While working on MBS-3494, I've noticed that there are some duplicate artist credits in the database. For example:
acoustid=> select * from artist_credit_name where artist_credit in (832431, 44954); artist_credit | position | artist | name | join_phrase ---------------+----------+--------+--------+------------- 44954 | 0 | 44954 | 590816 | 832431 | 0 | 44954 | 590816 | (2 rows) acoustid=> select * from artist_credit where id in (832431, 44954); id | name | artist_count | ref_count | created --------+--------+--------------+-----------+------------------------------- 44954 | 590816 | 1 | 3471 | 2011-05-16 16:32:11.963929+00 832431 | 590816 | 1 | 280 | 2011-06-21 18:59:35.986056+00 (2 rows)
For a funny example, see http://beta.musicbrainz.org/artist/89ad4ac3-39f7-470e-963a-56509c546377/aliases
I think this example is a bit more meaningful, because nearly all artist credit are duplicated: http://musicbrainz.org/artist/302bd7b9-d012-4360-897a-93b00c855680/aliases
In review at http://codereview.musicbrainz.org/r/1798/
For a funny example, see http://beta.musicbrainz.org/artist/89ad4ac3-39f7-470e-963a-56509c546377/aliases