The union keyword in C confuses me. From what I have read, unions are used to store different types of data at the same memory address. How can one memory address store different amounts of data. For example,
union Data {
int i;
float f;
char str[20];
};
The memory occupied from this struct is 20 bytes. How can the int, float and char array hold values all at once in the same memory address?
CodePudding user response：
How can unions store mutliple values in one memory address in C?
They cannot and they don't. Only one of the union members is stored at any one time.
Example:
union Data d = {.i = 42}; // d contains only i
printf("%d\n", d.i);
d.f = 3.145f; // d contains only f
printf("%f\n", d.f);
CodePudding user response：
The different members of a union all refer to the same memory address, Now, a variable of a union type can store an integer, a floating-point number, or a string of characters. It means a single variable, i.e., the same memory location, can be used to store multiple types of data.
When a variable is associated with a union, the compiler allocates the memory equals to the size of the largest member which is an array of characters of size 20 here so (20 bytes).
CodePudding user response：
A union only stores one thing at any given time - your union type can store either an int or a float or a string.
Enough space is set aside to store the largest member (in this case, the char array), subject to any alignment restrictions. If you do something like
union Data x;
x.i = 10;
x.f = 20.0;
strcpy( x.str, "vootie" );
then the only thing that's actually stored in x is the string "vootie" - the value of i was overwritten when you wrote to f, and the value of f was overwritten when you wrote to str.
CodePudding user response：
Unions do not store many values at the same location, only one. But in C you can access another union member reading (or writing) the same location. It is called union punning
union
{
unsigned u;
float f;
unsigned char uc[sizeof(float)];
}u;
int main(void)
{
if(sizeof(u.f) == sizeof(u.u))
{
u.f = 1.2345f;
printf("float value %f is representd as: 0x%x or bytes: 0xhhx 0xhhx 0xhhx 0xhhx ",
u.f, u.u, u.uc[0], u.uc[1], u.uc[2], u.uc[3]);
}
}