Fix nunique for MultiIndex, DataFrame, and all NA case with dropna=False
#15962
Conversation
mroeschke
added
bug
github-actions
bot
added
libcudf
| data = [np.nan, np.nan] | ||
| pd_ser = pd.Series(data) | ||
| cudf_ser = cudf.Series(data, nan_as_null=True) |
There was a problem hiding this comment.
I don't think we are actually testing nan case here or do you want to test only NA cases?
There was a problem hiding this comment.
I was hoping to just test the NA case here i.e. where dropna would impact the nunique count
There was a problem hiding this comment.
ah, okay. Can we use None then instead of np.nan?
There was a problem hiding this comment.
Sure thing. Changed to use None
| @@ -2162,3 +2162,14 @@ def test_multi_index_contains_hashable(): | |||
| lfunc_args_and_kwargs=((),), | |||
| rfunc_args_and_kwargs=((),), | |||
| ) | |||
|
|
|||
|
|
|||
| @pytest.mark.parametrize("array", [[1, 2], [1, np.nan], [np.nan] * 2]) | |||
There was a problem hiding this comment.
Same concern as below
As of 72b7ac3, it looks like the diff job is reporting a 0% change. https://github.com/rapidsai/cudf/actions/runs/9486926113 (Looks like there's still some noise in test failures rate between commits e.g. https://github.com/rapidsai/cudf/actions/runs/9473554707 vs https://github.com/rapidsai/cudf/actions/runs/9486926113) |
Co-authored-by: GALI PREM SAGAR <sagarprem75@gmail.com>
Co-authored-by: David Wendt <45795991+davidwendt@users.noreply.github.com>
Co-authored-by: Bradley Dice <bdice@bradleydice.com>
Co-authored-by: Bradley Dice <bdice@bradleydice.com>
|
/merge |
Description
Fixes 3 bugs with
nuniqueMultiIndex.nuniquereturning adictinstead of anint.nunique(dropna=False)with allNAs returning 0 instead of 1DataFrame.nuniquepreserving column class and type in the resultingSeries.indexChecklist