Seris and DataFrame creation typing #230
Conversation
There was a problem hiding this comment.
I'm a bit hesitant on these changes, because the type Iterable[Hashable] for index and columns might be too wide and include types that are not acceptable as an index or columns. An example is a set, which will soon be deprecated. There may be others as well.
In addition, the type used in frame.pyi for those arguments in the pandas source is Axes.
Having said that, if you just want to make a change for allowing a Scalar argument to the DataFrame.__new__(), please add a test for that case.
tests/test_frame.py
Outdated
| @@ -42,6 +42,7 @@ def test_types_init() -> None: | |||
| dtype=np.int8, | |||
| copy=True, | |||
| ) | |||
| pd.DataFrame(0, index=[0, 1], columns=[0, 1]) | |||
There was a problem hiding this comment.
can you do check(assert_type(pd.DataFrame(0, index=[0, 1], columns=[0, 1]), pd.DataFrame), pd.DataFrame) here?
I know we don't have it above, but we're putting this in with all new tests.
Series.__new__, DataFrame.__new__.DataFrameto be created from a scalar and index/columns.