I followed the excellent musicbrainz-docker instructions and now I have a MusicBrainz database replica which is running quite nicely for me. Thank you to those who set up this mechanism.
However, I notice that the edit table in my replicated database appears to have zero rows. I am pretty sure that this is an undercount.
Is there a way to include the edit, edit_release, and edit_release_group tables in the replication to my database?
The underlying problem I am trying to solve is that I want to query which of the 2,000+ Release Groups without Releases originally had a Release, but an edit moved the Release to a different Release Group. I hope that the results of that query might be candidates for a merge of the first RG into the second.
I’m not sure if ALL edits will be replicated hourly (IIRC, only open edits are replicated?).
Until you receive a response from an MB developer, you could download mbdump-edit.tar.bz2 and mbdump-editor.tar.bz2 manually.
This is the complete edit history for the core database. If you want to see how metadata has evolved, make sure to grab this dump in addition to the core.
The history includes things like open and closed edits, edit notes, votes, and auto-editor elections. It does not include information about the people who made the edits. For that information, you’ll need the next item as well.
mbdump-editor.tar.bz2
This table includes non-personal user data about the people who’ve enacted the edits enumerated in the database above.
You’re right, the default MusicBrainz-Docker setup doesn’t include the edit-related tables in replication. Those tables are quite large and are usually excluded to keep the replication process lightweight. If you really need them, you’ll have to modify the replication script or use a full database dump instead of the replicated one. The full dumps from MusicBrainz include edit, edit_release, and edit_release_group tables, so importing from there would give you the complete history you’re looking for.
Thank you @proctor5. Re-reading MusicBrainz Database / Download again, it says, " mbdump-edit.tar.bz2 This is the complete edit history for the core database…".
More to the point, the Not included in replication section of the Replication page lists the edit, edit_release, and edit_release_group tables as not included.
I took a look at the Replication_Mechanics. As far as I can tell, there is no way for a replicator like me to pull more tables via the replication mechanism. The choice of which tables to replicate is entirely up to the main database which provides the replication packets.
So, I am downloading mbdump-edit.tar.bz2 and mbdump-editor.tar.bz2. Today that appears to be some 13-14GB of data. I hope that when I open them, I will see how to add them to my local replica of the database.
1 Like
I was able to load these dumps into my replicated database. It involved using a particular incantation to run MBImport.pl on my replicated MusicBrainz instance. See details in How to import edit/editor database dumps into a replicated database on musicbrainz-docker?.