How can I read and parse CSV files in C++?

Viewed 466074

I need to load and use CSV file data in C++. At this point it can really just be a comma-delimited parser (ie don't worry about escaping new lines and commas). The main need is a line-by-line parser that will return a vector for the next line each time the method is called.

I found this article which looks quite promising: http://www.boost.org/doc/libs/1_35_0/libs/spirit/example/fundamental/list_parser.cpp

I've never used Boost's Spirit, but am willing to try it. But only if there isn't a more straightforward solution I'm overlooking.

39 Answers

If you don't care about escaping comma and newline,
AND you can't embed comma and newline in quotes (If you can't escape then...)
then its only about three lines of code (OK 14 ->But its only 15 to read the whole file).

std::vector<std::string> getNextLineAndSplitIntoTokens(std::istream& str)
{
    std::vector<std::string>   result;
    std::string                line;
    std::getline(str,line);

    std::stringstream          lineStream(line);
    std::string                cell;

    while(std::getline(lineStream,cell, ','))
    {
        result.push_back(cell);
    }
    // This checks for a trailing comma with no data after it.
    if (!lineStream && cell.empty())
    {
        // If there was a trailing comma then add an empty element.
        result.push_back("");
    }
    return result;
}

I would just create a class representing a row.
Then stream into that object:

#include <iterator>
#include <iostream>
#include <fstream>
#include <sstream>
#include <vector>
#include <string>

class CSVRow
{
    public:
        std::string_view operator[](std::size_t index) const
        {
            return std::string_view(&m_line[m_data[index] + 1], m_data[index + 1] -  (m_data[index] + 1));
        }
        std::size_t size() const
        {
            return m_data.size() - 1;
        }
        void readNextRow(std::istream& str)
        {
            std::getline(str, m_line);

            m_data.clear();
            m_data.emplace_back(-1);
            std::string::size_type pos = 0;
            while((pos = m_line.find(',', pos)) != std::string::npos)
            {
                m_data.emplace_back(pos);
                ++pos;
            }
            // This checks for a trailing comma with no data after it.
            pos   = m_line.size();
            m_data.emplace_back(pos);
        }
    private:
        std::string         m_line;
        std::vector<int>    m_data;
};

std::istream& operator>>(std::istream& str, CSVRow& data)
{
    data.readNextRow(str);
    return str;
}   
int main()
{
    std::ifstream       file("plop.csv");

    CSVRow              row;
    while(file >> row)
    {
        std::cout << "4th Element(" << row[3] << ")\n";
    }
}

But with a little work we could technically create an iterator:

class CSVIterator
{   
    public:
        typedef std::input_iterator_tag     iterator_category;
        typedef CSVRow                      value_type;
        typedef std::size_t                 difference_type;
        typedef CSVRow*                     pointer;
        typedef CSVRow&                     reference;

        CSVIterator(std::istream& str)  :m_str(str.good()?&str:nullptr) { ++(*this); }
        CSVIterator()                   :m_str(nullptr) {}

        // Pre Increment
        CSVIterator& operator++()               {if (m_str) { if (!((*m_str) >> m_row)){m_str = nullptr;}}return *this;}
        // Post increment
        CSVIterator operator++(int)             {CSVIterator    tmp(*this);++(*this);return tmp;}
        CSVRow const& operator*()   const       {return m_row;}
        CSVRow const* operator->()  const       {return &m_row;}

        bool operator==(CSVIterator const& rhs) {return ((this == &rhs) || ((this->m_str == nullptr) && (rhs.m_str == nullptr)));}
        bool operator!=(CSVIterator const& rhs) {return !((*this) == rhs);}
    private:
        std::istream*       m_str;
        CSVRow              m_row;
};


int main()
{
    std::ifstream       file("plop.csv");

    for(CSVIterator loop(file); loop != CSVIterator(); ++loop)
    {
        std::cout << "4th Element(" << (*loop)[3] << ")\n";
    }
}

Now that we are in 2020 lets add a CSVRange object:

class CSVRange
{
    std::istream&   stream;
    public:
        CSVRange(std::istream& str)
            : stream(str)
        {}
        CSVIterator begin() const {return CSVIterator{stream};}
        CSVIterator end()   const {return CSVIterator{};}
};

int main()
{
    std::ifstream       file("plop.csv");

    for(auto& row: CSVRange(file))
    {
        std::cout << "4th Element(" << row[3] << ")\n";
    }
}

Solution using Boost Tokenizer:

std::vector<std::string> vec;
using namespace boost;
tokenizer<escaped_list_separator<char> > tk(
   line, escaped_list_separator<char>('\\', ',', '\"'));
for (tokenizer<escaped_list_separator<char> >::iterator i(tk.begin());
   i!=tk.end();++i) 
{
   vec.push_back(*i);
}

You can use Boost Tokenizer with escaped_list_separator.

escaped_list_separator parses a superset of the csv. Boost::tokenizer

This only uses Boost tokenizer header files, no linking to boost libraries required.

Here is an example, (see Parse CSV File With Boost Tokenizer In C++ for details or Boost::tokenizer ):

#include <iostream>     // cout, endl
#include <fstream>      // fstream
#include <vector>
#include <string>
#include <algorithm>    // copy
#include <iterator>     // ostream_operator
#include <boost/tokenizer.hpp>

int main()
{
    using namespace std;
    using namespace boost;
    string data("data.csv");

    ifstream in(data.c_str());
    if (!in.is_open()) return 1;

    typedef tokenizer< escaped_list_separator<char> > Tokenizer;
    vector< string > vec;
    string line;

    while (getline(in,line))
    {
        Tokenizer tok(line);
        vec.assign(tok.begin(),tok.end());

        // vector now contains strings from one row, output to cout here
        copy(vec.begin(), vec.end(), ostream_iterator<string>(cout, "|"));

        cout << "\n----------------------" << endl;
    }
}

If you DO care about parsing CSV correctly, this will do it...relatively slowly as it works one char at a time.

 void ParseCSV(const string& csvSource, vector<vector<string> >& lines)
    {
       bool inQuote(false);
       bool newLine(false);
       string field;
       lines.clear();
       vector<string> line;

       string::const_iterator aChar = csvSource.begin();
       while (aChar != csvSource.end())
       {
          switch (*aChar)
          {
          case '"':
             newLine = false;
             inQuote = !inQuote;
             break;

          case ',':
             newLine = false;
             if (inQuote == true)
             {
                field += *aChar;
             }
             else
             {
                line.push_back(field);
                field.clear();
             }
             break;

          case '\n':
          case '\r':
             if (inQuote == true)
             {
                field += *aChar;
             }
             else
             {
                if (newLine == false)
                {
                   line.push_back(field);
                   lines.push_back(line);
                   field.clear();
                   line.clear();
                   newLine = true;
                }
             }
             break;

          default:
             newLine = false;
             field.push_back(*aChar);
             break;
          }

          aChar++;
       }

       if (field.size())
          line.push_back(field);

       if (line.size())
          lines.push_back(line);
    }

You might want to look at my FOSS project CSVfix (updated link), which is a CSV stream editor written in C++. The CSV parser is no prize, but does the job and the whole package may do what you need without you writing any code.

See alib/src/a_csv.cpp for the CSV parser, and csvlib/src/csved_ioman.cpp (IOManager::ReadCSV) for a usage example.

You can use the header-only Csv::Parser library.

  • It fully supports RFC 4180, including quoted values, escaped quotes, and newlines in field values.
  • It requires only standard C++ (C++17).
  • It supports reading CSV data from std::string_view at compile-time.
  • It's extensively tested using Catch2.

Excuse me, but this all seems like a great deal of elaborate syntax to hide a few lines of code.

Why not this:

/**

  Read line from a CSV file

  @param[in] fp file pointer to open file
  @param[in] vls reference to vector of strings to hold next line

  */
void readCSV( FILE *fp, std::vector<std::string>& vls )
{
    vls.clear();
    if( ! fp )
        return;
    char buf[10000];
    if( ! fgets( buf,999,fp) )
        return;
    std::string s = buf;
    int p,q;
    q = -1;
    // loop over columns
    while( 1 ) {
        p = q;
        q = s.find_first_of(",\n",p+1);
        if( q == -1 ) 
            break;
        vls.push_back( s.substr(p+1,q-p-1) );
    }
}

int _tmain(int argc, _TCHAR* argv[])
{
    std::vector<std::string> vls;
    FILE * fp = fopen( argv[1], "r" );
    if( ! fp )
        return 1;
    readCSV( fp, vls );
    readCSV( fp, vls );
    readCSV( fp, vls );
    std::cout << "row 3, col 4 is " << vls[3].c_str() << "\n";

    return 0;
}

Since i'm not used to boost right now, I will suggest a more simple solution. Lets suppose that your .csv file has 100 lines with 10 numbers in each line separated by a ','. You could load this data in the form of an array with the following code:

#include <iostream>
#include <fstream>
#include <sstream>
#include <string>
using namespace std;

int main()
{
    int A[100][10];
    ifstream ifs;
    ifs.open("name_of_file.csv");
    string s1;
    char c;
    for(int k=0; k<100; k++)
    {
        getline(ifs,s1);
        stringstream stream(s1);
        int j=0;
        while(1)
        {
            stream >>A[k][j];
            stream >> c;
            j++;
            if(!stream) {break;}
        }
    }


}

Parsing CSV file lines with Stream

I wrote a small example of parsing CSV file lines, it can be developed with for and while loops if desired:

#include <iostream>
#include <fstream>
#include <string.h>

using namespace std;

int main() {


ifstream fin("Infile.csv");
ofstream fout("OutFile.csv");
string strline, strremain, strCol1 , strout;

string delimeter =";";

int d1;

to continue until the end of the file:

while (!fin.eof()){ 

get first line from InFile :

    getline(fin,strline,'\n');      

find delimeter position in line:

    d1 = strline.find(';');

and parse first column:

    strCol1 = strline.substr(0,d1); // parse first Column
    d1++;
    strremain = strline.substr(d1); // remaining line

create output line in CSV format:

    strout.append(strCol1);
    strout.append(delimeter);

write line to Out File:

    fout << strout << endl; //out file line

} 

fin.close();
fout.close();

return(0);
}

This code is compiled and running. Good luck!

My simple and fast contribution, if I may. No Boost.

Accepts delimiters and delimiters within delimiters, as long as in pairs or away from a separator.

#include <iostream>
#include <vector>
#include <fstream>

std::vector<std::string> SplitCSV(const std::string &data, char separator, char delimiter)
{
  std::vector<std::string> Values;
  std::string Val = "";
  bool VDel = false; // Is within delimiter?
  size_t CDel = 0; // Delimiters counter within delimiters.
  const char *C = data.c_str();
  size_t P = 0;
  do
  {
    if ((Val.length() == 0) && (C[P] == delimiter))
    {
      VDel = !VDel;
      CDel = 0;
      P++;
      continue;
    }
    if (VDel)
    {
      if (C[P] == delimiter)
      {
        if (((CDel % 2) == 0) && ( (C[P+1] == separator) || (C[P+1] == 0) || (C[P+1] == '\n') || (C[P+1] == '\r') ))
        {
          VDel = false;
          CDel = 0;
          P++;
          continue;
        }
        else
          CDel++;
      }
    }
    else
    {
      if (C[P] == separator)
      {
        Values.push_back(Val);
        Val = "";
        P++;
        continue;
      }
      if ((C[P] == 0) || (C[P] == '\n') || (C[P] == '\r'))
        break;
    }
    Val += C[P];
    P++;
  } while(P < data.length());
  Values.push_back(Val);
  return Values;
}

bool ReadCsv(const std::string &fname, std::vector<std::vector<std::string>> &data,
  char separator = ',', char delimiter = '\"')
{
  bool Ret = false;
  std::ifstream FCsv(fname);
  if (FCsv)
  {
    FCsv.seekg(0, FCsv.end);
    size_t Siz = FCsv.tellg();
    if (Siz > 0)
    {
      FCsv.seekg(0);
      data.clear();
      std::string Line;
      while (getline(FCsv, Line, '\n'))
        data.push_back(SplitCSV(Line, separator, delimiter));
      Ret = true;
    }
    FCsv.close();
  }
  return Ret;
}

int main(int argc, char *argv[])
{
  std::vector<std::vector<std::string>> Data;
  ReadCsv("fsample.csv", Data);
  return 0;
}

If you're using Visual Studio / MFC, the following solution may make your life easier. It supports both Unicode and MBCS, has comments, doesn't have dependencies other than CString, and works well enough for me. It doesn't support line breaks embedded within a quoted string, but I don't care so long as it doesn't crash in that case, which it doesn't.

The overall strategy is, handle quoted and empty strings as special cases, and use Tokenize for the rest. For quoted strings, the strategy is, find the real closing quote, keeping track of whether pairs of consecutive quotes were encountered. If they were, use Replace to convert the pairs to singles. No doubt there are more efficient methods but performance wasn't sufficiently critical in my case to justify further optimization.

class CParseCSV {
public:
// Construction
    CParseCSV(const CString& sLine);

// Attributes
    bool    GetString(CString& sDest);

protected:
    CString m_sLine;    // line to extract tokens from
    int     m_nLen;     // line length in characters
    int     m_iPos;     // index of current position
};

CParseCSV::CParseCSV(const CString& sLine) : m_sLine(sLine)
{
    m_nLen = m_sLine.GetLength();
    m_iPos = 0;
}

bool CParseCSV::GetString(CString& sDest)
{
    if (m_iPos < 0 || m_iPos > m_nLen)  // if position out of range
        return false;
    if (m_iPos == m_nLen) { // if at end of string
        sDest.Empty();  // return empty token
        m_iPos = -1;    // really done now
        return true;
    }
    if (m_sLine[m_iPos] == '\"') {  // if current char is double quote
        m_iPos++;   // advance to next char
        int iTokenStart = m_iPos;
        bool    bHasEmbeddedQuotes = false;
        while (m_iPos < m_nLen) {   // while more chars to parse
            if (m_sLine[m_iPos] == '\"') {  // if current char is double quote
                // if next char exists and is also double quote
                if (m_iPos < m_nLen - 1 && m_sLine[m_iPos + 1] == '\"') {
                    // found pair of consecutive double quotes
                    bHasEmbeddedQuotes = true;  // request conversion
                    m_iPos++;   // skip first quote in pair
                } else  // next char doesn't exist or is normal
                    break;  // found closing quote; exit loop
            }
            m_iPos++;   // advance to next char
        }
        sDest = m_sLine.Mid(iTokenStart, m_iPos - iTokenStart);
        if (bHasEmbeddedQuotes) // if string contains embedded quote pairs
            sDest.Replace(_T("\"\""), _T("\""));    // convert pairs to singles
        m_iPos += 2;    // skip closing quote and trailing delimiter if any
    } else if (m_sLine[m_iPos] == ',') {    // else if char is comma
        sDest.Empty();  // return empty token
        m_iPos++;   // advance to next char
    } else {    // else get next comma-delimited token
        sDest = m_sLine.Tokenize(_T(","), m_iPos);
    }
    return true;
}

// calling code should look something like this:

    CStdioFile  fIn(pszPath, CFile::modeRead);
    CString sLine, sToken;
    while (fIn.ReadString(sLine)) { // for each line of input file
        if (!sLine.IsEmpty()) { // ignore blank lines
            CParseCSV   csv(sLine);
            while (csv.GetString(sToken)) {
                // do something with sToken here
            }
        }
    }

I've got a way quicker solution, was originally intended for this question:

How to pull specific part of different strings?

But it was closed obviously. I'm not about to throw this away though:

#include <iostream>
#include <string>
#include <regex>

std::string text = "\"4,\"\"3\"\",\"\"Mon May 11 03:17:40 UTC 2009\"\",\"\"kindle2\"\",\"\"tpryan\"\",\"\"TEXT HERE\"\"\";;;;";

int main()
{
    std::regex r("(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")(\".*\")");
    std::smatch m;
    std::regex_search(text, m, r);
    std::cout<<"FOUND: "<<m[9]<<std::endl;

    return 0;
}

Just pick out whichever match you want from the smatch collection by index. Regex is bliss.

Like everyone puts his solution, here is mine using template, lambda and tuple.

It can convert any CSV with wanted columns to a C++ vector of tuple.

It works by defining each CSV line element type in a tuple.

You also need to define the std::string to type conversion Formatter lambda for each element (using std::atod for example).

Then you got a vector of this struct corresponding to your CSV data.

You can reuse this easily to match any CSV structure.

StringsHelpers.hpp

#include <string>
#include <fstream>
#include <vector>
#include <functional>

namespace StringHelpers
{
    template<typename Tuple>
    using Formatter = std::function<Tuple(const std::vector<std::string> &)>;

    std::vector<std::string> split(const std::string &string, const std::string &delimiter);

    template<typename Tuple>
    std::vector<Tuple> readCsv(const std::string &path, const std::string &delimiter, Formatter<Tuple> formatter);
};

StringsHelpers.cpp

#include "StringHelpers.hpp"

namespace StringHelpers
{
    /**
     * Split a string with the given delimiter into several strings
     *
     * @param string - The string to extract the substrings from
     * @param delimiter - The substrings delimiter
     *
     * @return The substrings
     */
    std::vector<std::string> split(const std::string &string, const std::string &delimiter)
    {
        std::vector<std::string> result;
        size_t                   last = 0,
                                 next = 0;

        while ((next = string.find(delimiter, last)) != std::string::npos) {
            result.emplace_back(string.substr(last, next - last));
            last = next + 1;
        }

        result.emplace_back(string.substr(last));

        return result;
    }

    /**
     * Read a CSV file and store its values into the given structure (Tuple with Formatter constructor)
     *
     * @tparam Tuple - The CSV line structure format
     *
     * @param path - The CSV file path
     * @param delimiter - The CSV values delimiter
     * @param formatter - The CSV values formatter that take a vector of strings in input and return a Tuple
     *
     * @return The CSV as vector of Tuple
     */
    template<typename Tuple>
    std::vector<Tuple> readCsv(const std::string &path, const std::string &delimiter, Formatter<Tuple> formatter)
    {
        std::ifstream      file(path, std::ifstream::in);
        std::string        line;
        std::vector<Tuple> result;

        if (file.fail()) {
            throw std::runtime_error("The file " + path + " could not be opened");
        }

        while (std::getline(file, line)) {
            result.emplace_back(formatter(split(line, delimiter)));
        }

        file.close();

        return result;
    }

    // Forward template declarations

    template std::vector<std::tuple<double, double, double>> readCsv<std::tuple<double, double, double>>(const std::string &, const std::string &, Formatter<std::tuple<double, double, double>>);
} // End of StringHelpers namespace

main.cpp (some usage)

#include "StringHelpers.hpp"

/**
 * Example of use with a CSV file which have (number,Red,Green,Blue) as line values. We do not want to use the 1st value
 * of the line.
 */
int main(int argc, char **argv)
{
    // Declare CSV line type, formatter and template type
    typedef std::tuple<double, double, double>                          CSV_format;
    typedef std::function<CSV_format(const std::vector<std::string> &)> formatterT;

    enum RGB { Red = 1, Green, Blue };

    const std::string COLOR_MAP_PATH = "/some/absolute/path";

    // Load the color map
    auto colorMap = StringHelpers::readCsv<CSV_format>(COLOR_MAP_PATH, ",", [](const std::vector<std::string> &values) {
        return CSV_format {
                // Here is the formatter lambda that convert each value from string to what you want
                std::strtod(values[Red].c_str(), nullptr),
                std::strtod(values[Green].c_str(), nullptr),
                std::strtod(values[Blue].c_str(), nullptr)
        };
    });

    // Use your colorMap as you  wish...
}

A minor edition to @sastanin's solution, so that it can deal with newlines within quotes.

std::vector<std::vector<std::string>> readCSV(std::istream &in) {
    std::vector<std::vector<std::string>> table;

    while (!in.eof()) {
        CSVState state = CSVState::UnquotedField;
        std::vector<std::string> fields {""};
        size_t i = 0; // index of the current field
        for (char c : row) {
            switch (state) {
                case CSVState::UnquotedField:
                    switch (c) {
                        case ',': // end of field
                                  fields.push_back(""); i++;
                                  break;
                        case '"': state = CSVState::QuotedField;
                                  break;
                        default:  fields[i].push_back(c);
                                  break; }
                    break;
                case CSVState::QuotedField:
                    switch (c) {
                        case '"': state = CSVState::QuotedQuote;
                                  break;
                        default:  fields[i].push_back(c);
                                  break; }
                    break;
                case CSVState::QuotedQuote:
                    switch (c) {
                        case ',': // , after closing quote
                                  fields.push_back(""); i++;
                                  state = CSVState::UnquotedField;
                                  break;
                        case '"': // "" -> "
                                  fields[i].push_back('"');
                                  state = CSVState::QuotedField;
                                  break;
                        case '\n': // newline
                                  table.push_back(fields);
                                  state = CSVState::UnquotedField;
                                  fields = vector<string>{""};
                                  i = 0;
                        default:  // end of quote
                                  state = CSVState::UnquotedField;
                                  break; }
                    break;
            }
        }
    }
    return table;
}

CSV file are text files consist of lines, each line is consist of tokens sererated by comma. while there are something you should know when parsing:

(0) The file is encoded with "CP_ACP" code page. you should use the same encoding page to decode the file contents.

(1) the CSV lost "composite cells" information (likes rowspan > 1) , so when it is read back to excel, the composite cell information is lost.

(2) the cell text can be quoted by """ at head and tail, and literal quote char will become double quotes. so the closing matching quote char must be a quote char not followed by another quote char. for eg, if a cell has a comma, it must be quoted in csv, because comma make sense in csv.

(3) when the cell content have multiple lines, it will be quoted in CSV, in this case, you parser must keep reading next sereral lines in CSV file, until you got a closing quote char matching the first quote char, make sure the current logical line is read complete before parsing the line's tokens.

for eg: in csv file, the following 3 physical lines is one logical line consist of 3 tokens:

    --+----------
    1 |a,"b-first part
    2 |b-second part
    3 |b-third part",c
    --+----------
Related