Uh oh!
There was an error while loading. Please reload this page.
silence the dask dataframe upstream-dev errors - #4757
Conversation
Uh oh!
There was an error while loading. Please reload this page.
max-sixty
commented
Jan 3, 2021
Thanks @keewis How should we think about changing the test code vs xfailing? IIUC the current test code should work, but doesn't because of the pandas issue. Would we expect to revert this after pandas fixes its issue? |
Co-authored-by: Maximilian Roos <5635139+max-sixty@users.noreply.github.com>
keewis
commented
Jan 3, 2021
that's a good point. I'm really not sure: in this case we don't change the behavior of the test, but in general it would probably be better to xfail. |
dcherian
commented
Jan 3, 2021
It's a bug in constructing an expected dataframe. Seems OK to change it and not revert once pandas fixes its behaviour. |
max-sixty
left a comment
There was a problem hiding this comment.
in this case we don't change the behavior of the test
I see, SGTM
Some of the
to_dask_dataframetests fail on the upstream-dev CI because the behavior ofpandas.DataFramechanged (see pandas-dev/pandas#38645). For now we can silence these by computing alldaskarrays we pass topandas.DataFrame.isort . && black . && mypy . && flake8