Purpose of declaring an struct variable inside union - c

Question

I have a code like this
typedef struct
{
unsigned char arr[15]; //size = 15bytes
unsigned char str_cks; //size = 1byte
}iamstruct; //Total Size = 16bytes
typedef union
{
iamstruct var;
unsigned char union_cks[16];
}iamunion; //Total Size = 16bytes
static iamunion var[2];
int main()
{
printf("The size of struct is %d\n",sizeof(iamstruct)); //Output = 16
printf("The size of union is %d\n",sizeof(iamunion)); //Output = 16
var[1].union_cks[1] = 2;
printf("%d",var[1].union_cks[1] ); // Output =2
return 0;
}
I'm confused with struct variable declaration inside the union and how it works?.
What is the main purpose of doing this & How it improves accessibility?
Please share your ideas.
Thanks in advance.
I understood something now from the below code. Here memory allocated is 16bytes and its all shared by an individual member of union.
typedef struct
{
unsigned char str_cks1;
unsigned char str_cks2;
unsigned char str_cks3;
unsigned char str_cks4;
unsigned char str_cks5;
unsigned char str_cks6;
unsigned char str_cks7;
}iamstruct;
typedef union
{
iamstruct var;
unsigned char union_cks[7];
}iamunion;
static iamunion var[7];
int main()
{
int i = 0;
printf("The size of struct is %d\n",sizeof(iamstruct));
for(i=0;i<7;i++)
{
var[i].var.str_cks1 = (i*1);
var[i].var.str_cks2 = (i*2);
var[i].var.str_cks3 = (i*3);
var[i].var.str_cks4 = (i*4);
var[i].var.str_cks5 = (i*5);
var[i].var.str_cks6 = (i*6);
var[i].var.str_cks7 = (i*7);
}
for(i=0;i<7;i++)
{
printf("%d\t",var[i].var.str_cks1);
printf("%d\t",var[i].var.str_cks2);
printf("%d\t",var[i].var.str_cks3);
printf("%d\t",var[i].var.str_cks4);
printf("%d\t",var[i].var.str_cks5);
printf("%d\t",var[i].var.str_cks6);
printf("%d\t",var[i].var.str_cks7);
printf("\n");
}
return 0;
}
Output:
enter image description here

Answer 1

A struct represents the values corresponding to the cartesian product o the types of all its fields. These values are the ordered concatenation of the field values and all are present in every single struct value. In this sense, you'll see all the values in the fields as a tuple, or sequence, of values, each of the type of the field they represent.
On the contrary, a union represents an alternative of every field inside, so the whole set of values of the type is the plain union of each of the types inside it.
So, composition (of union and struct) ensures that you can set an ordered sequence of values(struct) or see the struct field as a single alternative to the union. Easy, right! :) (you can also have a union as a field of a struct, meaning this time that the field in the sequence is an alternative of possible sets of values.
Let's see it with an example. Let's assume you have a variable which is supposed to store Real or Complex values. For Real you just use a plain float value (I will over complex this on purpose, to see how it expands) and for Complex we'll use it two double values (this selection has nothing to do with the other alternative and the lose of precision of a float against a double) You can use:
struct complex {
double real_part, imaginary_part;
};
and then
union {
float real_number;
struct complex complex_number;
} my_variable;
then you can access my_variable.real_number as the single precision float value, and my_variable.complex_number.real_part and my_variable.complex_number.imaginary_part as the double precision double real part and imaginary part of a complex number.
Beware that this is not a way to convert values from real to complex or viceversa. Indeed, in this example, both types of values have different representation internally, and you'll mangle your data if you store a single precision float real number on the variable and try to access it as a complex number (you'll have to externally manage the kind of value you have stored in the variable in order to know how to access it) The set of values storable in the variable will be the whole set of float values for real numbers, plus(or also) the whole set of double pairs or real parts and imaginary parts that conform the complex numbers. This is where the union reserved word was taken from.
It is important to consider that a type represents the set of values storable in a variable of that type. In this way, a struct allows you to store a value of each of the types that the fields represent, and you can store all of them at the same time on the variable, while a union only allows you to decide which type (and which field) you'll use to store only a single value of any of those field alternatives, and no more than one.

Answer 2

In the C programming language described in the second edition of K&R's book, a struct or union is a sequence of bytes, and a struct or union member is a means of interpreting some of the storage within the struct or union as that type. Within a struct, all members are assigned to disjoint regions of storage, while all union members use the same storage.
If the members of the structures are stored consecutively, then given:
struct s1 {char a[3],b[5]; };
struct s2 {char a[5],b[3]; };
union { struct s1 v1; struct s2 v2; } u;
the storage assigned to u.v1.b[3] would also be assigned to u.v2.b[1], so a write to either would effectively set the value in both.
The C Standard allows for other dialects of C which would impose additional restrictions on when objects may be read or written, either in cases where it would allow implementations to generate more efficient code or otherwise benefit their customers, or in cases where it would hurt an implementation's customers but the implementer doesn't care. There has never been any consensus about exactly what additional restrictions should be expected, and because the authors of the C Standard assume that implementers will seek to benefit their customers, there was never any real effort to formulate rules that would not rely upon such benevolence.