[Python][Packaging] Windows wheels fail due to `Cannot locate timezone 'UTC': Timezone database not found at "C:\Users\ContainerAdministrator\Downloads\tzdata"`

[raulcd]

Describe the bug, including details regarding any error messages, version, and platform.

I was trying to build Windows wheels on #36925 and found the following error:
https://github.com/ursacomputing/crossbow/actions/runs/5735683765/job/15543833056#step:9:774
It does seems that nightly Windows wheels have also been failing due to this for the last couple of days:

• wheel-windows-cp310-amd64
• wheel-windows-cp311-amd64
• wheel-windows-cp38-amd64
• wheel-windows-cp39-amd64

This seems to fail since: #36942 was merged.

 _________________ test_sequence_timestamp_from_int_with_unit __________________

    @pytest.mark.pandas
    def test_sequence_timestamp_from_int_with_unit():
        # TODO(wesm): This test might be rewritten to assert the actual behavior
        # when pandas is not installed
    
        data = [1]
    
        s = pa.timestamp('s')
        ms = pa.timestamp('ms')
        us = pa.timestamp('us')
        ns = pa.timestamp('ns')
    
        arr_s = pa.array(data, type=s)
        assert len(arr_s) == 1
        assert arr_s.type == s
>       assert repr(arr_s[0]) == (
            "<pyarrow.TimestampScalar: '1970-01-01T00:00:01'>"
        )

Python310\lib\site-packages\pyarrow\tests\test_convert_builtin.py:1355: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
pyarrow\scalar.pxi:537: in pyarrow.lib.TimestampScalar.__repr__
    ???
Python310\lib\site-packages\pyarrow\compute.py:262: in wrapper
    return func.call(args, options, memory_pool)
pyarrow\_compute.pyx:367: in pyarrow._compute.Function.call
    ???
pyarrow\error.pxi:144: in pyarrow.lib.pyarrow_internal_check_status
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

>   ???
E   pyarrow.lib.ArrowInvalid: Cannot locate timezone 'UTC': Timezone database not found at "C:\Users\ContainerAdministrator\Downloads\tzdata"

pyarrow\error.pxi:100: ArrowInvalid
____________________________ test_timestamp_scalar ____________________________

    def test_timestamp_scalar():
>       a = repr(pa.scalar("0000-01-01").cast(pa.timestamp("s")))

Python310\lib\site-packages\pyarrow\tests\test_scalars.py:158: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
pyarrow\scalar.pxi:537: in pyarrow.lib.TimestampScalar.__repr__
    ???
Python310\lib\site-packages\pyarrow\compute.py:262: in wrapper
    return func.call(args, options, memory_pool)
pyarrow\_compute.pyx:367: in pyarrow._compute.Function.call
    ???
pyarrow\error.pxi:144: in pyarrow.lib.pyarrow_internal_check_status
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

>   ???
E   pyarrow.lib.ArrowInvalid: Cannot locate timezone 'UTC': Timezone database not found at "C:\Users\ContainerAdministrator\Downloads\tzdata"

Component(s)

Packaging, Python

[raulcd]

@danepitkin do you know if this is related to the pytz = pytest.importorskip("pytz") that we add on some of our tests?

[danepitkin]

Yes, I believe so! I'm not super familiar with the root cause, but that's how I remember fixing these types of errors before.