Finding anagrams in a word list

I have a word list and a file containing a number of anagrams. These anagrams are words found in the word list. I need to develop an algorithm to find the matching words and produce them in an output file. The code I have developed so far has only worked for the first two words. In addition, I can't get the code to play nice with strings containing numbers anywhere in it. Please tell me how I can fix the code.

Code:

#include <iostream>
#include <fstream>
#include <string>
using namespace std;

int main (void)
{
    int x = 0, y = 0;
    int a = 0, b = 0;
    int emptyx, emptyy;
    int match = 0;
    ifstream f1, f2;
    ofstream f3;
    string line, line1[1500], line2[50];
    size_t found;

    f1.open ("wordlist.txt", ios::in);
    f2.open ("file.txt", ios::in);
    f3.open ("output.txt", ios::out);

    //stores content into string arrays
    if (f1.is_open() && f2.is_open())
    {
        while (f1.eof() == 0)
        {
            getline (f1, line);
            line1[x] = line;
            x++;
        }

        while (f2.eof() == 0)
        {
            getline (f2, line);
            line2[y] = line;
            y++;
        }

        //finds position of last elements
        emptyx = x-1;
        emptyy = y-1;

        //matching algorithm
        for (y = 0; y <= emptyy; y++)
        {
            for (x = 0; x <= emptyx; x++)
            {
                if (line2[y].length() == line1[x].length())
                {
                    for (a = 0; a < line1[x].length(); a++)
                    {
                        found = line2[y].find(line1[x][a]);
                        if (found != string::npos)
                        {
                            match++;
                            line2[y].replace(found, 1, 1, '.');

                            if (match == line1[x].length())
                            {
                                f3 << line1[x] << ", ";
                                match = 0;
                            }
                        }
                    }
                }
            }
        }

        f1.close();
        f2.close();
        f3.close();
    }

    //file access error
    else
        cout << "Input error.";

    return 0;
}

EDIT: Oops, I left the x and y uninitialized, causing the code to crash

> line2[y].replace(found, 1, 1, '.');
The problem with this, is that if line2[y] does not end up being a match, you've trashed it for matching against something later on in the list.

Also, see the FAQ on why using feof() in a loop is bad.

while ( getline (f2, line) )
should be used.

Thanks! I implemented a different method, which involves creating a set of new lists with all the letters and words sorted alphabetically, which works.

However, I still can't figure out how to store strings which contain number(s) in them, so I decided to cheat my way out by removing them from the list. What is a good solution for this?

You would have to post your code, and attach some short example data files which show the problem.
There is nothing special about letters or digits in your original code (at least).

You're right! I tried again and apparently it works as it should with the new algorithm. It didn't work with the above which was why I removed them. Thanks for the help!