As a user of the RDF dumps I find misleading to have a list of prefixes that are not actually used in the graph it represents.
This might require doing two passes on the data and might not be doable on large dataset like wikidata.
As a user of the RDF dumps I find misleading to have a list of prefixes that are not actually used in the graph it represents.
This might require doing two passes on the data and might not be doable on large dataset like wikidata.
This doesn't affect any of our current technical solutions and is only about the shape of the data itself - unused prefixes do not affect the update process.
This isn't really WDQS related. @Lydia_Pintscher: can you track this on your side as needed?