Fork me on GitHub

Model usage

Model itself ensures the persistence, removal and refresh behavior.

Persist and Remove#

Call persist() or remove() method to process the passed entity. Persist internally handles the needed insert or update database call. At the end of your work, you should confirm your changes by flushing them. Flushing is internally implemented as a transaction commit; transactions are automatically opened with the first persist or remove call. Not only newly created entities, but also already persisted entities must be persisted once more to promote its changes to the storage. This behavior differs from Doctrine, which automatically saves all changes in attached entities.

$user = $model->users->getById(1);
$user->isEmailSubscribed = false;

// or

To remove call remove() method.

$user = $model->users->getById(1);
// or

You may totally disable cascade behavior by passing the second optional false argument. However, this may lead to inconsistences what is stored in database and what is not, then it's your responsibility to call persist properly.

$author = new Author();
$book = new Book();
$book->author = $author;

// with auto-cascade

// or without auto-cascade
$model->persist($author, false);
$model->persist($book, false);


In some use-cases it is needed to refresh the entity data from the storage, where they may have been changed. However, just repeated fetching entity from repository does not return updated entity though it may run new database select query. This is due to Identity map and Orm data constistency. To solve the refresh need Orm provides Model::refreshAll() method which will refresh all entitites from the storage.

$book = $model->books->getById(1);

sleep(60); // sleep for one minuty

// $book is updated with the latest data from database

Also, some entities may be changed but not persisted. Calling refreshAll() in such case will throw an ` exception. You may allow data override by passing `true argument to the method.

$book = $model->books->getById(1);
$book->title = 'Test';


$book->title = 'Changed title';

assert($book->title === 'Test');
// the title may be actually different, becouse another proces may have changed it
// the "Changed title" value is discareded and replaced by the actual database value


Batch processing is often memory demanding. To free some memory, you may use IModel::clear() method. Calling clear will clear all caches and references to all fetched entities, it also nulls their values & data. Be aware that you should never access these entity after calling the clear method. Also, be carefull not to store any references to this entities.

$lastId = 0;
do {

    $fetched = $model->books->findBy(['id>' => $lastId])->orderBy('id')->limitBy(1000)->fetchAll();
    foreach ($fetched as $book) {
        // do the work

        $lastId = $book->id;

    // release the entites from the memory

} while (!empty($fetched));

// optinally, the final memory release
unset($fetched, $book);