[VL] Improve sort operator performance
- Dominant language
- Scala
- Stars
- 1.6k
- Forks
- 657
- Avg merge
- 2d 14h
- Merged PRs (30d)
- 80
Description
**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**
Currently gluten will use `std::stable_sort` in Window/SMJ/Orderby cases. The `stable_sort` is not optimal and it maybe slower than vanilla Spark in some cases.
**Describe the solution you'd like**
The performance could be improved by several factors:
- Std::sort -> tim_sort
- Null check is not optimal - could skip null check if no null in row vector
- Single key/multiple keys sort
- Enable radix sort for numeric types - radix sort is much faster
- Do inplace sort if no sort payloads
**Describe alternatives you've considered**
N/A
**Additional context**
N/A
Contributor guide
Assessment
This issue has not been assessed yet.