Wikimedia Belgium
bewikimedia
https://be.wikimedia.org/wiki/Wikimedia_Belgium
MediaWiki 1.45.0-wmf.5
first-letter
Media
Special
Talk
User
User talk
Wikimedia
Wikimedia talk
File
File talk
MediaWiki
MediaWiki talk
Template
Template talk
Help
Help talk
Category
Category talk
TimedText
TimedText talk
Module
Module talk
Translations
Translations talk
Mul language
0
4540
28748
26362
2025-06-14T21:31:39Z
Geertivp
171
impact on Wikidata Query (example)
28748
wikitext
text/x-wiki
For a substantial list of instances '''multiple languages''' typically have identical labels, aliases, and/or descriptions.
This impacts [[Wikidata storage]], and the runtime performance of [[Wikidata Query]], [[ListeriaBot]], [[Pywikibot|Pywikbot]], and other [[Wikimedia tools|Wikimedia Tools]].
== Technical impact ==
As a consequence [[Wikidata]] has storage issues, and [[Wikidata Query]] has performance problems with redundant labels, aliases and descriptions, taking up redundant storage, memory, and processing time. In addition to that, the item transaction history will grow much larger due to unnecessary duplication of labels.
Therefore in 2024 a new functionality was introduced: the ''mul labels'' and ''mul aliases''. If a specific language does not have a label, the mul label is shown (without notice) instead. There is fewer emphasis on the English language. Therefore, as a principle, every item could have a mul label. The description should still be filled for each language, except when the description label would be the same for all instances of that type (e.g. firstname, lastname, category, etc.).
The item description for the mul language itself is never registered (because mul is not a real language).
== Examples ==
* Names of persons, firstnames, lastnames → labels only to be stored for the mul language (no description); the same for aliases
* Descriptions of instances of firstnames, lastnames, categories, → should not be stored; to be retrieved at runtime from the instance language label
* Scolar articles and ISBN editions (the title should be only registered in the mul label, because a publication normally has only a single language). The actual language of the publication should be registered with {{P|P407}}
== Impacted functions ==
* Item creation, item update
* Retrieval of labels (should take mul label as last resort)
* [[Pywikibot|Pywikibot scripts]]
* Personal [[Babel registry]] (not required any more since January 2025; the mul language is automatically activated for everybody)
* [[Wikidata Query]]
* [[OpenRefine]]: register mul label via schema for persons, organisations, projects
* [[QuickStatements]]
=== Wikidata query ===
<syntaxhighlight lang="sparql">
SELECT ?item ?itemLabel WHERE {
SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],mul,en". }
?item wdt:P6104 wd:Q134895452.
}
</syntaxhighlight>
== Wikidata project documentation ==
* [[d:Help:Default_values_for_labels_and_aliases|Only the mul-label is stored]] (Wikidata)
* [[phab:T303677|Do not store obvious item descriptions]] (Phabricator)
== Related problems ==
The same problem holds with inverse and derived statements.
* [[Inverse statement]]s
* [[Derived statement]]s (see Preferences → Gadgets → RelatedItems; M:1 statements should not be stored but retrieved from 1:M relationship)
[[Category:Wikidata]]
[[Category:Pywikibot scripts]]
[[Category:Wikidata Query]]
[[Category:OpenRefine]]
[[Category:ListeriaBot]]
[[Category:QuickStatements]]
tou91e9lbd5e7eyb7qq98x6jigff48g
Category:ListeriaBot
14
7864
28749
2025-06-14T21:32:00Z
Geertivp
171
cat
28749
wikitext
text/x-wiki
[[Category:Wikidata]]
95a54nfv29lt6er4guzs1x7guhd3i6v