Alec, I'm a little confused (big news!) ... Have (more) bear with me, please ...
Apologize in advance for all these issues, (not sure if they correspond to the "duty" of support).
A question about the easiest solution: what is the consequence to the reviewer's page of not migrating those data?
About the second solution, which seemed to me (am I right?) the most correct to avoid errors in future upgrades, I found 256 duplicate entries in user_settings - some with up to 3 "locales".
In most cases:
- Relating to "pt_BR" and "en_US" locales (respectively, default locales of the only two journals);
- Only in the few "setting_value" is <> ' ';
- Ditto the "assoc_type" > 0;
So I ask, if possible:
- What is the consequence of this deletion on the page of the reviewers;
- What is the best criterion as follows: priority "setting_value" <> ' ' or default locale of the reviewer (hard work on this case, but...