Homec4science
Diffusion libphutil (production) Browse

libphutil/src/utils/PhutilBufferedIterator.php
42d7ebd477f9production

Last Change
Raw File

PhutilBufferedIterator.php
View Options

<?php
/**
* Simple iterator that loads results page-by-page and handles buffering. In
* particular, this maps well to iterators that load database results page
* by page and allows you to implement an iterator over a large result set
* without needing to hold the entire set in memory.
*
* For an example implementation, see @{class:PhutilExampleBufferedIterator}.
*
* @task impl Methods to Implement
* @task config Configuration
* @task iterator Iterator Implementation
*/
abstract class PhutilBufferedIterator extends Phobject implements Iterator {
private $data;
private $pageSize = 100;
private $naturalKey;
/* -( Methods to Implement )----------------------------------------------- */
/**
* Called when @{method:rewind} is invoked. You should reset any internal
* cursor your implementation holds.
*
* @return void
* @task impl
*/
abstract protected function didRewind();
/**
* Called when the iterator needs a page of results. You should load the next
* result page and update your internal cursor to point past it.
*
* If possible, you should use @{method:getPageSize} to choose a page size.
*
* @return list<wild> List of results.
* @task impl
*/
abstract protected function loadPage();
/* -( Configuration )------------------------------------------------------ */
/**
* Get the configured page size.
*
* @return int Page size.
* @task config
*/
final public function getPageSize() {
return $this->pageSize;
}
/**
* Configure the page size. Note that implementations may ignore this.
*
* @param int Page size.
* @return this
* @task config
*/
final public function setPageSize($size) {
$this->pageSize = $size;
return $this;
}
/* -( Iterator Implementation )-------------------------------------------- */
/**
* @task iterator
*/
final public function rewind() {
$this->didRewind();
$this->data = array();
$this->naturalKey = 0;
$this->next();
}
/**
* @task iterator
*/
final public function valid() {
return (bool)count($this->data);
}
/**
* @task iterator
*/
final public function current() {
return end($this->data);
}
/**
* By default, the iterator assigns a "natural" key (0, 1, 2, ...) to each
* result. This method is intentionally nonfinal so you can substitute a
* different behavior by overriding it if you prefer.
*
* @return scalar Key for the current result (as per @{method:current}).
* @task iterator
*/
public function key() {
return $this->naturalKey;
}
/**
* @task iterator
*/
final public function next() {
if ($this->data) {
$this->naturalKey++;
array_pop($this->data);
if ($this->data) {
return;
}
}
$data = $this->loadPage();
// NOTE: Reverse the results so we can use array_pop() to discard them,
// since it doesn't have the O(N) key reassignment behavior of
// array_shift().
$this->data = array_reverse($data);
}
}
c4science · Help