Can't cast +00:00 to Europe/Brussels

[MarcoGorelli]

Polars version checks

• [X] I have checked that this issue has not already been reported.

• [X] I have confirmed this bug exists on the latest version of Polars.

Issue description

Can't cast +00:00 timezone to Europe/Brussels

Reproducible example

In [13]: pl.Series(['2020-01-01']).str.strptime(pl.Datetime).dt.with_time_zone('+01:00').dt.cast_time_zone('Europe/Brussels')
---------------------------------------------------------------------------
ComputeError                              Traceback (most recent call last)
Cell In[13], line 1
----> 1 pl.Series(['2020-01-01']).str.strptime(pl.Datetime).dt.with_time_zone('+01:00').dt.cast_time_zone('Europe/Brussels')

File ~/tmp/.311venv/lib/python3.11/site-packages/polars/internals/series/utils.py:98, in call_expr.<locals>.wrapper(self, *args, **kwargs)
     96     expr = getattr(expr, namespace)
     97 f = getattr(expr, func.__name__)
---> 98 return s.to_frame().select(f(*args, **kwargs)).to_series()

File ~/tmp/.311venv/lib/python3.11/site-packages/polars/internals/dataframe/frame.py:5592, in DataFrame.select(self, exprs)
   5504 def select(
   5505     self: DF,
   5506     exprs: (
   (...)
   5511     ),
   5512 ) -> DF:
   5513     """
   5514     Select columns from this DataFrame.
   5515
   (...)
   5589
   5590     """
   5591     return self._from_pydf(
-> 5592         self.lazy().select(exprs).collect(no_optimization=True)._df
   5593     )

File ~/tmp/.311venv/lib/python3.11/site-packages/polars/utils.py:394, in deprecated_alias.<locals>.deco.<locals>.wrapper(*args, **kwargs)
    391 @functools.wraps(fn)
    392 def wrapper(*args: P.args, **kwargs: P.kwargs) -> T:
    393     _rename_kwargs(fn.__name__, kwargs, aliases)
--> 394     return fn(*args, **kwargs)

File ~/tmp/.311venv/lib/python3.11/site-packages/polars/internals/lazyframe/frame.py:1168, in LazyFrame.collect(self, type_coercion, predicate_pushdown, projection_pushdown, simplify_expression, no_optimization, slice_pushdown, common_subplan_elimination, streaming)
   1157     common_subplan_elimination = False
   1159 ldf = self._ldf.optimization_toggle(
   1160     type_coercion,
   1161     predicate_pushdown,
   (...)
   1166     streaming,
   1167 )
-> 1168 return pli.wrap_df(ldf.collect())

ComputeError: Could not parse timezone: 'Europe/Brussels'

Expected behavior

shape: (1,)
Series: '' [datetime[μs, Europe/Brussels]]
[
        2020-01-01 01:00:00 CET
]

Also, the error message should probably say

ComputeError: Could not parse timezone: '+01:00'

as that's the part which can't be parsed by chrono_tz?

Installed versions

---Version info---
Polars: 0.15.16
Index type: UInt32
Platform: Linux-5.10.102.1-microsoft-standard-WSL2-x86_64-with-glibc2.35
Python: 3.11.1 (main, Dec  7 2022, 01:11:34) [GCC 11.3.0]
---Optional dependencies---
pyarrow: 10.0.1
pandas: 1.5.3
numpy: 1.24.1
fsspec: <not installed>
connectorx: <not installed>
xlsx2csv: <not installed>
deltalake: <not installed>
matplotlib: <not installed>

[FObersteiner]

related: https://github.com/pola-rs/polars/issues/6338 - It seems there is some confusion, what cast_time_zone and with_time_zone should do.

[ritchie46]

Assuming

with_time_zone: set tz (localize)
cast_time_zone: convert tz,

All values are stored in UTC in memory by polars. Upon showing data to the users, the UTC times are corrected for the timezone when showed to users/exporting.

with_time_zone leaves the values as is and simply sets a different timezone on the column cast_time_zone take the current time zone and the given time zone and shifts the underlying values by the difference.

That's the idea at least. :)

[FObersteiner]

@ritchie46 sorry, didn't want to cause more confusion than there actually is (post edited).

[MarcoGorelli]

Kinda on the way to having something working for this, getting there 💪