You could use a self-join:
SELECT r.id, r.recipe_id, r.rating FROM recipes_ratings r LEFT JOIN recipes_ratings r2 ON r.recipe_id = r2.recipe_id AND r2.rating>r.rating WHERE r2.rating IS NULL
Please see a fiddle here.
To be more specific I have a table with recipes. In another table I keep ratings associated with these recipes. Each rating has an unique id, the second column is the recipe's id, the third column is the rating (from 1-5).
Expected result is the recipe_id, where all ratings associated with that recipe_id are the highest.
I hope you understand, I have no idea how to approach such query properly.
Hopefully I've posted my question properly, it's my first one.