[Python] Use tzdata timezone database on Windows

[asfimport]

In ARROW-13168, we enabled setting the path of the text-based database engine at runtime. This allowed R to use the tzdb package for the timezone database, since it uses the text format.

However, it doesn't seem like tzdata Python package ships that text format. They do have a "compact" text format, which might be compatible with our vendored date library. Otherwise, we'd likely have to wait for binary format support in HowardHinnant/date#564

Reporter: Will Jones / @wjones127
Watchers: Rok Mihevc / @rok

Related issues:

• [C++] Timezone database configuration and access (is related to)

_{Note: This issue was originally created as ARROW-16054. Please see the migration documentation for further details.}

[asfimport]

Antoine Pitrou / @pitrou:
I doubt the compact text format is compatible with the date library, it's just an input format for generating the binary files (IIUC). But worth trying probably.

[asfimport]

Michael Sterling:
It's very possible I'm missing something but would it be possible to enable the setting a timezone-path for Python as well even if users have to figure out how to get the timezone database in the correct format?

[asfimport]

Antoine Pitrou / @pitrou:
@jorisvandenbossche @AlenkaF @milesgranger ^

[asfimport]

Rok Mihevc / @rok:
See https://stackoverflow.com/a/74292266/10615379 for a workaround.

[AlenkaF]

PR to skip the failing tests on Windows that require a tz database: #36996 (interim solution)

[h-vetinari]

Are things still being backported to the branch for 13.0? If so, I think #36996 would make sense, as it will cause test failures on windows otherwise.

CC @raulcd

[raulcd]

Are things still being backported to the branch for 13.0? If so, I think #36996 would make sense, as it will cause test failures on windows otherwise.

CC @raulcd

From my understanding the original error was introduced on #36323 which is not part of 13.0.0

[h-vetinari]

From my understanding the original error was introduced on #36323 which is not part of 13.0.0

Ah OK great! Thanks for the info.

Background: I remember seeing a tzdata failure when I had built from source a while ago, but I had already tested one of the rc's, and it passed as well on windows.