Inconsistent behavior for Interval type with document

[yyy1000]

Describe the bug

When I take #9524, I found the behavior of the function is inconsistent with the actual CLI output.
Docs:
https://github.com/apache/arrow-datafusion/blob/eebdbe8be61ba1944b7a7e14d3edeed0259a7d74/docs/source/user-guide/sql/data_types.md?plain=1#L32-L38
Output from local CLI:
DataFusion CLI v35.0.0
❯ select arrow_typeof(interval '1 month');
+---------------------------------------------------------------------+
| arrow_typeof(IntervalMonthDayNano("79228162514264337593543950336")) |
+---------------------------------------------------------------------+
| Interval(MonthDayNano) |
+---------------------------------------------------------------------+
1 row in set. Query took 0.001 seconds.

To Reproduce

Run the same sql

Expected behavior

No response

Additional context

I'm researching whether it's because the docs is outdated or there's some issue related to Interval type

[yyy1000]

take

[alamb]

I don't think this is related to arrow_typeof, I think think it is that the interval syntax always returns a type of IntervalMonthDayNano ... for reasons I can't remember now.

[yyy1000]

I don't think this is related to arrow_typeof, I think think it is that the interval syntax always returns a type of IntervalMonthDayNano ... for reasons I can't remember now.

Yeah, I also think so.
I found the reason is #5801
So I think for this issue, I can just update the document.
Also I have a question, when IntervalYearMonth will be used? If it will not be used maybe we can remove this type? @alamb

[alamb]

Also I have a question, when IntervalYearMonth will be used? If it will not be used maybe we can remove this type? @alamb

I don't understand this question -- I think the idea is that IntervalYearMonth is the most general interval type that can support most non fixed width intervals (e.g. 1 month vs 28, 29, 30, 31 days)