // Do something with cell value
        // std::string value;
        // cell.read_value(value);
      }
    }
  }
}

Performance Benchmark

This benchmark measures the average execution time (of 5 runs after 3 warmup runs) for csv2 to memory-map the input CSV file and iterate over every cell in the CSV. See benchmark/main.cpp for more details.

cd benchmark
g++ -I../include -O3 -std=c++11 -o main main.cpp
./main <csv_file>

System Details

Type	Value
Processor	Apple M1 Pro (8-core or 10-core CPU)
Installed RAM	16.0 GB
OS	MacOS Tahoe 26.6 (25G70)
C++ Compiler	Apple clang version 21.0.0 (clang-2100.3.34.2)

Results (as of 02 OCT 2026)

Dataset	File Size	Rows	Cols	Cells	Time
Denver Crime Data	111 MB	386,866	20	7,737,320	0.077s
AirBnb Paris Listings	196 MB	59,882	96	5,748,672	0.125s
2015 Flight Delays and Cancellations	574 MB	5,819,080	31	180,391,480	1.051s
StackLite: Stack Overflow questions	870 MB	17,203,825	7	120,426,775	0.629s
Used Cars Dataset	1.4 GB	426,881	26	11,098,906	0.683s
Title-Based Semantic Subject Indexing	3.7 GB	12,834,027	4	51,336,108	2.526s
Bitcoin tweets - 16M tweets	4 GB	16,889,766	9	152,007,894	3.026s

Reader API

Here is the public API available to you:

template <class delimiter = delimiter<','>, 
          class quote_character = quote_character<'"'>,
          class first_row_is_header = first_row_is_header<true>,
          class trim_policy = trim_policy::trim_whitespace>
class Reader {
public:
  
  // Use this if you'd like to mmap and read from file
  bool mmap(string_type filename);

  // Use this if you have the CSV contents in std::string already
  bool parse(string_type contents);

  // Shape
  size_t rows() const;
  size_t cols() const;
  
  // Row iterator
  // If first_row_is_header, row iteration will start
  // from the second row
  RowIterator begin() const;
  RowIterator end() const;

  // Access the first row of the CSV
  Row header() const;
};

Here's the Row class:

// Row class
class Row {
public:
  // Get raw contents of the row
  void read_raw_value(Container& value) const;
  
  // Cell iterator
  CellIterator begin() const;
  CellIterator end() const;
};

and here's the Cell class:

// Cell class
class Cell {
public:
  // Get raw contents of the cell
  void read_raw_value(Container& value) const;
  
  // Get converted contents of the cell
  // Handles escaped content, e.g., 
  // """foo""" => ""foo""
  void read_value(Container& value) const;
};

CSV Writer

This library also provides a basic csv2::Writer class - one that can be used to write CSV rows to file. Here's a basic usage:

#include <csv2/writer.hpp>
#include <vector>
#include <string>
using namespace csv2;

int main() {
    std::ofstream stream("foo.csv");
    Writer<delimiter<','>> writer(stream);

    std::vector<std::vector<std::string>> rows = 
        {
            {"a", "b", "c"},
            {"1", "2", "3"},
            {"4", "5", "6"}
        };

    writer.write_rows(rows);
    stream.close();
}

Writer API

Here is the public API available to you:

template <class delimiter = delimiter<','>>
class Writer {
public:
  
  // Construct using an std::ofstream
  Writer(output_file_stream stream);

  // Use this to write a single row to file
  void write_row(container_of_strings row);

  // Use this to write a list of rows to file
  void write_rows(container_of_rows rows);

Compiling Tests

mkdir build && cd build
cmake -DCSV2_BUILD_TESTS=ON ..
make
cd test
./csv2_test

Generating Single Header

python3 utils/amalgamate/amalgamate.py -c single_include.json -s .

Contributing

Contributions are welcome, have a look at the CONTRIBUTING.md document for more information.

License

The project is available under the MIT license.