Thread: Terminator in integer Array.

Hello

I have to create program in which user has to choose how many numbers he would like to enter for example 3. then program ask about these numbers.(Array needs to have space=100) After that program has to add the terminator with value (-1) on the next free space in the array. (that is my main problem).

For example, if the user enters that they have 5 values, the system will then ask them5 times to enter a number to be placed in the arrayOnce the user has finished entering their values, the system should add a -1 in to thenext free element of the array. This will act as a 'terminator'.

in the same program system should ask the user what element of the arraythey want to replace. It should then ask what number they want to put in theirchosen array element.The program should then replace the users chosen array element with the value theyentered.

Main problem which i have i do not know how to use terminator, when i cant declare variables like int /code a[5]={1,2,3,4\-1};, like a[100]; and user has to enter the values. /code

Code:

#include <stdio.h>
#include <ctype.h>
void As5GradeB()
{

  int a[100];
  char opt;
  int i;
  float mean;
  int sum = 0;
  int num;

  opt = tolower(opt);
  printf("Enter number of elements \n");
  scanf("%d", &num);
  for (i = 0; i < num; i = i + 1) {
    printf("Enter value of the elements%d\n", i + 1);
    scanf("%d", &a);
  }


  while ((opt = getchar()) != '\n' && opt != EOF);

  while (opt != 'x') {


    printf("\n Please enter your choice from the following menu:\n");
    printf("A – Repopulate array \n");
    printf("B - Display all values \n");
    printf("C – Replace one number \n");
    printf("D – Calculate the mean \n");
    printf("E - Find largest number\n");
    printf("X - Exit the program \n");
    scanf("%c", &opt);

    opt = tolower(opt);
    if (opt == 'a') {
      for (i = 0; i < num; i = i + 1) { // reading 5 values from user and storing it in array
        printf(" Enter the value number %d\n\n", i + 1);
        scanf("%d", &a);
      }
      printf("The new array is : \n");
      for (i = 0; i < num; i++) {
        printf("%d\n", a);
      }
    } else if (opt == 'b') {
      printf("You choose to display all values \n\n");
      for (i = 0; i < num; i++) {
        printf("%d\n", a);
      }
    } else if (opt == 'c') {
      int ele;
      int val;
      printf("You choose to replace one number\n");
      printf("Which element of the array You want to replace ? \n");
      scanf("%d", &ele);
      printf("What is the replaced value ?");
      scanf("\n%d", &val);
      for (i = 0; i < 100; i++) {
        if (a == ele) {
          a = val;
        }
      }
      printf("%d", a);
    } else if (opt == 'd') {
      sum = 0;
      for (i = 0; i < num; ++i) {
        sum += a;
      }
      mean = sum / num;
      printf
          ("Mean = sum of  the elements devided on the number of elements\n = %.2f\n",
           mean);
    }

    else if (opt == 'x') {
      printf("Thank you for using the program");
      break;
    } else if (opt == 'e') {
      printf("You choosed to find the largest element  \n\n");
      // storing the largest number to arr[0]
      for (i = 0; i < num; ++i) {
        if (a[0] < a)
          a[0] = a;
      }
      printf("Largest element = %d\n", a[0]);
    } else {
      printf("Wrong menu option- try again\n");
    }
    while ((opt = getchar()) != '\n' && opt != EOF);


  }
}

> scanf("%d", &a);
All your references to your array should have a subscript.
As in
scanf("%d", &a[i]);

No, there are 10 different places where you make the same mistake.

Code:

$ gcc -Wall -Wextra -c main.c
main.c: In function ‘As5GradeB’:
main.c:18:11: warning: format ‘%d’ expects argument of type ‘int *’, but argument 2 has type ‘int (*)[100]’ [-Wformat=]
     scanf("%d", &a);
           ^
main.c:40:15: warning: format ‘%d’ expects argument of type ‘int *’, but argument 2 has type ‘int (*)[100]’ [-Wformat=]
         scanf("%d", &a);
               ^
main.c:44:16: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘int *’ [-Wformat=]
         printf("%d\n", a);
                ^
main.c:49:16: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘int *’ [-Wformat=]
         printf("%d\n", a);
                ^
main.c:60:15: warning: comparison between pointer and integer
         if (a == ele) {
               ^
main.c:61:13: error: assignment to expression with array type
           a = val;
             ^
main.c:64:14: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘int *’ [-Wformat=]
       printf("%d", a);
              ^
main.c:68:13: warning: assignment makes integer from pointer without a cast [-Wint-conversion]
         sum += a;
             ^
main.c:83:18: warning: comparison between pointer and integer
         if (a[0] < a)
                  ^
main.c:84:16: warning: assignment makes integer from pointer without a cast [-Wint-conversion]
           a[0] = a;
                ^

I guess it would be something like this

Code:

    if (opt == 'a') {
      for (i = 0; i < num; i = i + 1) { // reading 5 values from user and storing it in array
        printf(" Enter the value number %d\n\n", i + 1);
        scanf("%d", &a[i]);
      }
      a[i] = -1;  // I'll be back
    } else if (opt == 'b') {
      printf("You choose to display all values \n\n");
      // pretend we don't know what num is any more, but there is a terminator
      for (i = 0; a[i] != -1 ; i++) {
        printf("%d\n", a[i]);
      }

Originally Posted by Yonut

Does it have to be a terminator? I am experiencing the same problem with sending an array of structs. I ignore the first struct and start filling the array from array[1]. Then when done I put the total items of the array in array[0]. Then I don't have to check for a terminator, I have the number of items for a loop limit, and I can do this:

Code:

new_array = &array[1];

Which allows me to use the array as normal. Just an idea.

That's not a good idea: it means that you need a member in the struct object just for storing the number of items, and this member is uselessly repeated for all the struct objects even though most of them don't need it. Either that, or you reuse a suitable member, but that's code obfuscation.

There are at least 4 better approaches:

• The use of a sentinel value (terminator), which of course only makes sense if you can designate such a sentinel value (typically a pointer or some zero-initialised value; in the original poster's case it is a negative value because only the non-negative values are used as per normal)
• Pass/keep track of the number of items in a separate variable
• Pass/keep track of a pointer to one past the end of the array of items in a separate variable
• Declare a struct to wrap the array of items and the number of items

Originally Posted by Yonut

Are you thinking that there is a special field for this data? In the struct, there is an 'int' used for 'mode'. The first struct is informational an does not have a mode, so I use it to store to total of the items in the array. It is a field normally used for the remaining data.

That's what I mean by "Either that, or you reuse a suitable member, but that's code obfuscation."

Basically, instead of properly designing a way to convey the number of items, you're obfuscating your code by relying on implicit knowledge that a particular member that normally serves an entirely different purpose is now the number of items for that particular array element. Furthermore, this only works if there is such a member to begin with, otherwise the obfuscation is even worse because you'll have to reinterpret certain bytes of the first array element as an integer.

EDIT:

Originally Posted by Yonut

I did have it setup that way before (using separate variables) I started using a function pointer in my struct. There are 4 functions that can be executed using the function pointers, and for the most part, sending one struct as an argument instead of a struct and a title and an amount, helps the other functions that don't care about the title and amount.

For instance, I have the main function that accepts the struct array. In the first struct is the title for what it is displaying and the amount of items that follow in the array. Then I can just send a struct as an argument. Now there is a function that requires the struct but not the information in the first element. It doesn't care about it. But if they were separate variables, I would need to include the following (to stop compiler from complaining):

So what you should have done is something like this:

Code:

struct X
{
    // ...
};

struct Container
{
    char title[TITLE_SIZE];
    struct X xs[XS_SIZE];
    size_t xs_count;
};

// ...
struct Container container;

Now, you can pass a pointer to struct Container where you need it, or directly pass container.xs to functions that only need the array (but then you will probably also pass container.xs_count as if you need the array, you probably also need the count of the number of elements in use for the array).

Having said that, I don't want to discourage you from using your idea in other contexts where it is not necessarily obfuscation. For example, if you were trying to write your own version of malloc, storing the number of bytes allocated just before the bytes that are actually allocated could be a reasonable strategy to allow for deallocation.

Anyway, we have gone off topic from danlopek14q's discussion, so if you want to discuss this further, let's start a new topic.