从C ++中的.csv文件读取双打

Question

Hello I'm having a bit of trouble getting my values to read properly from a .csv file. I want to populate a vector with 1000 user defined objects (DetectedParticle), with their variables assigned from the file. The object has 4 member variables (x0,x1,x2,x3) of type double. In the file each line ought to correspond to a single entry in the vector, and each column of the line needs to be read into its respective variable in the object.

The file format is like this (excluding the comments):

wordswordswordswordswords
2.231,23.52,123.4,213.2   //first object to be created with these values
23213,23123,41234,45236  //second
21323,123123,123123,2435  //third
.
.
.
23434,234234,234234,234 //1000th

The problem is it only reads in every other line into the vector. So the vector will have an object in from the first line after the words, and then it'll have the third line but it'll miss out the second line. As a result I have a vector of size 499 rather than size 1000. It is successfully assigning the column to the correct member variable though, it's just not doing it enough times! Here's the code snippet:

std::vector<DetectedParticle>  populate(std::string file){   //DetectedParticle is my user defined type with membervariables x0,x1,x2,x3
std::vector<DetectedParticle> temp;  //creates a vector of Detected Particles for later returning
std::ifstream fin("detectedpositrons.csv");  //ifstream to read from
std::string linestr;     //string for the stream to be read into            

if (!fin.is_open())  {  std::cerr << "failed to open file\n"; } // check file is open, error message if not

fin.ignore ( std::numeric_limits<std::streamsize>::max(), '\n' ); //ignores first 


while ( std::getline(fin, linestr) ){     //
    int i = 0;
    DetectedParticle temppart;

while ( std::getline(fin,linestr,',') )
{
    if (i == 4){temp.push_back(temppart);  break;}   //when it gets to teh end of the line, past the 4th column, it breaks and should go to the next line 
    float holder ;                                               //float for the string to be held in
    std::istringstream(linestr) >> holder;            //converts string to float
    if(i == 0){ temppart.x0 = holder; i++;}         //if first column, sets x0 equal to the value
    else if (i == 1){ temppart.x1 = holder; i++;}  //if second column, sets x1 = to value
    else if (i == 2){ temppart.x2 = holder; i++;}  //if 3rd column sets x2= value
    else if (i == 3){ temppart.x3 = holder; i++;}  //if last column it populates the last value x3 with that value.

}
}
return temp;   //returns the populated vector
}

Sorry if it's not easy to follow or the most intuitive way to read a csv file or the nicest looking method but I came up with it myself so that's to be expected! Thanks in advance!

Answer 1

The problem is you read a line, do nothing with it, and then read the next line to parse:

while ( std::getline(fin, linestr) ) // first read
{
    ...
    while ( std::getline(fin,linestr,',') ) // second read
    {
        // you are doing stuff here
    }
}

I think you want to just discard the header row (based on your description), so you should have a single std::getline call outside of a loop, and then a looped std::getline call to read in the double values:

std::getline(fin, linestr); // header row - throw away
while (std::getline(fin, linestr))
{
    istringstream iss(linestr);
    // parse the stringstream into your vectors - I wouldn't use getline here, but you could
}

Answer 2

You're discarding the line you've got in your outer loop ( while ( std::getline(fin, linestr) ){ ). You're abusing getline to chop input into constituent parts - that's be swell, but you're not using the line you've already read.

If you insist on this hack, you need to create a temporary string stream that takes the line you've read in the outer loop, and then run getline in the inner loop on that temporary string stream , not on fin .

Answer 3

Use Boost Spirit .

Here is an example you can adapt. It does parse the input line looking for 4 comma separated double values and populate a std::vector of std::tuple<double, double, double, double> :

#include <boost/spirit/include/qi.hpp>
#include <boost/spirit/home/phoenix/object/construct.hpp>
#include <boost/spirit/home/phoenix/container.hpp>
#include <boost/spirit/include/phoenix_core.hpp>
#include <iostream>
#include <string>
#include <tuple>

typedef std::tuple<double, double, double, double> particle;
typedef std::vector<particle> Particles;

template <typename Iterator>
bool populate_vector(Iterator first, Iterator last, Particles& vector)
{
    namespace qi = boost::spirit::qi;
    namespace ascii = boost::spirit::ascii;
    namespace phoenix = boost::phoenix;

    bool r = qi::phrase_parse(first, last,
        (
            (qi::double_ >> ',' >> qi::double_>> ',' >> qi::double_ >> ',' >> qi::double_)
            [ 
                phoenix::push_back(phoenix::ref(vector),
                    phoenix::construct<particle>(qi::_1, qi::_2 , qi::_3, qi::_4))
            ]
        ), qi::space);

    return r;
}

int main()
{
    std::string str;
    Particles particles;

    while (getline(std::cin, str))
    {
        if (str.empty() || str[0] == 'q' || str[0] == 'Q')
        {
            break;
        }

        if (populate_vector(str.begin(), str.end(), particles))
        {
            std::cout << "Parsing succeeded: " << particles.size() << std::endl;
        }
        else
        {
            std::cout << "Parsing failed." << std::endl;
        }
    }

    return 0;
}

Interaction example:

1.26,1.23,1.6,152
Parsing succeeded: 1
1,2,3,4
Parsing succeeded: 2