[FEA] GPU-accelerated Feather support.
Open
0 - Backlog
cuIO
feature request
libcudf
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.**
A user with existing Feather data files asked about GPU-accelerated Feather support. cuDF currently supports CPU-based Feather reading but does not use GPU acceleration.
**Describe the solution you'd like**
`cudf.read_feather` should be GPU accelerated.
**Describe alternatives you've considered**
Converting files to another format like Parquet is probably ideal for performance and can leverage existing GPU accelerated I/O. This conversion can be done with pandas.
**Additional context**
Feather format description: https://github.com/wesm/feather/blob/master/doc/FORMAT.md
Contributor guide
Assessment
This issue has not been assessed yet.