MayaFlux::Kakshya::DataFieldTraversal< Array > Class Template Reference

On-demand windowing and field operations over an indexed array. More...

#include <DataFieldTraversal.hpp>

Collaboration diagram for MayaFlux::Kakshya::DataFieldTraversal< Array >:

Public Member Functions
void	advance (size_t rows=0)
	Move forward without passing the end of the array.
bool	at_end () const
	Whether there are no elements after the current position.
size_t	batch_size () const noexcept
	Current window size.
auto	current () const
	Borrow the current window, including a short final window.
	DataFieldTraversal (const Array &array, size_t batch_size=1) noexcept
	Borrow an array and set the initial window size.
template<typename Fn >
void	for_each (Fn &&fn) const
	Apply a callable to each available element in the array.
template<typename Fn >
void	for_each_in_window (Fn &&fn) const
	Apply a callable to each available element in the current window.
size_t	position () const noexcept
	Current zero-based element position.
template<typename Acc , typename Fn >
Acc	reduce (Acc init, Fn &&fn) const
	Fold all available elements into an accumulator.
size_t	remaining () const
	Number of elements after the current position.
void	seek (size_t row)
	Set the position, clamped to the end of the array.
void	set_batch_size (size_t rows) noexcept
	Set the window size for subsequent operations.
template<typename ValueT > requires requires(const Array& array, std::string_view name) { array.template to_nddata<ValueT>(name); }
auto	to_nddata (std::string_view field_name) const
	Derive homogeneous NDData from one field across the entire array.
template<typename ValueT > requires requires(const Array& array, std::string_view name) { array.slice(size_t {}, size_t {})->template to_nddata<ValueT>(name); }
auto	window_to_nddata (std::string_view field_name) const
	Derive homogeneous NDData from one field in the current window.

Private Attributes
const Array *	m_array
size_t	m_batch_size
size_t	m_position {}

Detailed Description

template<FieldArray Array>
class MayaFlux::Kakshya::DataFieldTraversal< Array >

On-demand windowing and field operations over an indexed array.

A traversal maintains an independent position and batch size without owning the array. Field projection is available when the array and its slices provide to_nddata<T>(field_name). The same traversal works with any FieldArray; neither it nor the array needs to name a concrete record format.

Usage:

DataFieldTraversal traversal(items, 64);
while (!traversal.at_end()) {
    auto scores = traversal.window_to_nddata<double>("score");
    traversal.for_each_in_window([](const auto& item) {
        auto name = item.text("name");
    });
    traversal.advance();
}

Template Parameters

Array

Indexed array satisfying FieldArray.

Note

Borrowed elements and windows remain subject to the array's own lifetime and invalidation rules. Keep the array at a stable address while using a traversal.

Definition at line 46 of file DataFieldTraversal.hpp.

---

The documentation for this class was generated from the following file:

• DataFieldTraversal.hpp