What are all the different methods of accessing array elements?

Answer 1

#include<iostream>

#include<cassert>

int main()

{

char str[] = "The quick brown fox jumped over the lazy dog.";

// Access the 5th element ('q' @ index 4) in 6 different ways:

char a = *(str+4);

char b = *(4+str);

char c = str[4];

char d = 4[str];

char e = *(&str[2]+2);

char f = *(2+&str[2]);

// Point to 7th element and access the 5th element:

char* p = &str[6];

char g = *(p-2);

// Test for equality:

assert(a==b);

assert(b==c);

assert(c==d);

assert(d==e);

assert(e==f);

assert(f==g);

assert(g=='q');

}

Explanation:

There is really just one version, a, but just as x+y == y+x, it is commutative thus we get version b as well. Note that the array name alone always returns a reference to the start of the array. Once we factor out all the pointer arithmetic, we are essentially just returning the dereferenced value of an address within the array.

Version c is the conventional notation using the subscript operator. However the subscript operator is merely sugar-coating (a notational convenience). Behind the Scenes, we're actually executing a.

Version d is an unusual form that is rarely seen but is perfectly valid. Just as str[4] is functionally equivalent to *(str+4), so 4[str] is functionally equivalent to *(4+str), thus d is the same as b (which is the same as a).

Versions e and f are similar to a and b respectively, but instead of taking the address of the start of the array, we now take the address of the 3rd element (index 2) and then offset by 2 elements to get to the 5th element. In these versions, we can also use a negative offset to move backwards from an element, so long as we remain within the bounds of the array. These versions can also be notated more conventionally using version c; str[2+2] which is the same as a.

Version g is similar to e except we now store the address of the 7th element (index 6) in p and then use pointer arithmetic to obtain the offset of the 5th element. However, this version is not commutative so we cannot use *(2-p), but is functionally the same as a.

There are other variants similar to these 7 but they all ultimately come back to version a once you factor out all the pointer arithmetic. In other words, no method is better or worse than any other method. However version c is favoured for its readability alone, regardless of whether the offset is a compile-time constant or needs to be calculated at runtime. All others are academic but give a better understanding of what's really going on behind the scenes.