Support as_datetime::<DataType::Interval~>

[NiwakaDev]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

I'd like to call as_datetime::<DataType::Interval~> to address apache/datafusion#7097.

Is it possible to return Some(~), not None?:

arrow-rs/arrow-array/src/temporal_conversions.rs

Line 216 in 03d0505

DataType::Interval(_) => None,

Also, is it possible to uncomment the following lines?:

arrow-rs/arrow-array/src/types.rs

Lines 299 to 301 in 03d0505

	// impl ArrowTemporalType for IntervalYearMonthType {}
	// impl ArrowTemporalType for IntervalDayTimeType {}
	// impl ArrowTemporalType for IntervalMonthDayNanoType {}

Describe the solution you'd like

Describe alternatives you've considered

Additional context

[tustvold]

A DateTime is a point in time, whereas an interval is not, I'm not entirely sure how this would work?

[Jefffrey]

I don't think this is the correct way to go about implementing the extract(x from interval) functionality, since as mentioned there isn't a well defined way to convert an interval into a datetime.

Would probably need to modify the existing temporal kernels:

arrow-rs/arrow-arith/src/temporal.rs

Lines 225 to 232 in c578570

	/// Extracts the years of a given temporal primitive array as an array of integers
	pub fn year<T>(array: &PrimitiveArray<T>) -> Result<Int32Array, ArrowError>
	where
	T: ArrowTemporalType + ArrowNumericType,
	i64: From<T::Native>,
	{
	time_fraction_internal(array, "year", |t| t.year())
	}

To be able to extract from intervals directly, instead of attempting to convert intervals to datetime to then leverage existing functionality of these kernels (which only currently work for datetimes).

That or implement the logic entirely on Datafusion's side.

Note there is an issue on Datafusion repo that relates to this functionality: apache/datafusion#6327