themlc / bwarm-data-load Goto Github PK

Hello. I noticed the table unclaimed_works contain the isrc column as a varchar(11) and the percentileforprioritisation as an integer, whereas the data in recent dump files are in fact varchar(12) (as an ISRC should be) and float.

To make this data run on my database server (I use Postgres so I adapted the scripts) I had to perform the following requests:

ALTER TABLE unclaimed_works ALTER COLUMN isrc TYPE VARCHAR(12);

ALTER TABLE unclaimed_works ALTER COLUMN percentileforprioritisation TYPE FLOAT;

https://github.com/TheMLC/bwarm-data-load/blob/main/MySQL/create_database.sql#L165
https://github.com/TheMLC/bwarm-data-load/blob/main/MySQL/create_database.sql#L175

I notice the same errors occur in the SQL Server create_database.sql script.

Mysql server version: 8.0.23 Homebrew

Running the create_database.sql script causes a max key length error.

I suspect this is related to declaring the FeedProvidersWorkId as a 3000 length VARCHAR and then attempting to set it as the primary key of the table.

Perhaps an auto-incrementing INT would work better in this case? (What if different FeedProviders have the same WorkId? e.g. Spotify has The Beatles' "Penny Lane" with a WorkID of 123 while Apple has WorkId 123 assigned to The Red Hot Chili Peppers' "Californication")

This will likely be a concern for other tables as well as they also reference 3000 length VARCHARs as their primary keys.