Thread: File management, help please!

Hi, Im dealing with file management now, and I need some help here. Well what I need is to get the first integer from a file and assign it to a variable and the others integers to an array. Example

Thats my file content 5 4 6 7 8 0 and that would be the code:

Code:

struct Array{
    int n;
    int *v;
};
struct Array ArrayCreaFichero(FILE * f){


struct Array array1;
f = fopen("array.txt","r");
array1.n = 5 (the first value taken from the file)
array1.v = [4,6,7,8,0](the rest of values taken from the file)

The good way to read data from a file, into a variable or an array, depends on the way the data is arranged. All the details are important.

Post an example of the file your program is to read data from, (just a half dozen lines should be enough), and show what you've done to try and make it work. Then we know we're on the same page.

Originally Posted by Adak

The good way to read data from a file, into a variable or an array, depends on the way the data is arranged. All the details are important.

Post an example of the file your program is to read data from, (just a half dozen lines should be enough), and show what you've done to try and make it work. Then we know we're on the same page.

The file would be a .txt and it would be like this:

37 4 5 6 3 2 1 6 9 0 1 3 4 5 6 8 1 8
3 4 5 1 3 8 2 4 2 2 8 1 8 3 4 8 2 8

The first value would be the size of the array. What did I do? just what I posted, nothing more because I have no idea what function I have to use to get one value (the first one here)

Logic like this worked OK.

Code:

   if(!fp) {
      printf("Error opening data.txt file!\n");
      return 1;
   }
   i=fscanf(fp, "%d", &numdigits);
   if(i==1) {
      digits = malloc(numdigits * sizeof(int));
      if(!digits) {
         printf("Error allocating memory!\n");
         return 1;
      }
      for(i=0;i<numdigits;i++) {
         fscanf(fp, "%d", &digits[i]);
         printf("%d ",digits[i]);
      }

   }

Add in the other parts, and you should be off and running.

Adak: Perhaps it might be better to abort in all error cases, just like the code does if the file cannot be opened? And maybe add some comments, too?

Say,

Code:

    /* Verify the file was opened. */
    if (!fp) {
        printf("Error opening data.txt file!\n");
        return 1;
    }

    /* Read the number count from the file. */
    if (fscanf(fp, "%d", &numdigits) != 1) {
        printf("Cannot read the number count from data.txt!\n");
        return 1;
    }
    if (numdigits < 1) {
        printf("The number count in data.txt was too small!\n");
        return 1;
    }

    /* Allocate memory for numdigits integer numbers. */
    digits = malloc(numdigits * sizeof *digits);
    if (!digits) {
        printf("Out of memory (asked for %d numbers).\n");
        return 1;
    }

    /* Read numdigits integers into the allocated array. */
    for (i = 0; i < numdigits; i++) {
        if (fscanf(fp, "%d", &digits[i]) != 1) {
            printf("Error reading number %d of %d from data.txt!\n", i + 1, numdigits);
            return 1;
        }

        /* Print the read number to standard output. */
        printf("%d\n", digits[i]);
    }

Or is this actually going too far? It is almost all the necessary code..

Originally Posted by grumpy

Your line 20 exhibits undefined behaviour, Nominal.

Good catch, thanks! Obviously, that should have been

Code:

    printf("Out of memory (asked for %d numbers).\n", numdigits);

but the stuff after the comma was lost somewhere. And I'm too late to edit it now. Ouch.

Originally Posted by grumpy

It's usually a good idea to write error messages to stderr, too.

Fully agreed. For the file open case, I'd personally recommend

Code:

if (!fp) {
    fprintf(stderr, "Cannot open data.txt: %s.\n", strerror(errno));
    return 1;
}

because fopen() sets errno based on the reason for the failure when it fails. (That requires two more additional includes, <string.h> and <errno.h>.) This way, if the open fails, the user sees also why it failed.

Originally Posted by grumpy

You're structured the code so it tests for failure, whereas I'd assume failure is the default condition, and test for success. That may seem backward, but it can often result in more concise code that needs less comments.

Here, the four points (plus the fifth point inside the loop) are basically fatal, there not being any obvious way to deal with the error.

I thought it would be easier for a new programmer to understand the checks, if they are done immediately and in sequence. Also, minimum changes to Adak's example. (As to comments, I know they're not spot on, but I did try to describe the intent, not the actions the code does.)

The way I'd personally develop the exercise further would be to move all this into a function, then adding the necessary error cleanup to avoid leaking memory or file handles in error cases. In that case, your structure is definitely better, as the cleanups are nicely ordered, and nested at the same depth as the corresponding open or allocation. You can add the cleanups to the code I showed, too, but it wouldn't be as concise.