[bog_gob]

I you accept MB theory of db schema, recordings (a unique audio) do not have an associated date.  MB uses a "release date" and i don't see anything associated with a release group.  My guess is that original date is a derived value from the oldest date of a release in a release group.


scott s.
.

You are correct that there is no explicit date field associated with the recording.  Musicbrainz does still provide a mechanism for determining the recording date range via AREs e.g.: https://musicbrainz.org/recording/e2064c...58f3b9a9ce  you can clearly see the date of recording. We could scrape selective ARE data and store it in a new field, it could even be in the ARE form if that is preferable.

The work for the recording https://musicbrainz.org/work/62c39e8e-18...5f8e2b547d also has links to things like wikidata https://www.wikidata.org/wiki/Q864783. which can provide lots of rich date data.
As you can see here is a lot of additional (non date) data available as well for us to scrape apart from the date which we could associate with the song (the . I'm not suggesting that these fields need to mirror or overwrite existing ID3 tags. We could introduce fields specifically for the scraper to fill. The album/artist scraper I'm playing with grabs the wikidata entry for a musicbrainz release/artist and pulls the wikidata->wikipedia page data (in text form - choosing the appropriate language) into the description/ bio field, and uses other  wikidata fields e.g. the left field is the kodi artist data and the right is the wikidata field:
 
                    ('genre',    'P136'),   
                    ('gender',    'P21'),  
                    ('type',     'P31')
                    ('born',    'P569'),
                    ('died',    'P570'),
                    ('formed',    'P571'),
                    ('formed',    'P571'),
                    ('disbanded',    'P576')

There are similar fields for a recording e.g. publication date (P577), composer (P86) etc...