Presently event.mediawiki_page_html_content_change_v1 provides the page_title field as <prefix>:<page_title> when the namespace is not the main namespace.
spark-sql (default)> select page from event.mediawiki_page_html_content_change_v1 where year = 2026 and month = 5 and day = 9 and hour = 1 and changelog_kind = 'update' and wiki_id = 'frwiki' and page.page_id = 13459570 limit 1;
page
{"is_redirect":false,"namespace_id":6,"page_id":13459570,"page_title":"Fichier:Logo_Rugby_Club_Savoie_Rumilly_2018.png","redirect_page_link":{"interwiki_prefix":null,"is_redirect":null,"namespace_id":null,"page_id":null,"page_title":null},"revision_count":null}This is slightly asymmetric with the page_title found in other places like the page.page_title field in MediaWiki.
[commonswiki]> select page_id, page_namespace, page_title from page where page_id = 12444991; +----------+----------------+-----------------------------------------+ | page_id | page_namespace | page_title | +----------+----------------+-----------------------------------------+ | 12444991 | 6 | KirstenboshBotGard09292010OverviewN.JPG | +----------+----------------+-----------------------------------------+
It's handy to join on the page_title without the namespace prefix, particularly if the page_id is many hops away in a join, or someone is operating on a list of page titles and didn't have the page_id readily available. That said, the namespace prefix can indeed be handy sometimes, especially when one struggles to determine the name of the namespace.
This is a request to add two fields for ergonomics:
- namespace_prefix: the plaintext prefix string for the namespace (in the example above this may be Fichier or File, for example, maybe good to have the main namespace localized name as well instead of implicit as an empty string, too?).
- page_title_noprefix: the page title stripped of its namespace prefix string. This will also make it so that titles that happen to contain a : character can't be as easily visually confused as potentially bearing a namespace (e.g., enwiki's CD:UK).
Maybe there's a more straightforward way, but after talking with @Ottomata it was suggested to file a task to see what could be workable.