Between Polars and Spark Dataframe APIs (not Koalas) as well as the occasional dplyr, I will gladly abandon Pandas.