error: aggregate value used where an integer was expected

Question

I am having following union

union data {
     uint64_t val;
     struct{
     ....
     }
};

and I have a function

func(union data mydata[])
{
    printf("%llu",(uint64_t)mydata[0]); // Here is the error
}

When i compile this code it is giving following error

error: aggregate value used where an integer was expected

Solution 1

You are failing to access a field of the indexed union array: mydata[0] is a value of type union data, and can't be cast to uint64_t.

You need to access the proper union member:

printf("%" PRIu64, mydata[0].val);

to select the uint64_t value. No need for the cast.

Also: Use PRIu64 to portably print 64-bit values, you can't assume that %llu is the right format specifier.

Solution 2

The "dirty" solution to access the first member of the n^th element of the array of unions pointed to by mydata as an int64_t without knowing its name is:

#include <inttypes.h>
#include <stdio.h>

union data 
{
  uint64_t val;
  ...
};

func(union data mydata[])
{ 
  size_t n = 3;
  printf("%"PRIu64, *((uint64_t *)(mydata + n));     
}

This works as the first member's address of a union or struct is guaranteed to be the same as the address of the union or struct itself.

Comments

[Grijesh Chauhan]

I think suggest OP PRIu64 also

[unwind]

@GrijeshChauhan True, I edited that fix into the answer. Thanks.

[radsdau]

Vote for alerting me to the existence of PRIu64 etc :)