Defined in header <iterator>
template<class I>
  concept random_access_iterator =
    std::bidirectional_iterator<I> &&
    std::derived_from</*ITER_CONCEPT*/<I>, std::random_access_iterator_tag> &&
    std::totally_ordered<I> &&
    std::sized_sentinel_for<I, I> &&
    requires(I i, const I j, const std::iter_difference_t<I> n) {
      { i += n } -> std::same_as<I&>;
      { j +  n } -> std::same_as<I>;
      { n +  j } -> std::same_as<I>;
      { i -= n } -> std::same_as<I&>;
      { j -  n } -> std::same_as<I>;
      {  j[n]  } -> std::same_as<std::iter_reference_t<I>>;
(since C++20)

The concept random_access_iterator refines bidirectional_iterator by adding support for constant time advancement with the +=, +, -=, and - operators, constant time computation of distance with -, and array notation with subscripting.

Iterator concept determination

Definition of this concept is specified via an exposition-only alias template /*ITER_CONCEPT*/.

In order to determine /*ITER_CONCEPT*/<I>, let ITER_TRAITS<I> denote I if the specialization std::iterator_traits<I> is generated from the primary template, or std::iterator_traits<I> otherwise:

  • If ITER_TRAITS<I>::iterator_concept is valid and names a type, /*ITER_CONCEPT*/<I> denotes the type.
  • Otherwise, if ITER_TRAITS<I>::iterator_category is valid and names a type, /*ITER_CONCEPT*/<I> denotes the type.
  • Otherwise, if std::iterator_traits<I> is generated from the primary template, /*ITER_CONCEPT*/<I> denotes std::random_access_iterator_tag.
  • Otherwise, /*ITER_CONCEPT*/<I> does not denote a type and results in a substitution failure.

Semantic requirements

Let a and b be valid iterators of type I such that b is reachable from a, and let n be a value of type std::iter_difference_t<I> equal to b - a. random_access_iterator<I> is modeled only if all the concepts it subsumes are modeled and:

  • (a += n) is equal to b.
  • std::addressof(a += n) is equal to std::addressof(a).
  • (a + n) is equal to (a += n).
  • (a + n) is equal to (n + a).
  • For any two positive integers x and y, if a + (x + y) is valid, then a + (x + y) is equal to (a + x) + y.
  • a + 0 is equal to a.
  • If (a + (n - 1)) is valid, then --b is equal to (a + (n - 1)).
  • (b += -n) and (b -= n) are both equal to a.
  • std::addressof(b -= n) is equal to std::addressof(b).
  • (b - n) is equal to (b -= n).
  • If b is dereferenceable, then a[n] is valid and is equal to *b.
  • bool(a <= b) is true.
  • Every required operation has constant time complexity.

Equality preservation

An expression is equality preserving if it results in equal outputs given equal inputs.

  • The inputs to an expression consist of its operands.
  • The outputs of an expression consist of its result and all operands modified by the expression (if any).

In specification of standard concepts, operands are defined as the largest subexpressions that include only:

The cv-qualification and value category of each operand is determined by assuming that each template type parameter denotes a cv-unqualified complete non-array object type.

Every expression required to be equality preserving is further required to be stable: two evaluations of such an expression with the same input objects must have equal outputs absent any explicit intervening modification of those input objects.

Unless noted otherwise, every expression used in a requires-expression is required to be equality preserving and stable, and the evaluation of the expression may modify only its non-constant operands. Operands that are constant must not be modified.

Implicit expression variations

A requires-expression that uses an expression that is non-modifying for some constant lvalue operand also implicitly requires additional variations of that expression that accept a non-constant lvalue or (possibly constant) rvalue for the given operand unless such an expression variation is explicitly required with differing semantics. These implicit expression variations must meet the same semantic requirements of the declared expression. The extent to which an implementation validates the syntax of the variations is unspecified.


Unlike the LegacyRandomAccessIterator requirements, the random_access_iterator concept does not require dereference to return an lvalue.


Demonstrates a possible implementation of std::distance via C++20 concepts.

#include <iterator>
namespace cxx20 {
   template<std::input_or_output_iterator Iter>
   constexpr std::iter_difference_t<Iter> distance(Iter first, Iter last)
       if constexpr(std::random_access_iterator<Iter>)
           return last - first;
           std::iter_difference_t<Iter> result{};
           for (;first != last;++first)
           return result;
int main() {
    static constexpr auto il = { 3, 1, 4 };
    static_assert(cxx20::distance(il.begin(), il.end()) == 3);
    static_assert(cxx20::distance(il.end(), il.begin()) == -3);

See also

specifies that a forward_iterator is a bidirectional iterator, supporting movement backwards

© cppreference.com
Licensed under the Creative Commons Attribution-ShareAlike Unported License v3.0.