[FEA] Move the implementation of building an empty nested column to the native
Open
feature request
Java
Spark
- Dominant language
- C++
- Stars
- 9.8k
- Forks
- 1.1k
- Avg merge
- 3d 6m
- Merged PRs (30d)
- 278
Description
Currently building an empty nested column is done in Java layer, the performance is not good, and it creates a lot of unnecessary Java objects.
We should move it to the native after figuring out a way to pass the nested data type to the native.
This requirement is from https://github.com/rapidsai/cudf/pull/8173#discussion_r627530793
Contributor guide
Assessment
This issue has not been assessed yet.