Thanks for the answer.
As I assume Liquibase is doing lookups on the table, a unique index would make sense especially if the table does grow that much. Without an index the database always needs to scan the entire table if Liquibase checks if a changeset has already been applied. So the impact without an index is actually bigger than with the index.
And I don’t see why this would make the code more complex? Liquibase can already create a unique index in database independent way ().