fixed bug where pd.NA was being cast to NaN during formatting #55754
Conversation
passed new test locally
|
im not sure, is it normal that so many CI steps are failing? (it's my first time contributing to pandas) |
Two things to do when the CI is failing: look at the failure message of (some) builds and see if the builds are passing on main. For the first, I'm seeing: https://github.com/pandas-dev/pandas/actions/runs/6703704571/job/18214736331?pr=55754#step:4:226
This is likely not due to your PR 😄 Recent commits to main also have 4 failing. Once you believe the failed builds are not due to your PR and you're not seeing them fail on the most recent commit to main, try merging main into this branch. It will run all the builds again. |
pandas/io/formats/format.py
Outdated
| self.tr_frame = self.tr_frame.iloc[ | ||
| chain(range(row_num), range(_len - row_num, _len)) | ||
| ] |
There was a problem hiding this comment.
Thanks - this is definitely a lot better. I think we can do a little better by using NumPy instead.
self.tr_frame = self.tr_frame.iloc[
np.hstack([np.arange(row_num), np.arange(_len - row_num, _len)])
]
| @@ -171,6 +171,15 @@ def test_repr_truncation(self): | |||
| with option_context("display.max_colwidth", max_len + 2): | |||
| assert "..." not in repr(df) | |||
|
|
|||
| def test_repr_truncation_preserves_na(self): | |||
| # https://github.com/pandas-dev/pandas/issues/55630 | |||
| with option_context("display.max_rows", 10): | |||
There was a problem hiding this comment.
If you set this to 2 instead of 10, I think we could then test for the entire string.
|
it looks like main is having the same CI failures as here |
|
Looks good - I think CI issues will be resolved when you merge main. |
mroeschke
added
Missing-data
|
Thanks @dominiquegarmier! |
replaced
pd.concatwithDataFrame.dropto copypd.NAmore faithfully (see linked issue)pd.NAnot printed properly in REPL when DataFrame gets collapsed #55630