sitemap.xml does not include content translations, only the original ones (the content in its original language)

Hello,

I have the same problem when using different domains:

https://shoppinginromania.ro/sitemap.xml
https://shoopinginromania.com/sitemap.xml

Most Articles have translations. The articles are different content types.

Best regards.

Same problem

(Original Language) is included in its sitemap (domain/es/sitemap.xml). However its translation is not included in domain/en/sitemap.xml

After debugging the module in detail I observed this is not build to use translations at all. The reason for this is that it's using the entity id with the node type as a primary key, so even if you force a translated entity it will identify and overwrite the previous language. To handle this I did the following:

1) in xmlsitemap.module I modified the xmlsitemap_xmlsitemap_process_entity_links function to handle translations:

function xmlsitemap_xmlsitemap_process_entity_links($entity_type, array $ids) {
	
  $entities = entity_load_multiple($entity_type, $ids);
  $anonymous_user = new AnonymousUserSession();

  foreach ($entities as $entity) {
    $link_storage = \Drupal::service('xmlsitemap.link_storage');
    $link = $link_storage->create($entity);
    $link_storage->save($link);
	if($entity_type == "node" || $entity_type == "taxonomy_term")
	{
		$languages = \Drupal::languageManager()->getLanguages();
		$current_language = $entity->language()->getId();
		foreach($languages as $language){
			if($current_language != $language->getId())
			{	
			  try{
				$translation = $entity->getTranslation($language->getId());
				$link_storage = \Drupal::service('xmlsitemap.link_storage');
				$link = $link_storage->create($translation, $language->getId());
				$link_storage->save($link);
			  }catch(Exception $e){
				  //handle exceptions;
			  }
			}
		}
	}
  }
}

In src/XmlSitemapLinkStorage.php I was forced to modify the create method to handle a different id, by appending the language id as well.

 public function create(EntityInterface $entity,$is_translation = false) {
    if (!isset($entity->xmlsitemap)) {
      $entity->xmlsitemap = array();
      if ($entity->id() && $link = $this->load($entity->getEntityTypeId(), $entity->id())) {
        $entity->xmlsitemap = $link;
      }
    }

    $settings = xmlsitemap_link_bundle_load($entity->getEntityTypeId(), $entity->bundle());
    $uri = $entity->url();

    $entity->xmlsitemap += array(
	'type' => $entity->getEntityTypeId(),
	'id' => (string) $entity->id(),
	'subtype' => $entity->bundle(),
	'status' => $settings['status'],
	'status_default' => $settings['status'],
	'status_override' => 0,
	'priority' => $settings['priority'],
	'priority_default' => $settings['priority'],
	'priority_override' => 0,
	'changefreq' => isset($settings['changefreq']) ? $settings['changefreq'] : 0,
    );
	

    if (method_exists($entity, 'getChangedTime')) {
      $entity->xmlsitemap['lastmod'] = $entity->getChangedTime();
    }

    $url = $entity->url();
    // The following values must always be checked because they are volatile.
    $entity->xmlsitemap['loc'] = $uri;
    $entity->xmlsitemap['access'] = isset($url) && $entity->access('view', $this->anonymousUser);
    $language = $entity->language();
	if($is_translation == false){
		$entity->xmlsitemap['language'] = !empty($language) ? $language->getId() : LanguageInterface::LANGCODE_NOT_SPECIFIED;
	}else{
		$id = (string) $entity->id();
		$id .= '_'.$is_translation;
		$entity->xmlsitemap['id'] = $id;
		$entity->xmlsitemap['language'] = $is_translation;
	}
	
    return $entity->xmlsitemap;
  }

Please use this at your own risk, but for me it is working properly: http://shoppinginromania.ro/sitemap.xml and http://shoppinginromania.com/sitemap.xml

@multidimension could you please create a patch?

I will try. I am very limited in terms of time. Also I am not sure this is the best method, but for the time being this works with existing websites. For new websites I would suggest looking into a schema update as well or adding a separate indexing as a primary key.

Patch to support for content translation as in #4

resubmitting patch.

For multilingual sites, simple_sitemap works perfectly. I found it after spent several hours trying to make xmlsitemap run on a multilingual site with no success.

For what I can see, nearly everything is in 8.x-1.x-dev now. Only thing missing was the $is_translation = false parameter from public function create() so that later this variable is undefined.

Do we even need the patch in #11 now?

I believe this is resolved by #2941164: using rel="alternate" rather than multiple sitemaps by language context (Drupal 8) now.

Our SEO expert says, that we still should generate one sitemap per language. This SEO thing is always confusing.
I will just leave my patch here. It provides a new option "Include non-existent translations".

Hello,

<<>>

We have the same kind of problems.
We have 3 domains and each domain has 2 or 3 languages (see below).
The default language is English (common to the 3 domains).
The 2 other languages are not properly indexed into Google.
I am confused with all comments and changes.
I have three questions:

1) Should we create specific sitemaps by languages?
2) Which module(s) and patch(es) would you install for this?
3) How to make a language more important for a specific country?

Thanks community !!

www.abcompany.be/fr
www.abcompany.be/nl (more important language)
www.abcompany.be/en

www.abcompany.fr/fr (more important language)
www.abcompany.fr/en

www.abcompany.nl/nl (more important language)
www.abcompany.nl/en