Difference between union and structure

[anukalp]

I've been searching on internet and trying to understand how union is different then structure in c language

structure is use to hold different type of data type

#include <stdio.h>

struct student{
    char *name;
    int age;
}record;
int main()
{
    
     record.name = "Ceena";
     record.age = 35;

     printf("student name : %s", record.name);
     printf("\nStudent Age: %d", record.age);
    
 return 0;
}

In this link https://www.geeksforgeeks.org/difference-structure-union-c/ there are the six reason I don't understand second reason memory allocation for union and structure

Can anyone help me to understand how union is different then structure in c language ?

 

[(*steve*)]

In a struct the variables all occupy their own place in memory. In a union they all overlap.

 

[anukalp]

In a struct the variables all occupy their own place in memory. In a union they all overlap.

can you explain little bit how they will overlap ?

 

[(*steve*)]

They will occupy the same memory locations.

 

[anukalp]

They will occupy the same memory locations.

Example for structure

struct student{
    int roll_number;
    float mark ;
}record;

A struct will allocate space in memory for each of it's members at same time

0001 for int
0002 for float


Example for union

union student{
    int roll_number;
    float mark ;
}record;

A union will use the same space in memory for each of it's members

A union will allocate space in memory for any one of it's members at one time

first it will allocate memory space 0001 for int and later it will use same memory space 0001 for float

Does it happen ?

 

[(*steve*)]

first it will allocate memory space 0001 for int and later it will use same memory space 0001 for float

Not later, at the same time. All variables are always available.

Your code might use this to save space in certain cases. In my experience it is more often used to overlay structures which may contain different data at different times.

For example a set of 4 byres might contain one long integer or 4 separate bytes depending on context. An old style ip address is often presented this way.

 

[anukalp]

Not later, at the same time. All variables are always available.
.

I didn't understand

consider example

union student{
    int roll_number;
    float mark ;
}record;

memory space
0001
0002
0003
0004
.......

0010

how the union will allocate memory space ?

 

[(*steve*)]

how the union will allocate memory space ?

both variables will start at the same location in memory. The size of the union will be that of the largest member.

 

[Harald Kapp]

Here's a good description of unions in C including an example what can go wrong when a union is handled without the required care.

 

[(*steve*)]

Unfortunately that sure contains the rather misleading information:

"You can define a union with many members, but only one member can contain a value at any given time"

Because the variables occupy the same memory, placing a value in one will generally place a value in all of them. However, there is no guarantee that the other fields will contain meaningful or even valid data.

 

[anukalp]

Unfortunately that sure contains the rather misleading information:

"You can define a union with many members, but only one member can contain a value at any given time"

Because the variables occupy the same memory, placing a value in one will generally place a value in all of them. However, there is no guarantee that the other fields will contain meaningful or even valid data.

I had asked you before, I thought you would explain by giving examples.

I still do not understand, can you explain by giving examples,

 

[(*steve*)]

No. You need to learn yourself.

Implement something yourself and see what happens.

I'd recommend unioning together an integer and a byte.

Put values into one and see what happens to the other.

To understand this you need to understand the size and the layout of data types on your machine. Do you have this basic understanding?

 

[anukalp]

No. You need to learn yourself.

Implement something yourself and see what happens.

look post 7, didn't I try to do that ?

To understand this you need to understand the size and the layout of data types on your machine. Do you have this basic understanding?

sizeof gives the size of a variable

#include<stdio.h>
int main()
{
    printf("%d\n",sizeof(char));
    printf("%d\n",sizeof(int));
    printf("%d\n",sizeof(float));
    printf("%d", sizeof(double));
    return 0;
}

output 1 4 4 and 8

Machine take 1 byte to store character
Machine take 4 byte to store integer value
Machine take 4 byte to store float value
Machine take 8 byte to store big data

 

[(*steve*)]

But what exactly is stored in those bytes?

If you set your integer to -75, what would the character be?

Similarly, if you set the integer to zero, then placed an "A" in the character variable, what value would the integer have?

You need to know this if you're using unions in one particular way.

If all you want to do is save space, and you know that you can't use the variables at the same time, then this information isn't necessary.

 

[anukalp]

But what exactly is stored in those bytes?

If you set your integer to -75, what would the character be?

Similarly, if you set the integer to zero, then placed an "A" in the character variable, what value would the integer have?

I am more confused by you're answer

 

[(*steve*)]

Ok. I'll stop then.

 

[Harald Kapp]

consider example

union student{
int roll_number;
float mark ;
}record;

memory space
0001
0002
0003
0004
.......

0010

how the union will allocate memory space ?

According to post #13 both integer or float occupy 4 bytes. Assume your memory is 8 bit wide and the starting address of the union is 0x0001. This then means addresses 0x0001 to 0x0003 will be used to store the value of roll_number or mark. Note that both variables occupy the exact same addresses!
When you access the union using the syntax record.roll_number the content of memory addresses 0x001 to 0x003 wil be interpreted as an integer number.
When you access the union using the syntax record.mark the content of memory addresses 0x001 to 0x003 wil be interpreted as a float number.
This may (and typically will) lead to unpredictable behavior of a program if the program(mer) doesn't take care to evaluate the contents of a union in the right context. Read the link in my post #9.

A general computer is not aware of data types. Memory content is just a sequence of bits and bytes. It is a matter of interpretation what you make of these. A byte with hex value 0x41 may represent for example:

• an integer number 65 (decimal)
• an ASCII character ('A')
• a set of bits (1000 0001)
• or whatever you define it to represent.

Therefore mixing data types in a union requires careful handling. When you put a float number into your union record (record.mark), you should not access it as a record.roll_number as the sequence of bits found there may make no sense at all when interpreted as an integer. And vice versa.

 

[anukalp]

Ok. I'll stop then.

I didn't mean that,

But what exactly is stored in those bytes?

Similarly, if you set the integer to zero, then placed an "A" in the character variable, what value would the integer have?
.

int integer = 0;
char character = 'A'

character variable take 1 byte to store 'A' in memory
integer take 4 byte to store integer value 0 in memory

Each has different size to store their value

They will occupy the same memory location ?

 

[(*steve*)]

@Harald Kapp I think you mean memory locations x0001 to x0004. Also it may be important if the machine is big big endian, little little endian, big little endian or little big endian for longs as to which byte is first.

@anukalp my explanations are not helping you. I do not wish to confuse you further. Perhaps someone else can put things a way that works for you.

 

[gorgon]

Here's a good description of unions in C including an example what can go wrong when a union is handled without the required care.

Did you read this description? From what you write, I don't think you did.
This text explains very good all sides of the 'union' problem. You could test this out yourself from the examples given.
You should also read about, and test how the 'structure' declaration works to understand the difference.

When starting to learn C, 'union' is not the first declaration to use. Depending on what program types you do, you may never need it at all.