csvf is a C++ library for csv file (especially large csv file) processing. Functionalities such as csv parsing, reading, writing, splitting, combining etc, are implemented as C++ classes, functions, and standalone executable programs. Various csv format can be automatically detected and parsed, including different separator, end of line, quote rules etc.

Table of Contents

• Installation
• Usage
  • Parse csv file
  • Write csv file
  • Split csv file
  • Combine csv files

Installation

1. Prerequisites
   • CMake
   • C++ compiler
   • Boost
2. Download source code

   git clone https://github.com/gongliyu/csvf

3. Build and install

   mkdir csvf-build
   cd csvf-build
   cmake -DCMAKE_INSTALL_PREFIX=/path/to/install /path/to/csvf/source
   make
   make install

Usage

Parse csv file

#include <iostream>
#include <iomanip>
#include <csvf/csvf.hpp>

int main(int argn, const char** argv)
{
    csvf::reader reader("abc.csv");
    while (reader) {
        auto record = reader.read_record();
        for (auto& field : record) {
            std::cout<<field<<"\t";
        }
        std::cout<<std::endl;
    }
}

The member function of csvf::reader::read_record() will return a record as std::vectorstd::string_view (if C++ compiler supports) or std::vectorboost::string_view.

Write csv file

#include <iostream>
#include <iomanip>
#include <csvf/csvf.hpp>

int main(int argn, const char** argv)
{
    csvf::reader writer("abc.csv");
    std::vector<std::string> record1{"hello", "hi", "nihao"};
    std::list<std::string> record2{"1","2","3"};
    writer.write_record(record1).write_record(record2);
    writer.close();
}

The member function csvf::writer::write_record could accept any container with element type std::string or std::string_view.

Split csv file

std::string infname("abc.csv");
std::vector<std::string> outfnames{"abc1.csv", "abc2.csv", "abc3.csv"};
csvf::split_by_record(infname, outfnames);

will split abc.csv by record, i.e. each result small csv file will contain different rows of the original file.

On the other hand,

csvf::split_by_field(infname, outfnames);

will split abc.csv by field, i.e. abc1.csv, abc2.csv and abc3.csv will contain different columns of the original file abc.csv.

There is also a standalone executable program csvf_split could be used to do the same job

csvf_split --input abc.csv --output abc1.csv abc2.csv abc3.csv --by record

Combine csv files

Just as split files, there are functions and standalone program to combine multiple csv files to a single csv file

csvf::combine_by_record(infnames, outfname);
csvf::combine_by_field(infnames, outfname);

csvf_combine --input abc1.csv abc2.csv abc3.csv --output abc.csv --by record