Thread: Merging Files

I have a program that splits a large, unsorted, file into two smaller files, sorts the two files, then merges them into "final.txt".

The user is asked to enter a guess as to how many lines exist in the original file. The program works fine except when the user's guess is too small (so small that "file1.txt" contains only 1 or 2 lines of text and "file2.txt" contains the majority). In this case the last line of "file1.txt" is written to "final.txt" twice. Any ideas... I know the problem lies in this function?

Code:

void merge_files(void) {
   char temp1[80], temp2[80];
   int comparison;
   FILE *input1, *input2, *output;
   input1=fopen("file1.txt", "r");
   input2=fopen("file2.txt", "r");
   output=fopen("final.txt", "w");
   fgets(temp1, 80, input1);
   fgets(temp2, 80, input2);
   while (!feof(input1) && !feof(input2)) {
      comparison=strcmp(temp1, temp2);
      if (comparison<0) {
         fprintf(output, "%s", temp1);
         fgets(temp1, 80, input1);
      } else {
         fprintf(output, "%s", temp2);
         fgets(temp2, 80, input2);
      }
   }
   while (!feof(input1)) {
      fprintf(output, "%s", temp1);
      fgets(temp1, 80, input1);
   }
   while (!feof(input2)) {
      fprintf(output, "%s", temp1);
      fgets(temp1, 80, input2);
   }
   fclose(input1);
   fclose(input2);
   fclose(output);
}

Cheers...

Salem mentions this problem almost everyday....

I was talking with Salem about this very code in a thread a while back. It's taken me this long to find out there actually was a problem with it (the case where the user wildly underestimates the number of lines of text in his file).

Dont bother with calls to feof() but instead explicitly check for EOF

Firstly: I don't know how to do that (maybe fgets returns EOF?)

Secondly: How would I change the above loop so that it quits when one of the files returns EOF, but continues reading from the other?

Thanks...

A better way is to remove the feof call altogether, and go with

Code:

int ch;
while ( (ch=fgetc(fp)) != EOF ) {
  // do stuff
}

Thanks... but does fgets also return a value (and does this also have to be an int?)?

ie will this work for strings

Code:

int x1, x2;
while ( x1=fgets(temp1, 80, input1) != EOF && x2=fgets(temp2, 80, input2) != EOF) {}

Ahhhh... that's brilliant... thanks for that...

So, what I have learned today is:

fgetc() returns an integer EOF (-1 as far as I can see)

fgets() returns a NULL character pointer

fgets() doesn't empty the buffer it created after reading the string


At least I hope so...