CODEWITHSUNDEEP
c
solutery7
solutery7

Reputation: 21

Understanding Unions in C

Hello I am trying to understand how unions in C work. I have created a float called temperature with the number 12345678.

This converts to binary (25 bits): 101111000110000101001110

In the union I created called temp_union, I created a float variable (value2) and an integer array with 4 bytes called value1.

I then store the temperature into the union value2.

When I display value1[0], shouldn't this print the first 8 bits of the float number? And value[1] the next 8 bits, value[2] the next 8 bits and so on..

So displaying value1[0] as an integer, would be 78 (01001110)
Displaying value1[1] as an integer, would be 97 (01100001)
Displaying value1[2] as an integer, would be 188 (10111100)

Instead, I get the following:

value1[0]: 1262248270                                                                                                                          
value1[1]: 32766                                                                                                                               
value1[2]: 0                                                                                                                                   
value1[3]: 0

My code is below:

 #include <stdio.h>

    int main()
    {
    float temperature = 12345678;

    union union_data_type {
      int  value1[4];
      float value2;
    };

    union union_data_type temp_union;

    temp_union.value2 = temperature;

    printf("\n Temperature float value: ");
    printf("%f", temp_union.value2);

    printf("\n  Value 0: ");
    printf("%i", temp_union.value1[0]);


    printf("\n  Value 1: ");
    printf("%i", temp_union.value1[1]);

    printf("\n  Value 2: ");
    printf("%i", temp_union.value1[2]);

    printf("\n  Value 3: ");
    printf("%i", temp_union.value1[3]);


    }

Upvotes: 0

Views: 262

Answers (2)

Serge
Serge

Reputation: 12344

size of float usually is 32 bit, which is 4 bytes. 'int' is usually 4 bytes as well, so I assume that what you have.

So, in your case the union consists of a 4 bytes float and 16 bytes of int array. only first 4 bytes (value1[0]) of the array overlaps with the float member.

your union then looks like the following in memory

  addr     int value1[4]    float value2
 00000000  [0]  4 bytes       4 bytes
 00000004  [1]  4 bytes       ---
 00000008  [2]  4 bytes       ---
 0000000C  [3]  4 bytes       ---

where addr here represents an offset from the stack location provided to you,

When you initialize the float, you also initialize value1[0]. The rest remains uninitialized and is just trash.

float is usually represented by encoding defined in IEEE 754 standard (https://en.wikipedia.org/wiki/Single-precision_floating-point_format) and the bits of it are interepreted by the cpu and the print services. Same bits for int are interpreted differently.

So, your result shows value1[0] which is an int representation of the float bits, and uninitialized values for other members of the array.

Upvotes: 0

R.S.
R.S.

Reputation: 276

When I display value1[0], shouldn't this print the first 8 bits of the float number? And value[1] the next 8 bits, value[2] the next 8 bits and so on..

No. An int usually is 32bit long. So in your case with temp_union.value1[0] you get the first 32 bit of your float value. If you want to acces one byte at a time change your union to

union union_data_type {
  uint8_t  value1[4];
  float value2;
};

For that you will have to #include <stdint.h>. But regarding your understanding of how a union works, you are correct.

Upvotes: 1

Related Questions