Getline和EOF

Question

I am trying to read a file. The file contents have a newline between words in a sentence and two newlines between sentence. I can only read one sentence. I have tried to put a EOF as a delimiter in getline but it seems not to work. Does anyone have any suggestions on how to resolve this?

The file contents are:

County

Grand

Jury

said Friday an investigation of Atlanta's recent primary
election produced `` . no evidence '' . that any
irregularities took place . . The jury further said in
term-end presentments that the City Executive Committee

But what's get printed is:

County

Grand Jury said Friday an investigation of Atlanta's
recent primary election produced `` . no evidence '' . that any irregularities took place . .

string line;
string a, b;
ifstream infile("myFile");

 while (getline(infile, line))
{
    istringstream iss(line);

    if (!(iss >> a >> b)) { break; } // error

    cout << a << b << endl;
}

Answer 1

#include <iostream>
#include <vector>
#include <boost/tokenizer.hpp>

using namespace std;

typedef boost::tokenizer<boost::char_separator<char>,
        std::istreambuf_iterator<char> >
    tokenizer;

void printPhrase(const vector<string>& _phrase) {
    if (!_phrase.empty()) {
        vector<string>::const_iterator it = _phrase.begin();
        cout << "Phrase: \"" << *it;
        for(++it; it != _phrase.end(); ++it)
            cout << "\", \"" << *it;
        cout << "\"" << endl;
    } else
       cout << "Empty phrase" << endl;
}

int main() {
    boost::char_separator<char> sep("", "\n", boost::drop_empty_tokens);
    istreambuf_iterator<char> citer(cin);
    istreambuf_iterator<char> eof;
    tokenizer tokens(citer, eof, sep);

    int eolcount = 0;
    vector<string> phrase;
    for (tokenizer::iterator it = tokens.begin(); it != tokens.end(); ++it) {
        if (*it == "\n") {
            eolcount ++;
            if (eolcount > 1 && eolcount % 2 == 0) { // phrase end
                printPhrase(phrase);
                phrase.clear();
            }
        } else {
            eolcount = 0;
            phrase.push_back(*it);
        }
    }
    if (!phrase.empty())
        printPhrase(phrase);
    return 0;
}

The basic idea is to keep newlines in the output, count them and if there're 2, 4, .. even number of sequential newlines print words collected so far. A non-newline token breaks the sequence, and this token is added into the accumulator.