Rewrite UDAF reversed expression name
- Dominant language
- Rust
- Stars
- 9.3k
- Forks
- 2.4k
- Avg merge
- 3d 7h
- Merged PRs (30d)
- 344
Description
I think it would be good to eventually move this to a method in `https://github.com/apache/datafusion/pull/11611` though I agree this is good for now. Maybe we can file a ticket to track
_Originally posted by @alamb in https://github.com/apache/datafusion/pull/11611#discussion_r1688699754_
I think we could extend to rewrite the whole expression name, and it could be more straightforward what the name is rewritten
```rust
let mut name = self.name().to_string();
// If the function is changed, we need to reverse order_by clause as well
// i.e. First(a order by b asc null first) -> Last(a order by b desc null last)
if self.fun().name() == reverse_udf.name() {
} else {
replace_order_by_clause(&mut name);
}
replace_fn_name_clause(&mut name, self.fun.name(), reverse_udf.name());
```
Something like
```rust
let mut name = self.name().to_string();
name = self.fun.reverse_name(name)
```
Add `reverse_name` in `AggregateUDFImpl`
```rust
trait AggregateUDFImpl {
fn reverse_name(&self) -> String {
...
}
}
```
Specificially we need to rewrite name for First/Last. For array_agg, the name should be the same.
Contributor guide
Research direction
Start by locating the existing reversed-expression handling that calls replace_order_by_clause and replace_fn_name_clause, then read AggregateUDFImpl and the First, Last, and array_agg implementations. The change is complete when reverse_name centralizes expression-name rewriting, First and Last produce the correct reversed names, and array_agg keeps its name unchanged.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- databases
- Issue type
- Refactor
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100