Thread: Reading .txt files with whitespaces within data-fields

I'm using the following code to read .txt files and print to console:

Code:

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


using namespace std;


int main (){


    fstream File;
    string name, address;
    int age;
    vector<tuple<string,string,int>>v;
    File.open("F:\\test.txt");
    if(File.is_open()){
       string line;
       while(getline (File, line)){
         stringstream stream(line);
                 while(stream>>name>>address>>age){
                v.emplace_back(name, address, age);
            }
        }
    }
      for(auto& itr : v){
        cout<<"name: "<<get<0>(itr)<<", address: "<<get<1>(itr)<<", age: "<<get<2>(itr)<<"\n";
      }
}

When my .txt file is:
Name Address Age
John London 45
Jane Cambridge 31
Output:

Code:

name: John, address: London, age: 45
name: Jane, address: Cambridge, age: 31

Note that the first line of the .txt file (Name Address Age) is not printed in this case but if I remove the first line and start the .txt file with actual names, addresses, ages then still the first entry is printed fine. Why this dichotomy?


And when my .txt file is:
Name Address Age
John Smith Abbey Road, London 45
Jane Doe Huntingdon Road, Cambridge 31


The output is essentially nothing:

Code:

Process returned 0 (0x0)   execution time : 0.025 s
Press any key to continue.

I understand that having whitespaces within data-fields is breaking down sstringstream but how could I then best write the program so that it can read strings with whitespaces within them into a single string and then move on to the next variable which could be another string with whitespaces within (as in this example) or some other datatype? It seems that the stringstream ctor does not take a delimiter, otherwise I might have tried ...

Code:

 stringstream stream(line, ';');

... with the semi-colon delimiting the datafields in the .txt file.


Thanks as ever.

@Salem - thanks for the headsup, this has get me going but I haven't reached the final solution yet. My initial effort has been on the following lines:

0. read the file a line at a time and for each line generate the tokens, token, with find(delimiter) as described in the link sent
1. intialize a stringstream object, stream, with token
2. declare an int counter, t, and a temp tuple, MyTuple
3. assign stream to MyTuple upon generation with get<t>(MyTuple) = stream;
3. t++
4. and when the entire line has been read, v.push_back(MyTuple) where v is the vector<tuple<string,string,int>>
5. read next line and repeat above ...

But there's a problem at step 3 because std::get<t>(MyTuple) would only work if t was const. I'm showing the code as it stands below and commented out the line that doesn't work. If anybody can help me out from here on that would be most appreciated. Thanks

Code:

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

int main (){
fstream File;
vector<tuple<string,string,int>>v;
File.open("F:\\test.txt");
    if(File.is_open()){
       string line;
       string delimiter = ";";
       size_t pos = 0;
       string token;
            while(getline (File, line)){
            tuple<string, string, int> MyTuple{};
                while((pos = line.find(delimiter))!=string::npos){
                token = line.substr(0, pos);
                stringstream stream(token);
                int t = 0;
                get<t>(MyTuple) = stream;//DOESN'T WORK!!!!//
                t++;
                line.erase(0, pos + delimiter.length());
                }
                v.push_back(MyTuple);
            }
        }
File.close();

      for(auto& itr : v){
        cout<<"name: "<<get<0>(itr)<<", address: "<<get<1>(itr)<<", age: "<<get<2>(itr)<<"\n";
      }
}

@laserlight : this is brilliant, thank you so much.

Is there a specific term to describe the getline() && getline() ... technique? It seems very handy and I'd like to read up a bit more on this.

PS: FYI, both the following lines:

Code:

 auto stream = std::stringstream(line);
auto File = fstream("F:\\test.txt");

give error messages that the corresponding ctor's are deleted

Is there a specific term to describe the getline() && getline() ... technique?

No, getline() has a delimiter parameter. You can click the link to know more.

Laserlight is using a newer C++11 feature. You can still declare objects the normal way.

Thanks, I later realized that the short-circuit evaluation also works with my original vector<tuple> approach:

Code:

vector<tuple<string,string,int>>v;
getline(stream, name, delimiter) && // reads stream into name upto delimiter;
            getline(stream, address, delimiter) && // reads stream into address upto delimiter;
            stream >> age;
                v.emplace_back(name, address, age);