alamb commented on issue #15363:
URL: https://github.com/apache/datafusion/issues/15363#issuecomment-2758210504
In general, adding Catalog/Schema information to functions makes sense to
me, but I think it would be a non trivial effort
> However, suppose I want to register a custom avg for Duration. In that
case, it means that I need to re-implement the whole avg function because the
new implementation will override the previous one when registered.
Another approach could be to implement your own `avg` function as a wrapper
around `avg` that called into the provided implementation. For example
```rust
/// An average that also supports duration
pub struct AvgWithDuration {
// standard DataFusion average impmenetation
inner: Average
}
/// implement aggregate udf for new average, call
impl AggregateUDFImpl for AvgWithDuration {
...
fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
// handle duration
if arg_types[0] == DataType::Duration {
Ok(DataType::Duration)
}
// otherwise, fall back to inner implementation
self.inner.return_type(arg_types)
}
fn accumulator(&self, acc_args: AccumulatorArgs) -> Result<Box<dyn
Accumulator>> {
// handle duration, with accumulator implemented for Duration
if arg_types[0] == DataType::Duration {
Ok(Box::new(DurationAccumulator(...))
}
// otherwise, fall back to inner implementation
self.inner.return_type(arg_types)
}
...
}
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
