Add option to sort by dictionary keys in sort kernels
- Dominant language
- Rust
- Stars
- 3.6k
- Forks
- 1.3k
- Avg merge
- 2d 18h
- Merged PRs (30d)
- 169
Description
**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**
There are two use cases for this feature:
- Some storage providers or engines are able to guarantee that dictionary keys are already sorted and so sorting could be more efficient by using the keys instead of looking up corresponding strings.
- For the PARTITION BY part of window functions the data does not have to be sorted by the strings, sorting by the keys also ensures a partitioning
**Describe the solution you'd like**
Add a flag `assume_sorted_dictionary` to `SortOptions`. In `sort_to_indices` this flags gets used in the branch for dictionary types and if it is set we sort the keys as a primitive array. The same distinction also needs to be implemented in `build_compare` for the `lexsort_to_indices` kernel.
**Additional context**
Once this is implemented, the window function logic in DataFusion could be adjusted to take advantage of it.
Contributor guide
Research direction
Start by reading SortOptions and the sort_to_indices and build_compare entry points, focusing on their dictionary-type handling. Implement the assume_sorted_dictionary behavior for both sorting paths and verify that dictionary keys determine ordering when the option is enabled while existing behavior remains unchanged otherwise.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- data
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100