still very small yet, but Nim's dataframe library (datamancer) has a dplyr api (and it is fast): https://github.com/SciNim/Datamancer

Being in Nim, it will be easy also to add sweet DSLs.