Most recordings aren't associated with Works


I’m new to MusicBrainz and evaluating it for use in my web app. I’ve noticed that most recordings aren’t associated with a work. For example, if I do the following query on the DB:

select count(*)
from recording left join l_recording_work on l_recording_work.entity0 =
where l_recording_work.entity1 is null;

I get roughly 14.5 million. Only about 2 million recordings are associated with a work.

Is there any plan to associate most of these recordings with works some day?


Yes – feel free to do it. :stuck_out_tongue:

But in earnest, the data in MusicBrainz is edited by volunteers. While there certainly is the general plan to have a complete database of all music one day, including all works and their associations with recordings, there can’t be an ETA.


Haha! Well, I’ve done a few by hand so far, but at this rate it will take me a few lifetimes to finish!

I just wondered whether there was some way of doing it in bulk that was planned (e.g. by merging another DB).

For now I’ll have to create fake work records based on recordings, and make artist_credit the “writer” of each fake “work”.


Huh ? Please don’t do that. You can’t make assumption the writer of a work is the artist credited in the recording. Only create work if needed (for exemple if you know a recording is a cover of a work, and original recording has no work associated with yet).
If you want to create work, and don’t know who is the writer or composer or anything about it, just create it without these. But please don’t add “fake” data, it serves no purpose and will require even more work to fix.


To be clear, I’m not talking about creating fake data in the MB DB. I have my own DB (for my own web app) which will contain a subset of the MB DB, using a different schema.

Rest assured I’ve only added a few works to MB and I’ve made sure the artists creating it are correct.


Oh ok, i missed the fact you are using your own database :wink:


It is very unlikely since past experiences of importing other DBs have rarely been satisfactory in terms of data quality.

At least, it is very likely that MusicBrainz UI will evolve (or even get revamped) for the best. Works editing can be improved and highlighted, see userscripts.