To someone unfamiliar with the pandas api I think the SQL version is by far the most readable. It's just a regular join with the `asof` keyword in front. The

To someone unfamiliar with the pandas api I think the SQL version is by far the most readable. It's just a regular join with the `asof` keyword in front. The pandas version assumes both fields have the same name, it separates the on and by conditions in a way that's not necessary in SQL version, and imo the "backwards" keyword is a bit ugly and error prone.