is it a language problem though? it's just lack of tooling.
The dataframe paradigm (a good example being polars) is another good alternative that's more composable (imo).
It is true. I still hate it. I think because it always offers 10 different ways to do the same thing. So it is just too much to remember.