.mean_horizontal() does not work with pl.datetime (but .mean() does)

[Elvynzs]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import polars as pl
import datetime

df = pl.DataFrame({"t1":pl.datetime_range(datetime.datetime(2000,1,1),datetime.datetime(2001,1,1), eager=True),
                  "t2":pl.datetime_range(datetime.datetime(2000,1,1),datetime.datetime(2001,1,1), eager=True) + datetime.timedelta(seconds=3600)})
df.mean() # OK !
df.mean_horizontal() # NOK !

Log output

thread 'polars-2' panicked at crates/polars-core/src/frame/mod.rs:2585:26:
called `Option::unwrap()` on a `None` value
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Error message :

PanicException                            Traceback (most recent call last)
Cell In[2], line 7
      4 df = pl.DataFrame({"t1":pl.datetime_range(datetime.datetime(2000,1,1),datetime.datetime(2001,1,1), eager=True),
      5                   "t2":pl.datetime_range(datetime.datetime(2000,1,1),datetime.datetime(2001,1,1), eager=True) + datetime.timedelta(seconds=3600)})
      6 df.mean() # OK !
----> 7 df.mean_horizontal() # NOK !

File C:\...\polars\dataframe\frame.py:9131, in DataFrame.mean_horizontal(self, ignore_nulls)
   9099 def mean_horizontal(self, *, ignore_nulls: bool = True) -> Series:
   9100     """
   9101     Take the mean of all values horizontally across columns.
   9102 
   (...)
   9129     ]
   9130     """
-> 9131     return wrap_s(self._df.mean_horizontal(ignore_nulls)).alias("mean")

PanicException: called `Option::unwrap()` on a `None` value

Issue description

I did not find this exact issue being reported, but it is releated to #13599

Expected behavior

I expected .mean_horizontal to produce similar result to .mean for datetime objects.

Installed versions

--------Version info---------
Polars:               0.20.31
Index type:           UInt32
Platform:             Windows-10-10.0.22000-SP0
Python:               3.11.8 | packaged by conda-forge | (main, Feb 16 2024, 20:40:50) [MSC v.1937 64 bit (AMD64)]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          <not installed>
connectorx:           <not installed>
deltalake:            <not installed>
fastexcel:            <not installed>
fsspec:               <not installed>
gevent:               <not installed>
hvplot:               <not installed>
matplotlib:           3.8.4
nest_asyncio:         1.6.0
numpy:                1.26.4
openpyxl:             3.1.2
pandas:               2.2.2
pyarrow:              15.0.2
pydantic:             <not installed>
pyiceberg:            <not installed>
pyxlsb:               <not installed>
sqlalchemy:           <not installed>
torch:                <not installed>
xlsx2csv:             <not installed>
xlsxwriter:           <not installed>

[DougieHauser]

Seems like mean_horizontal is using sum_horizontal, which doesn't support addition operations on dates

polars.exceptions.InvalidOperationError: add operation not supported for dtypes `datetime[μs]` and `datetime[μs]

It's possible to add this support by casting the columns to &DataType::Float64 (and back again after applying the mean).
This would make sense when applying mean_horizontal on dates, but not for sum_horizontal since all timestamps start at 1970, and results would seem wrong (for example, adding (2000, 1, 1) + (2000, 1, 1) would result in (2030, 1, 1))

I would be happy to try doing a first time contribution if I get a clear understanding of expected behavior. ℹ️ 🙏