Leonardus Chen

How's the refactoring going? If you're short on time and if it is at a state where you can offload some work, I'd be happy to help out

Thanks for the updates!

I think we can skip calling `get_adapted_type` altogether and just use whatever datatype `MetaData.reflect` gives us. The reflection gives the most specific datatype by default. From [docs](https://docs.sqlalchemy.org/en/14/core/reflection.html#reflecting-with-database-agnostic-types): >When the columns...

This is because `pandas.testing` API is only available on later versions of pandas