-
Notifications
You must be signed in to change notification settings - Fork 87
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Running a full test class fails when each test succeeds #900
Comments
Thanks for the report! |
We're on pandas 1.5.3. It's installed through pip for some reason, although I don't know why. |
Thanks - downgrading to 1.5.3 didn't change the outcome. I will have a closer look somewhere in the next days. |
Thanks for taking a look! Let me know how else I can help. |
I still cannot reproduce the problem. I tested your example in the CI, and it runs fine in Ubuntu, MacOS and Windows, under Python 3.8 to 3.12. The environment is not the same, as I used a different Linux, the packages had been installed via pip instead of conda, and I did not install the bunch of other packages in your environment, but none of these should make a difference, so I'm a bit lost... As long as I'm not able to reproduce this I cannot promise any fix. I will let you know if I can think of some additional tests you can do to clarify this. |
@JonathanDCohen - I have tried a couple of times to reproduce or understand this, but without success. |
I'll see if I can't make a small reproduction |
I had recently a number of issues related to the dynamic patcher cleanup. def setUp(self) -> None:
self.setUpPyfakefs(use_dynamic_patch=False) You need to update |
I'll give it a shot!
…On Thu, Mar 7, 2024 at 1:32 PM mrbean-bremen ***@***.***> wrote:
I had recently a number of issues related to the dynamic patcher cleanup.
You could check if your problem is also related to this by switching off
the dynamic patcher:
def setUp(self) -> None:
self.setUpPyfakefs(use_dynamic_patch=False)
You need to update pyfakefs to the latest version for this, as this
option was not available in the version you used.
If either all tests pass, or no test passes, the problem is indeed related
to the dynamic patcher.
—
Reply to this email directly, view it on GitHub
<#900 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAZFEUA47AE7W7PTJPOYIB3YXCXERAVCNFSM6AAAAAA6OFFB32VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOBUGE4DENRTGM>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
I just published a new release - please use this if you are going to re-test the problem. |
Describe the bug
There is a strange interaction going on between OpenPyxl and Pyfakefs via Pandas. I can read and write DataFrames as normal using the fake filesystem and
pd.read_excel
et al just fine when the tests are run in isolation. However, when I run the entire test class, one test will succeed and the rest will fail.I have already used the solution in #716 to remove the use of LXML to read and write the Excel files.
How To Reproduce
test_a
passes,test_b
andtest_c
fail with essentially the same output:Your environment
Please run the following in the environment where the problem happened and
paste the output.
The text was updated successfully, but these errors were encountered: