retaining punctuation in c++

Question

Can anybody help me? In the code i have instead of erasing the punctuation as shown, i need to strip it out to be re-used later on when the word that precedes it has been expanded from its abbreviated form. Any help would be greatly appreciated. N.B This is only one function of the code, to display the vectors and open the infile etc... have been placed in separate functions.

void readText2(ifstream &infile, vector <string> &textFile2 ) 
{
    while( !infile.eof() )
    {
        string fileWord;
        infile >> fileWord;
        for (size_t i=0; i !=fileWord.length(); i++)
        {
            fileWord[i]=toupper(fileWord[i]);
        }

        for (size_t i=0; i < fileWord.length(); i++)
        {
            if (ispunct (fileWord[i]))
            {
                fileWord.erase(i); 
            }
        }
        textFile2.push_back(fileWord);
    }
    infile.close();
}

It is unclear what you are trying to do. Can you give an overall high level description of the problem you are trying to solve? — David Rodríguez - dribeas, Mar 31 '13 at 17:01
i'm reading the infile string by string into string fileWord, but instead of it storing the punctuation at the end of the string i want it to be taken out and stored somewhere else therefore var fileWord will only hold letters NOT punctuation. — maccame, Mar 31 '13 at 17:03
So you dont need the punctuation? What is the problem you're solving? — DavidO, Mar 31 '13 at 17:15
**[`while (!stream.eof())` is wrong](http://stackoverflow.com/q/5605125/560648)** - which resource instructed you to do that? — Lightness Races in Orbit, Mar 31 '13 at 17:20
Also I can't find your question. Not sure why this has been upvoted. — Lightness Races in Orbit, Mar 31 '13 at 17:20
@didierc - that is my exactly my question - how would i do it?! — maccame, Mar 31 '13 at 18:08
@DavidO - the problem is how to store characters in a string, but create a new instance of the string when the infile read hits a punctuation mark. — maccame, Mar 31 '13 at 18:09
sorry, I meant : how are you planning to reuse it later on, ie. why do you need to reuse it later on? — didierc, Mar 31 '13 at 18:59

Erik van Velzen · Accepted Answer · 2013-04-01T23:32:30.080

I'd suggest traversing the string end-to-begin (to preserve the correct index numbers), and inside the loop store the index+punctuation charactar in an std::map<size_t,char>

Your function hacked, with comments where stuff changed:

// changed return type
vector<map<size_t,char>> readText2(ifstream &infile, vector <string> &textFile2 ) 
{
    vector<map<size_t,char>> vPunct;    // added

    while( !infile.eof() )
    {
        string fileWord;
        infile >> fileWord;
        for (size_t i=0; i !=fileWord.length(); i++)
        {
            fileWord[i]=toupper(fileWord[i]);
        }

        map<size_t,char> mPunct;        // construct a map
                                        // map stays empty if no punctuation is found

        for (size_t i=fileWord.length()-1; i>=0 ; i--)    // reversed loop
        {
            if (ispunct (fileWord[i]))
            {
                // store index + punctuation char
                mPunct.insert(pair<size_t,char>(i,fileWord[i]));     
                fileWord.erase(i); 
            }
        }
        textFile2.push_back(fileWord);      
        vPunct.push_back(mPunct);      // add map to vector
    }
    infile.close();
    return vPunct;   // return vector
}

The punctuation is now stored in a second vector, which the function returns. For example: a string at textFile[20] has its punctuation stored in vPunct[20].

It would be more reliable to store the punctuation together with the string in one vector<pair<string,map<size_t,char>>>, but that would alter the external behavior of the function, and I dont know if that is desireable for you.

wow - thankyou so much, that's really really helpful - AND understandable. :) — maccame, Apr 02 '13 at 08:24

retaining punctuation in c++

1 Answers1