Question

I’m currently busy with a project in which I have to copy a part from a file to another file, so I made a code for that using fread and fwrite. But I came across a problem: for testing purposes I made a code which should copy a whole file, but somehow the code creates copies that are larger than the original file. See the code I made below

             FILE *base_file;
             FILE *new_file;
             fpos_t curpos;
             int tmp;

             // Open the base file
             fopen_s(&base_file, "C:/base.dat", "rb");
             // Open the file which should contain the copy
             fopen_s(&new_file, "C:/new.dat", "w");

             // Get the filesize
             fseek(base_file, 0, SEEK_END);
             fgetpos(base_file, &curpos);
             fseek(base_file, 0, SEEK_SET);

             //Read and copy (it seems to go wrong here)
             for(int i = 0; i < curpos; i++){
                 fread (&tmp, 1, 1, base_file);
                 fwrite(&tmp, 1, 1, new_file);
             }

             fclose(base_file);
             fclose(new_file);

The base file is 525 kb, and the newfile is 527kb. As far as I could see the parts where this problem occurs is after parts where there are 7 nullbytes, and the copy somehow has added a ‘0D'(in Hex) after these parts. In ascii the ‘0D’ character is a ‘carriage return’. I was wondering what could be the reason that my copy code adds carriage returns into the file? As far as I know this script should just work as I just read the basefile, and directly copy it to the newfile, and the basefile doesn’t contain these carriage returns.

Answer 1

You’re opening the destination file in text mode, instead of binary mode, so the newline translation happens behind your back. Change the mode to "wb".

Other notes:

1. Use streams rather than stdio.
2. Don’t write byte-for-byte. Use a larger buffer, your method will take forever for larger files.