TST: add test for .unique() dtype preserving #29515
Conversation
alimcmaster1
added
the
Testing
pandas/tests/series/test_dtypes.py
Outdated
| @@ -506,6 +506,33 @@ def test_is_homogeneous_type(self): | |||
| assert Series([1, 2])._is_homogeneous_type | |||
| assert Series(pd.Categorical([1, 2]))._is_homogeneous_type | |||
|
|
|||
| @pytest.mark.parametrize( | |||
There was a problem hiding this comment.
this should go in pandas/tests/test_algos.py.
There was a problem hiding this comment.
Thanks, @jreback - moved the test there.
pandas/tests/series/test_dtypes.py
Outdated
| ([1, 2, 2], [1.0, 2.0], "float64"), | ||
| ([1, 2, 2], [1.0, 2.0], "complex64"), | ||
| ([1, 2, 2], [1.0, 2.0], "complex128"), | ||
| ([True, True, False], [True, False], "bool"), |
There was a problem hiding this comment.
I would really try to use the existing fixtures for this, e.g. any_numpy_dtype (and just cast the input in numpy before construction), skipping those which don't make sense.
There was a problem hiding this comment.
Thanks, @jreback - now using np.sctypes to extract all dtypes and skipping [bytes, str, np.void].
pandas/tests/test_algos.py
Outdated
| @pytest.mark.parametrize( | ||
| "data, uniques, dtype_list", | ||
| [ | ||
| ([1, 2, 2], [1, 2], np.sctypes["int"]), |
There was a problem hiding this comment.
i would rather not use numpy references here and instead use our fixtures or definitions in pandas/conftest.py they are much more comprehensive and include pandas types
There was a problem hiding this comment.
Thanks for the feedback, @jreback - now using any_numpy_dtype to extract the dtypes.
ganevgv
force-pushed
the
tests/unique_preserve_dtype_test
branch
from
94fa0a9 to
14793af
Compare
|
thanks @ganevgv |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diff