help needed with text analyzis

[vlaskiz]

Got a program code so far to:

#include <string>
#include <fstream>
#include <iomanip>
#include <iostream>

using namespace std;
//----------------------------------------------------------------------------
const char Cdat[] = "Data.txt";
const char Crez[] = "Rezult.txt";
const char Canalyzis[] = "Analyzis.txt";

//----------------------------------------------------------------------------
void TrimmText(const char dfv[], const char rfv[], const char afv[]);
void AnalyzeLine(string &lin, string & TwoNumeral, int & TwoStart,
                      unsigned int & TwoSum);
void EditLine(string &lin, int TwoStart, int TwoSum);

int main()
{
    TrimmText(Cdat, Crez, Canalyzis);
}
//----------------------------------------------------------------------------
// Text reading by line
// dfv - original text data file name, rfv - edited text data file name
// afv - analyzis data file name
void TrimmText(const char dfv[], const char rfv[], const char afv[])
{
    string TwoNumeral;
    int TwoStart;
    unsigned int TwoSum;
    ifstream fd(dfv);
    ofstream fr(rfv);
    ofstream fa(afv);
    string L;
    fa << "---------------------------\n";
    fa << "| Word | Start |  Sum |\n";
    fa << "---------------------------\n";
    while(!fd.eof()) 
    {
        getline (fd, L);
        {
            AnalyzeLine(L, TwoNumeral, TwoStart, TwoSum);
            fr << L << endl;
            if (TwoSum <= 9 && TwoSum != 0) 
            {
                fa << "| " << left << setw(5) << TwoNumeral << " | "
                    << right << setw(7) << TwoStart;
                fa << " | " << setw(5) << TwoSum << " |\n";
            }
        }
    }
    fa << "---------------------------\n" << endl;
    fd.close();
    fd.close();
    fa.close();
}
//----------------------------------------------------------------------------
// Function which finds words that are made of two numerals and its sum
// lin - line, in which search is being made
// TwoNumeral - word which is made of two numerals
// TwoStart - return its start
// TwoSum - returns its numerals sum
void AnalyzeLine(string &lin, string & TwoNumeral, int & TwoStart, unsigned int & TwoSum)
{
    string Skirt = " .,!?:;()\t";
    string Numbers = "0123456789";
    string Word;
    int zpr = 0, zpb = 0;
    int skaiciukai[2];
    char skirstymas[2];
    TwoNumeral = "";
    TwoStart = 0; TwoSum = 0;
    while ((zpr = lin.find_first_not_of(Skirt, zpb)) != string::npos) 
    {
        zpb = lin.find_first_of(Skirt, zpr);
        Word = lin.substr(zpr, zpb - zpr);
        if(Word.length()==2 && !Word.find_first_of(Numbers,0))
        {
            TwoNumeral = Word;
            TwoStart = zpr;
            skirstymas[0]=Word[0];
            skirstymas[1]=Word[1];
            char sk1=skirstymas[0];
            char sk2=skirstymas[1];
            skaiciukai[0]=atoi(&sk1);
            skaiciukai[1]=atoi(&sk2);
            TwoSum = skaiciukai[0]+skaiciukai[1];
            if (TwoSum <= 9 && TwoSum != 0)
            {
                EditLine(lin, TwoStart, TwoSum);
            }
        }
    }
}
//----------------------------------------------------------------------------
// function which deletes words that are made of two numerals and which sum is less or equal to 9
// lin - line, which is being changed
// TwoStart - the start of word made of two numerals

void EditLine(string & lin, int TwoStart, int TwoSum)
{
    string Numbers = "0123456789";
    if (Numbers.find_first_of(lin[TwoStart]) != string::npos) 
        {
            lin.erase(TwoStart, 2);
        }
}

I need it to delete words that are made of two numerals and which sum is less or equal to 9. (basicaly if its 18/15/71 it deletes that word, if its 91 it wont because its more than 9). Think im pretty much done getting program to do so, theres a problem - If there's more than 1 word that meets my criteria in single line, program returns only last word from that line to analyzis.txt while i need it to return every single one of it.. Any ideas?

Heres data file example:

Ne vieno antikos 18 autorių veikaluose 17 esama likę 567 nupasakojimų,
 kaip kvapiosios substancijos keliavo 1900 mylių aplink Viduržemio jūrą.
 
        ****  14  12  13 99 125 ***
 
  Iš šių 105 įvairių aprašymų matyti, jog Egiptas, ypač faraonų laikais,
  
  
   05 labai daug *** 76 kvapiųjų substancijų (kai kurių nūdienos  10 archeologų

 skaičiavimais gal net  **** 80 %) 01 atsigabendavo iš tolimų ir ne tokių  tolimų kraštų: iš Arabijos, iš Viduržemio jūros 
 55 salų – Kipro, Chijo ir Kretos, iš Nilo aukštupio. 
 Ypač 333 kvepalų importas 22 Egipte suklestėjo XIII a. pr. m. e, kai šalį valdė Ramzis Didysis.

Rezults file:

Ne vieno antikos  autorių veikaluose  esama likę 567 nupasakojimų,
 kaip kvapiosios substancijos keliavo 1900 mylių aplink Viduržemio jūrą.
 
        ****       99 125 ***
 
  Iš šių 105 įvairių aprašymų matyti, jog Egiptas, ypač faraonų laikais,
  
  
    labai daug *** 76 kvapiųjų substancijų (kai kurių nūdienos   archeologų

 skaičiavimais gal net  ****  %)  atsigabendavo iš tolimų ir ne tokių  tolimų kraštų: iš Arabijos, iš Viduržemio jūros 
 55 salų – Kipro, Chijo ir Kretos, iš Nilo aukštupio. 
 Ypač 333 kvepalų importas  Egipte suklestėjo XIII a. pr. m. e, kai šalį valdė Ramzis Didysis.

Analyzis file:

---------------------------
| Word | Start |  Sum |
---------------------------
| 17    |      37 |     8 |
| 25    |      24 |     7 |
| 10    |      64 |     1 |
| 01    |      33 |     1 |
| 22    |      27 |     4 |
---------------------------

also if i put 01 in front of 125 in data file, I get **** 99 1 *** in rezults file which shouldnt happen.

[Salem]

            skirstymas[0]=Word[0];
            skirstymas[1]=Word[1];
            char sk1=skirstymas[0];
            char sk2=skirstymas[1];
            skaiciukai[0]=atoi(&sk1);
            skaiciukai[1]=atoi(&sk2);

Well atoi() is going to break on those char pointers, because they're NOT pointing at strings with a \0 at the end.

    while(!fd.eof())
    {
        getline (fd, L);

See the FAQ on why using feof() to control a loop is bad.
Then use
while ( getline (fd, L) )

[vlaskiz]

Alright then, what am I supposed to change/do so program returns all the words it deleted to analyzis file? Since im getting only last one of single row right now, and is there any alternative for atoi?

[Salem]

Well atoi() might be OK, if you wrote your code properly so it was always referring to a string with a \0 at the end.

[Elysia]

I am unsure as to why you feel you have to assign individual characters:

skirstymas[0]=Word[0];
skirstymas[1]=Word[1];

...and not whole strings.
As to atoi, you can use atoi(Word.c_str()), or even better, the use of string streams:

std::stringstream str(Word);
str >> myint;

Or even better, boost's lexical_cast:

myint = boost::lexical_cast<int>(mystr);