core\common.py:496: ModuleNotFoundError

[dickreuter]

In circumstances that I cannot reproduce in a small code example, I get the following error:

When looking at common.py, it seems that main does indeed not exist. Does anybody know why and under what circumstances this is called?

_tcp\work\win-na-x64-release\py3\mre_venv3\lib\site-packages\pandas\core\base.py:61: in __str__

    return self.__unicode__()

_tcp\work\win-na-x64-release\py3\mre_venv3\lib\site-packages\pandas\core\frame.py:659: in __unicode__

    width, _ = console.get_console_size()

_tcp\work\win-na-x64-release\py3\mre_venv3\lib\site-packages\pandas\io\formats\console.py:68: in get_console_size

    if com.in_interactive_session():

_tcp\work\win-na-x64-release\py3\mre_venv3\lib\site-packages\pandas\core\common.py:503: in in_interactive_session

    return check_main()

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

 

    def check_main():

>       import __main__ as main

E       ModuleNotFoundError: No module named '__main__'

 

_tcp\work\win-na-x64-release\py3\mre_venv3\lib\site-packages\pandas\core\common.py:496: ModuleNotFoundError

[gfyoung]

I believe that's IPython-specific logic, but I'm 100% familiar with the rationale behind it.

We probably could except on ModuleNotFoundError in that check as well.

cc @jreback @TomAugspurger

[WillAyd]

Can you try this on master? I don't even see the function from your traceback there so I assume you are working with an older copy of pandas (though pd.show_versions() would also be helpful)

[dickreuter]

I’m using pandas 0.23.4, which I think is the latest in anaconda.

…

On 10 Jan 2019, at 05:23, William Ayd ***@***.***> wrote: Can you try this on master? I don't even see the function from your traceback there so I assume you are working with an older copy of pandas (though pd.show_versions() would also be helpful) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[TomAugspurger]

Under what circumstances is `__main__` not defined? If necessary, we can but that import in a `try / except` block. On Thu, Jan 10, 2019 at 2:18 AM Nicolas Dickreuter <[email protected]> wrote:

…

I’m using pandas 0.23.4, which I think is the latest in anaconda. > On 10 Jan 2019, at 05:23, William Ayd ***@***.***> wrote: > > Can you try this on master? I don't even see the function from your traceback there so I assume you are working with an older copy of pandas (though pd.show_versions() would also be helpful) > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#24690 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABQHIslhsyZg4vc-U9Kb7z6Gq9Pq-FzQks5vBvdLgaJpZM4Z4hGD> .

[dickreuter]

In my case I’m running pytest. When I run it individually on the failing test it passes, but then I run it in conjunction with a pylint test, then I get this error.

…

On 10 Jan 2019, at 12:39, Tom Augspurger ***@***.***> wrote: Under what circumstances is `__main__` not defined? If necessary, we can but that import in a `try / except` block. On Thu, Jan 10, 2019 at 2:18 AM Nicolas Dickreuter ***@***.***> wrote: > I’m using pandas 0.23.4, which I think is the latest in anaconda. > > > On 10 Jan 2019, at 05:23, William Ayd ***@***.***> wrote: > > > > Can you try this on master? I don't even see the function from your > traceback there so I assume you are working with an older copy of pandas > (though pd.show_versions() would also be helpful) > > > > — > > You are receiving this because you authored the thread. > > Reply to this email directly, view it on GitHub, or mute the thread. > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#24690 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ABQHIslhsyZg4vc-U9Kb7z6Gq9Pq-FzQks5vBvdLgaJpZM4Z4hGD> > . > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[TomAugspurger]

but then I run it in conjunction with a pylint test, then I get this

error. What do you mean by "pylint test"? Does `__main__` not existing seem strange to anyone else? If it happen as part of a normal workflow, then it'd be fine to catch that exception. But it's abnormal, I'd like to not catch it to avoid hiding a potentially real issue. On Thu, Jan 10, 2019 at 7:37 AM Nicolas Dickreuter <[email protected]> wrote:

…

In my case I’m running pytest. When I run it individually on the failing test it passes, but then I run it in conjunction with a pylint test, then I get this error. > On 10 Jan 2019, at 12:39, Tom Augspurger ***@***.***> wrote: > > Under what circumstances is `__main__` not defined? If necessary, we can > but that import in a `try / except` block. > > On Thu, Jan 10, 2019 at 2:18 AM Nicolas Dickreuter < ***@***.***> > wrote: > > > I’m using pandas 0.23.4, which I think is the latest in anaconda. > > > > > On 10 Jan 2019, at 05:23, William Ayd ***@***.***> wrote: > > > > > > Can you try this on master? I don't even see the function from your > > traceback there so I assume you are working with an older copy of pandas > > (though pd.show_versions() would also be helpful) > > > > > > — > > > You are receiving this because you authored the thread. > > > Reply to this email directly, view it on GitHub, or mute the thread. > > > > — > > You are receiving this because you were mentioned. > > Reply to this email directly, view it on GitHub > > < #24690 (comment)>, > > or mute the thread > > < https://github.com/notifications/unsubscribe-auth/ABQHIslhsyZg4vc-U9Kb7z6Gq9Pq-FzQks5vBvdLgaJpZM4Z4hGD > > > . > > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#24690 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABQHIiyry9J8JBd3cKXSE38zj7QWNVXLks5vB0IogaJpZM4Z4hGD> .

[dickreuter]

WIth pylint test I mean something like the attached. It's a test that runs pylint. The pandas error occurs in a different test, but it only occurs when this pylint test is active.

This PR fixes the issue: #24714

[dickreuter]

It looks like the source of the problem is this: https://github.com/PyCQA/pylint/blob/master/pylint/lint.py#L145

Pylint is mocking main and then pandas cannot find it. Apparently, this is related to a python bug.

[TomAugspurger]

Thanks for looking. So it's only an issue on Python2?

…

On Fri, Jan 11, 2019 at 6:26 AM Nicolas Dickreuter ***@***.***> wrote: It looks like the source of the problem is this: https://github.com/PyCQA/pylint/blob/master/pylint/lint.py#L145 Pylint is mocking main and then pandas cannot find it. Apparently, this is related to a python bug. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#24690 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABQHIhkSvQgXQj44X3r7z1EvS5wyqHUhks5vCILkgaJpZM4Z4hGD> .

[dickreuter]

No it's python3 in my case

[dickreuter]

Actually, the line causing the issue is this: https://github.com/PyCQA/pylint/blob/master/pylint/lint.py#L153. It's only called when pylint is used in multiprocessing.

Not sure pandas should rely on main existing. Probably a fix would make sense both in pylint and pandas?