std::ranges::fold_left_with_iter, std::ranges::fold_left_with_iter_result

Defined in header `<algorithm>`
Call signature
`template< std::input_iterator I, std::sentinel_for<I> S, class T, __indirectly_binary_left_foldable<T, I> F > constexpr /* see description */ fold_left_with_iter( I first, S last, T init, F f );`	(1)	(since C++23)
`template< ranges::input_range R, class T, __indirectly_binary_left_foldable<T, ranges::iterator_t<R>> F > constexpr /* see description */ fold_left_with_iter( R&& r, T init, F f );`	(2)	(since C++23)
Helper concepts
`template< class F, class T, class I > concept __indirectly_binary_left_foldable = /* see description */;`	(3)	(exposition only*)
Helper class template
`template< class I, class T > using fold_left_with_iter_result = ranges::in_value_result<I, T>;`	(4)	(since C++23)

Left-folds the elements of given range, that is, returns the result of evaluation of the chain expression:
f(f(f(f(init, x₁), x₂), ...), x_n), where x₁, x₂, ..., x_n are elements of the range.

Informally, ranges::fold_left_with_iter behaves like std::accumulate's overload that accepts a binary predicate.

The behavior is undefined if [first, last) is not a valid range.

1) The range is [first, last).

2) Same as (1), except that uses r as the range, as if by using ranges::begin(r) as first and ranges::end(r) as last.

3) Equivalent to:

Helper concepts

template< class F, class T, class I, class U >
concept /*indirectly-binary-left-foldable-impl*/ =
  std::movable<T> &&
  std::movable<U> &&
  std::convertible_to<T, U> &&
  std::invocable<F&, U, std::iter_reference_t<I>> &&
  std::assignable_from<U&,
    std::invoke_result_t<F&, U, std::iter_reference_t<I>>>;

(3A)

(exposition only*)

template< class F, class T, class I >
concept /*indirectly-binary-left-foldable*/ =
  std::copy_constructible<F> &&
  std::indirectly_readable<I> &&
  std::invocable<F&, T, std::iter_reference_t<I>> &&
  std::convertible_to<std::invoke_result_t<F&, T, std::iter_reference_t<I>>,
    std::decay_t<std::invoke_result_t<F&, T, std::iter_reference_t<I>>>> &&
    /*indirectly-binary-left-foldable-impl*/<F, T, I,
      std::decay_t<std::invoke_result_t<F&, T, std::iter_reference_t<I>>>>;

(3B)

(exposition only*)

4) The return type alias. See Return value section for details.

The function-like entities described on this page are niebloids, that is:

Explicit template argument lists cannot be specified when calling any of them.
None of them are visible to argument-dependent lookup.
When any of them are found by normal unqualified lookup as the name to the left of the function-call operator, argument-dependent lookup is inhibited.

In practice, they may be implemented as function objects, or with special compiler extensions.

Parameters

first, last	-	the range of elements to fold
r	-	the range of elements to fold
init	-	the initial value of the fold
f	-	the binary function object

Return value

Let U be std::decay_t<std::invoke_result_t<F&, T, std::iter_reference_t<I>>>.

1) An object of type ranges::fold_left_with_iter_result<I, U>.

The member ranges::in_value_result::in holds an iterator to the end of the range.
The member ranges::in_value_result::value holds the result of the left-fold of given range over f.

If the range is empty, the return value is obtained via the expression equivalent to return {std::move(first), U(std::move(init))};.

2) Same as (1) except that the return type is ranges::fold_left_with_iter_result<ranges::borrowed_iterator_t<R>, U>.

Possible implementations

class fold_left_with_iter_fn
{
    template<class O, class I, class S, class T, class F>
    constexpr auto impl(I&& first, S&& last, T&& init, F f) const
    {
        using U = std::decay_t<std::invoke_result_t<F&, T, std::iter_reference_t<I>>>;
        using Ret = ranges::fold_left_with_iter_result<O, U>;
        if (first == last)
            return Ret{std::move(first), U(std::move(init))};
        U accum = std::invoke(f, std::move(init), *first);
        for (++first; first != last; ++first)
            accum = std::invoke(f, std::move(accum), *first);
        return Ret{std::move(first), std::move(accum)};
    }
 
public:
    template<std::input_iterator I, std::sentinel_for<I> S, class T,
             __indirectly_binary_left_foldable<T, I> F>
    constexpr auto operator()(I first, S last, T init, F f) const
    {
        return impl<I>(std::move(first), std::move(last), std::move(init), std::ref(f));
    }
 
    template<ranges::input_range R, class T,
             __indirectly_binary_left_foldable<T, ranges::iterator_t<R>> F>
    constexpr auto operator()(R&& r, T init, F f) const
    {
        return impl<ranges::borrowed_iterator_t<R>>(
            ranges::begin(r), ranges::end(r), std::move(init), std::ref(f)
        );
    }
};
 
inline constexpr fold_left_with_iter_fn fold_left_with_iter;

Complexity

Exactly ranges::distance(first, last) applications of the function object f.

Notes

The following table compares all constrained folding algorithms:

Fold function template	Starts from	Initial value	Return type
`ranges::fold_left`	left	`init`	`U`
`ranges::fold_left_first`	left	first element	`std::optional<U>`
`ranges::fold_right`	right	`init`	`U`
`ranges::fold_right_last`	right	last element	`std::optional<U>`
`ranges::fold_left_with_iter`	left	`init`	(1) `ranges::in_value_result<I, U>` (2) `ranges::in_value_result<BR, U>`, where `BR` is `ranges::borrowed_iterator_t<R>`
`ranges::fold_left_first_with_iter`	left	first element	(1) `ranges::in_value_result<I, std::optional<U>>` (2) `ranges::in_value_result<BR, std::optional<U>>` where `BR` is `ranges::borrowed_iterator_t<R>`

Feature-test macro	Value	Std	Feature
`__cpp_lib_ranges_fold`	202207L	(C++23)	`std::ranges` fold algorithms

Example

#include <algorithm>
#include <cassert>
#include <functional>
#include <iostream>
#include <ranges>
#include <utility>
#include <vector>
 
int main()
{
    std::vector<int> v {1, 2, 3, 4, 5, 6, 7, 8};
 
    auto sum = std::ranges::fold_left_with_iter(v.begin(), v.end(), 6, std::plus<int>());
    std::cout << "sum: " << sum.value << '\n';
    assert(sum.in == v.end());
 
    auto mul = std::ranges::fold_left_with_iter(v, 0X69, std::multiplies<int>());
    std::cout << "mul: " << mul.value << '\n';
    assert(mul.in == v.end());
 
    // get the product of the std::pair::second of all pairs in the vector:
    std::vector<std::pair<char, float>> data {{'A', 2.f}, {'B', 3.f}, {'C', 3.5f}};
    auto sec = std::ranges::fold_left_with_iter
    (
        data | std::ranges::views::values, 2.0f, std::multiplies<>()
    );
    std::cout << "sec: " << sec.value << '\n';
 
    // use a program defined function object (lambda-expression):
    auto lambda = [](int x, int y){ return x + 0B110 + y; };
    auto val = std::ranges::fold_left_with_iter(v, -42, lambda);
    std::cout << "val: " << val.value << '\n';
    assert(val.in == v.end());
}

Output:

sum: 42
mul: 4233600
sec: 42
val: 42

References

C++23 standard (ISO/IEC 14882:2023):

27.6.18 Fold [alg.fold]

ranges::fold_left (C++23)	left-folds a range of elements (niebloid)
ranges::fold_left_first (C++23)	left-folds a range of elements using the first element as an initial value (niebloid)
ranges::fold_right (C++23)	right-folds a range of elements (niebloid)
ranges::fold_right_last (C++23)	right-folds a range of elements using the last element as an initial value (niebloid)
ranges::fold_left_first_with_iter (C++23)	left-folds a range of elements using the first element as an initial value, and returns a pair (iterator, optional) (niebloid)
accumulate	sums up or folds a range of elements (function template)
reduce (C++17)	similar to `std::accumulate`, except out of order (function template)

Docs

Docs4dev

Title here