Thread: Writing to a file checking if the input is new

This is a small function that takes a string (a name, an address or a mail) like "Mark" and writes this string in a file IFF there isn't a previous entry for "Mark". I want to write an entry a row.

The problem is that it seems that fprintf writes also some trash and the function doesn't really work because it tryes to check the presence of "mark" also in the trash that blows up everything.. the code is commented!

Any idea?

Code:

FILE* database;

int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, char * lpCmdLine, int  nCmdShow) { 
  
    database = fopen("database.txt", "a");
    writeDatabase("Mark");
    writeDatabase("Mark");
    writeDatabase("Bob");
    fclose(database);
  
  return 0; 
}

Code:

int writeDatabase(char* mail) {
    char token[MAX_PATH]= "\0";
    char tokentemp[MAX_PATH]= "\0";
    
    printf("string length %d\n", strlen(mail));
    // if the file is empty
    if ((fgets(token, MAX_PATH, database)) == NULL) {
         printf("empty\n");
         // then I will write the first string in the file
         fprintf(database, "%s\n", mail);
         printf("Written at the beginning -> %s\n", mail);
         system("PAUSE");
    } else {  // if the file contains some strings
              while(fgets(tokentemp, MAX_PATH, database)!= NULL) {
                   printf("Iterating on string %s\n", tokentemp);
                   // analyzing the single token, looking for a previous
                   // entry of the string
                   if (strstr(tokentemp, mail) != NULL) {
                          // previously written
                          printf("Found");
                          system("PAUSE");
                          return 0;
                   }
              }
              // otherwise the string is new, then write it
              fprintf(database, "%s\n", mail);
              printf("Written -> %s\n", mail);
              system("PAUSE");
      }
    return 0;
}

The output file should contain "mark" and "bob" in 2 different rows.. but it's not the case!!! it is huge, full of trash, like:

Code:

Mark
$=                                                                                                    Mark
$=

it is full of spaces (hundreds!) and strange things... ??

> if ((fgets(token, MAX_PATH, database)) == NULL) {
> fprintf(database, "%s\n", mail);
There is only one 'file pointer' per file, which means need to be a lot more careful when reading and writing the same file. This typically means you use lots of fseek() and fflush() calls to make sure the file pointer is in the right place, and the file is up to date.

Also, you normally use this for binary files with fixed-length records, not text files with variable length records.

I've corrected some bugs in my code and it quite works..
Once I call the function, the file is opened in read mode.. if nothign can be read, I close it, open it again in append mode and write the string. If something is readable, I check if some of the previous entry was the same of the one I have here, if not, I write it, otherwise I trash the string.

The problem is that It seems to fail at the second cycle because it doesn't check the first written string.. check out the code:

Code:

int writeDatabase(char* mail) {
    char startToken[MAX_PATH]= "\0";
    char tokentemp[MAX_PATH]= "\0";
    
    database = fopen("database.txt", "rt");
    printf("Want to write %s\n", mail);
    // if the file is empty
    if ((fgets(startToken, MAX_PATH, database)) == NULL) {
         //printf("empty\n");
         fclose(database); // closing read only
         // then I will write the first string in the file
         database = fopen("database.txt", "at"); // opening append
         fputs(mail,database);
         fputs("\n",database);
         fclose(database); // closing append
         printf("Written at the beginning -> %s\n", mail);
    } else {  // if the file contains some strings
              while(fgets(tokentemp, MAX_PATH, database)!= NULL) {
                   //printf("Iterating on string %s \n", tokentemp);
                   // analyzing the single token, looking for a previous
                   // entry of the mail
                   if (strstr(tokentemp, mail) != NULL) {
                          // previously written
                          printf("Found\n");
                          fclose(database); // closing read only
                          return 0;
                   }
              }
              // otherwise the mail is new, then write it
              fclose(database); // closing read only
              database = fopen("database.txt", "at"); // opening append
              fputs(mail,database);
              fputs("\n",database);
              printf("Written -> %s\n", mail);
              fclose(database); // closing append
      }
    return 0;
}

Code:

FILE* database;
int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, char* lpCmdLine, int nCmdShow) { 
  
    // creating the file
    database = fopen("database.txt", "w");
    fclose(database);
 
    writeDatabase("marco");
    writeDatabase("marco");
    writeDatabase("ciao");
    writeDatabase("ciao");
    writeDatabase("marco");
    writeDatabase("marco");
    writeDatabase("ciao");
    writeDatabase("miao");
    writeDatabase("miao");
    system("PAUSE");
  
  return 0; 
}

the debug (printf) output is :

Want to write marco
Written at the beginning -> marco
Want to write marco
Written -> marco
Want to write ciao
Written -> ciao
Want to write ciao
Found
Want to write marco
Found
Want to write marco
Found
Want to write ciao
Found
Want to write miao
Written -> miao
Want to write miao
Found

As you can seen, the second time I add "marco" it's added without succeding while checking if it was previously entered (it was entered the first time, just before!)

and the file contains:

marco
marco
ciao
miao

as you can see there is a duplicated marco..

any hints? thanx!

If the file isn't empty
> if ((fgets(startToken, MAX_PATH, database)) == NULL) {
This succeeds.

And then you read another line (that's 2) with this
> while(fgets(tokentemp, MAX_PATH, database)!= NULL)

With text files, the best thing to do is
- open current file for input
- open a new file for output
- while ( fgets() != NULL ) to read each line from the input file
- use program logic to decide whether to
= write the line out to the output file
= write something else to the output file
= not write anything to the output file

When you're done, close both files, then delete and rename.