[FEA] Remove `nan_as_null` in `pandas3` feature branch
Open
feature request
Python
- Dominant language
- C++
- Stars
- 9.8k
- Forks
- 1.1k
- Avg merge
- 3d 6m
- Merged PRs (30d)
- 278
Description
**Is your feature request related to a problem? Please describe.**
Now that `cudf` fully supports `float`(numpy), `Float`(nullable type) & `Arrow(float)` (arrow types), we should just remove `nan_as_null` to simplify this and close the gap between `cudf` and `pandas` float columns.
**Describe the solution you'd like**
After this change. I'd expect:
1. `float64/32` will only consist of `nan`'s
2. `Float64/32` & `Arrow(float)` will only consist of `NA`'s
Contributor guide
Assessment
This issue has not been assessed yet.