query to retrieve all translations and nodes of a language that has been removed from the system?

Question

I am experiencing an issue after a language has been removed from the system. Every time I go on the admin/content page on a drupal 8 site I get:

Error: Call to a member function getId() on null in Drupal\Core\Entity\Entity\EntityViewDisplay->buildMultiple() (line 254 of core/lib/Drupal/Core/Entity/Entity/EntityViewDisplay.php).

This is a known issue on drupal.org but it has not been addressed yet. I am afraid this may be a recurring issue and I am trying to build something to clean up content causing the issue (i.e. the content language was removed from the system).

I am looking for a Drupal query that would return all the nodes(entities)/translations which has a language that has been deleted from the system. I can load all content and loop through them to check $entity->language() == FALSE or $entity->language() == NULL but that does not seem to be performant even with using the batch API. Ideally, db query that returns the IDs of the offending content would be much better. Then I could update them to language neutral or delete the content.

I am wrote something that does something similar to the way node update translations after a language is deleted:

  /**
   * {@inheritdoc}
   */
  public function clearRevisionsLanguage(LanguageInterface $language) {
    return $this->database->update('node_revision')
      ->fields(['langcode' => LanguageInterface::LANGCODE_NOT_SPECIFIED])
      ->condition('langcode', $language->getId())
      ->execute();
  }

I repurposed the above to:

function some_helper_to_update_node_lang($lang_code) {
  \Drupal::database()->update('node_revision')
    ->fields(['langcode' => LanguageInterface::LANGCODE_NOT_SPECIFIED])
    ->condition('langcode', $lang_code)
    ->execute();
}

But the above function requires the language code, which is reasonable but I am not sure how to make this a more flexible in as much to return all content that has a removed language.

Answer 1

I ended executing this function to clean up the content:

function delete_non_existing_language_content() {
  // 1. Get enabled languages
  $languages = \Drupal::languageManager()->getLanguages();
  $lang_codes = array();
  foreach ($languages as $language) {
    $lang_codes[] = $language->getId();
  } 
  // 2. Get all the nodes that have langcode not in the language list.  
  $result = \Drupal::database()->select('node_field_data')
    ->fields('node_field_data')
    ->condition('langcode', $lang_codes, 'NOT IN')
    ->execute();

  // @todo think I can use fetchAllKeyed instead?
  $nids = array_keys($result->fetchAllAssoc('nid'));

  // 3.Delete all of them:
  // @todo Or possibly update them to language neutral code if possible:
  $storage_handler = \Drupal::entityTypeManager()->getStorage('node');
  $entities = $storage_handler->loadMultiple($nids);
  $storage_handler->delete($entities);

}

This worked for my use case, but I don't think it will be performant for large databases.