Allow zoneinfo objects

[mroeschke]

closes #915

[martindurant]

Suggested change

	if isinstance(dtype.tz, zoneinfo.ZoneInfo):
	if getattr(dtype.tz, "zone", False):

?

[martindurant]

update cibuildwheel to 2.16.5 ?

Aside from the wheels, a name is coming out as type int in writer, which is not allowed by parquet (in test_tz_zoneinfo)

[mroeschke]

Sorry about that delay here. I fix the test where the column name was coming out as an int

[martindurant]

There is some dask config option that says "don't use dask-expr" we'll need to set, since they didn't bother to implement fastparquet stuff (that already existed).

[martindurant]

Still hitting dask-expr code somehow:

../../../micromamba-root/envs/test_env/lib/python3.10/site-packages/dask_expr/_collection.py:3152: in to_parquet
    return to_parquet(self, path, **kwargs)
../../../micromamba-root/envs/test_env/lib/python3.10/site-packages/dask_expr/io/parquet.py:335: in to_parquet
    engine = _set_parquet_engine(engine=engine, meta=df._meta)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

engine = 'fastparquet', meta = Empty DataFrame
Columns: [unique, id]
Index: []

    def _set_parquet_engine(engine=None, meta=None):
        # Use `engine` or `meta` input to set the parquet engine
        if engine == "fastparquet":
>           raise NotImplementedError("Fastparquet engine is not supported")
E           NotImplementedError: Fastparquet engine is not supported

[mroeschke]

Added an xfail on test_sorted_row_group_columns_with_filters for now since it doesn't seem directly related to the zoneinfo support directly

[martindurant]

Would you also like to fix the

ValueError: assignment destination is read-only

errors from new pandas?

[mroeschke]

I would prefer to address that in a follow up to avoid to scope creep in this PR