从C ++文件中读取数字

Question

I want to get numbers from a text file by reading it symbol by symbol and output the numbers in another file.

Number - a word which consists only of numeric symbols.

Word - any string of symbols which is separated by dot, comma, brackets, exclamation mark, question mark or line break.

For example from file

in.txt:

masnd kasjd k!1234 564,7
7.,43, mb?? 67hh k4k 87 90.

I want to output

out.txt

1234
564
7
7
43
87
90

I tried to read the whole word to check if the word is a number, and if it is put the 'cursor' in the beginning of the word and output it in the out.txt

I somewhere messed up and all I get in the out.txt file is

out.txt

234
64

7
87
90

And also if I don't put a separating symbol in the end of the in.txt file it stucks in a loop somewhere, and when I close the program and try to open the out.txs file, notepad crushes

This is how far i have made it

#include <iostream>
#include <cctype>
#include <fstream>
using namespace std;
bool isStop(char c);
int main(){
    fstream fin;
    fstream fout;
    streampos p;
    char a,b=' ';
    fin.open("in.txt", ios::in);
    fout.open("out.txt", ios::out);
    fin.get(a);
    while(!fin.eof()){
        if(isdigit(a) && isStop(b)){
            p=fin.tellg();
            while(isdigit(a) && !fin.eof()){
            fin.get(a);
            }
            if(isStop(a) || fin.eof()){
                fin.seekg(p-1, ios::beg);
                fin.get(a);
                while(isdigit(a) && !fin.eof()){
                    fout.put(a);
                    fin.get(a);
                }
                fout.put('\n');
            }
        }
        b=a;
        fin.get(a);
    }
    fin.close();
    fout.close();
    return 0;
}
bool isStop(char c)
{
    return (c==' ' || c=='.' || c==',' || c=='(' || c==')' || c=='!' ||                             c=='?' || c=='\n');
}

Answer 1

I suspect your issue stems from a couple of things, like treating the space character as the only whitespace character (when there could be newlines '\\n' , tabs '\\t' , or carriage returns '\\r' ), and additionally using !fin.eof() as a loop check.

In fact, All the jumping around in the file appears unnecessary and difficult to read.

I propose the following simplified algorithm (much more C++-esque):

• Read a line

• Parse the line into words

• Verify a word is a valid number (according to your rules)

• if valid, write that number to file

Here's one possible implementation (using cin and cout instead of files):

---

Define your separators:

const std::string separators = ",.()! \t";

Read a line:

std::string next;
while(getline(std::cin, next))
{

Break it into words:

    std::string::iterator beg = std::begin(next);
    while(beg != std::end(next))
    {
        beg = std::find_if(beg, std::end(next), [](char c)
        {
            return std::isdigit(c);
        });

        std::string::iterator end = std::find_if(beg, std::end(next), [](char c)
        {
            return separators.find(c) != std::string::npos;
        });

        // now we have digit to separator (or end of string). So long as elements between beg (inclusive) and end(exclusive) are all digits, we're good

Verify the word is all digits from first digit to next separator:

        std::string::iterator it = beg;
        for(; it != end && isdigit(*it); ++it)
        {}

If verification passes, write to file:

        if (it == end) // they're all digits
        {
            std::cout << std::string(beg, end) << std::endl;
        }
        beg = end;
    } // end while loop (word separation)
} // end while loop (reading lines)

Live Demo

Answer 2

A bit more c++ like:

#include <fstream>
#include <string>
#include <algorithm>
#include <iterator>

using namespace std;
// string of separators 
static const string valid = ".,!? \n\0'";   

// check whether a character is in the separator
static bool isValid(char c) { 
  return (find(valid.begin(), valid.end(), c) != valid.end());
}

int main() {
  // open your input/output files
  ifstream inputf("./in.txt");
  ofstream outf("./out.txt");
  // save your input in a string
  string file((istreambuf_iterator<char>(inputf)), (istreambuf_iterator<char>()));
  int beg; // start of a number (index)
  std::string nb;
  // iterate over your string
  for (int i = 0; i < file.size() ;) {
    // while your char is an alpha char, do nothing
    while (isdigit(file[i]) == false) ++i;
    nb = "";
    beg = i;
    // search for a number
    while (isdigit(file[i]) == true)
      nb += file[i++];
    // check if the current number can be saved in the output file
    if (nb != "" && isValid(file[i]) && (beg == 0 || isValid(file[beg - 1])))
      outf << nb << "\n";
  }
  inputf.close(); outf.close();
}

And then your output file:

$> cat out.txt 
1234
564
7
7
43
87
90

Answer 3

I had to check the file symbol by symbol so I couldn't use getline, but in the end I figured it all out.

I saved the word in a string and afterwards checked if it was a number and if it was, i put it in the out file.

Here is the code:

#include <iostream>
#include <fstream>
using namespace std;
bool isStop(char c);
int main(){
    fstream fin;
    fstream fout;
    string s;
    char a,b=' ';
    fin.open("inF20.txt", ios::in);
    fout.open("outF20.txt", ios::out);
    while(fin.get(a)){
        if(isdigit(a) && isStop(b)){
                s+=a;
            while(fin.get(a) && isdigit(a)){
                s+=a;
            }
            if(isStop(a) || fin.eof()){
                for(int i=0;i<s.length();i++){
                    fout.put(s[i]);
                }
                fout.put('\n');
            }
        }
        s.clear();
        b=a;
    }
    fin.close();
    fout.close();
    return 0;
}
bool isStop(char c)
{
    return (c==' ' || c=='.' || c==',' || c=='(' || c==')' || c=='!' || c=='?' || c=='\n');
}