w3c / csvw

Documents produced by the CSV on the Web Working Group
Other
161 stars 57 forks source link

use column variables for language (future feature) #877

Open baskaufs opened 3 years ago

baskaufs commented 3 years ago

In the process of trying to use the Generating RDF from Tabular Data on the Web Recommendation with Wikidata, I discovered that it is not possible to specific the language of a language-tagged literal as a column variable. Apparently column variables can only be used within URI templates, and in a column description the value for the "lang" key is not a URI.

The practical implication of the requirement to hard-code the language is that there must be a separate table column for every different language tag that might be applied to a value for a particular property. A particular example is the Wikidata title property title (https://www.wikidata.org/wiki/Property:P1476), whose value in an RDF serialization is a language-tagged literal and is constrained to have a single value per subject. Under those circumstances, it would be sensible to have one column for the title string and another column for the language tag for the title string. If there were titles with 10 different languages in the table, there would only be 2 columns required. But under the current specification, 10 columns would be required (one for each language applied to a title string).