Change .str.to_datetime to default to microsecond precision for format specifiers "%f" and "%.f"

[gilnribeiro]

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

df = pl.DataFrame({"dates":["2022-08-31 00:00:00.0", "0920-09-18 00:00:00.0"]})

df.with_columns(pl.col("dates").str.to_datetime(format="%Y-%m-%d %H:%M:%S%.f"))

Log output

---------------------------------------------------------------------------
PanicException                            Traceback (most recent call last)
c:\Users\gilnr\Downloads\win_error.ipynb Cell 3 line 2
      1 print(df)
----> 2 df.with_columns(pl.col("dates").str.to_datetime(format="%Y-%m-%d %H:%M:%S%.f"))

File c:\Users\gilnr\Anaconda3\envs\r_env\lib\site-packages\polars\dataframe\frame.py:8235, in DataFrame.with_columns(self, *exprs, **named_exprs)
   8088 def with_columns(
   8089     self,
   8090     *exprs: IntoExpr | Iterable[IntoExpr],
   8091     **named_exprs: IntoExpr,
   8092 ) -> DataFrame:
   8093     """
   8094     Add columns to this DataFrame.
   8095 
   (...)
   8233 
   8234     """
-> 8235     return self.lazy().with_columns(*exprs, **named_exprs).collect(_eager=True)

File c:\Users\gilnr\Anaconda3\envs\r_env\lib\site-packages\polars\lazyframe\frame.py:1749, in LazyFrame.collect(self, type_coercion, predicate_pushdown, projection_pushdown, simplify_expression, slice_pushdown, comm_subplan_elim, comm_subexpr_elim, no_optimization, streaming, background, _eager)
   1746 if background:
   1747     return InProcessQuery(ldf.collect_concurrently())
-> 1749 return wrap_df(ldf.collect())

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

Issue description

This also seems to happen for all dates above "2262-09-18 00:00:00.0" and below "1677-09-18 00:00:00.0".

Note: I only tested the upper and lower boundary of the years until the exception was thrown. Month and day may vary.

Expected behavior

The expected behavior, to my mind, would be a conversion to Datetime. Especially in the cases of "1677-09-18 00:00:00.0" or "2262-09-18 00:00:00.0".

For the case of "0920-09-18 00:00:00.0" the original date that threw this exception, if indeed it should fail at least the error should be more descriptive than the existing one. For example, using the same exception that is thrown when the wrong format is applied pl.ComputeError

Installed versions

--------Version info---------
Polars:               0.20.3
Index type:           UInt32
Platform:             Windows-10-10.0.22621-SP0
Python:               3.9.2 (default, Mar  3 2021, 15:03:14) [MSC v.1916 64 bit (AMD64)]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          <not installed>
connectorx:           <not installed>
deltalake:            <not installed>
fsspec:               <not installed>
gevent:               <not installed>
hvplot:               <not installed>
matplotlib:           <not installed>
numpy:                1.20.2
openpyxl:             <not installed>
pandas:               <not installed>
pyarrow:              <not installed>
pydantic:             <not installed>
pyiceberg:            <not installed>
pyxlsb:               <not installed>
sqlalchemy:           <not installed>
xlsx2csv:             <not installed>
xlsxwriter:           <not installed>

[MarcoGorelli]

thanks for the report - for now I'd suggest explicitly specifying 'us' or 'ms' as your time unit:

   ...: df.with_columns(pl.col("dates").str.to_datetime(format="%Y-%m-%d %H:%M:%S%.f", time_unit='us'))
Out[3]:
shape: (2, 1)
┌─────────────────────┐
│ dates               │
│ ---                 │
│ datetime[μs]        │
╞═════════════════════╡
│ 2022-08-31 00:00:00 │
│ 0920-09-18 00:00:00 │
└─────────────────────┘