Etienne Bacher
Etienne Bacher
> Perhaps two functions may be needed on the Rust side for list, one to autocast and one not to autocast. Probably but I don't think it's related to this...
> Users who have been using dplyr for a few years would know that we had to use explicit type of NA because specifying NA in something like case_when() would...
I was thinking doing it on the Rust side. > if even one string is mixed in the list, the whole thing becomes string type. I imagine cases where implicit...
Does it work with NA as in the example of the first post?
> No, only NULLs are treated specially; use strict=FALSE to cast NA to NA_integer_. In this case, could we have `strict=FALSE` by default? I don't think many average users know...
> the Rd file is not generated It is generated, I just had to specify the `@name` otherwise it's named `Series_unnest`
Ah no my bad, the usage section is wrong
I'd rather find a solution here than coming back later for this.
No the Usage section in the docs is still wrong
I think it is overkill to import `rlang` just for a couple of functions, I never feel I'm missing this kind of functionalities when developing `polars`. My point of view...