GH1409 Improve Series.to_numpy typing #1476
Conversation
|
Seems like I should be able to restrict some of the types. |
cmp0xff
left a comment
cmp0xff
left a comment
There was a problem hiding this comment.
One comment to fix the warning on Windows. Others are just ideas.
| copy: bool = False, | ||
| na_value: Scalar = ..., | ||
| **kwargs: Any, | ||
| ) -> np_1darray[np.integer]: ... |
There was a problem hiding this comment.
In #1462 we are introducing np_1darray_anyint etc., which will be enforced in the stubs in the future. But I think this #1476 will be merged before #1462, and I can make the necessary changes in #1462.
There was a problem hiding this comment.
Prob let's make a change after, or hold on this one and I will make the changes once you merge the other ones, up to you, fine with both options!
tests/series/test_series.py
Outdated
| ], | ||
| dtype="datetime64[ns]", | ||
| ) | ||
| s7 = pd.Series(pd.PeriodIndex(dates, freq="M")) |
There was a problem hiding this comment.
There is a conversion function
| s7 = pd.Series(pd.PeriodIndex(dates, freq="M")) | |
| s7 = pd.PeriodIndex(dates, freq="M").to_series() |
tests/series/test_series.py
Outdated
|
|
||
| def test_to_numpy() -> None: | ||
| """Test Series.to_numpy for different types.""" | ||
| s1 = pd.Series(["a", "b", "c"], dtype=str) |
There was a problem hiding this comment.
s1 does not give a hint to the intention of the test. Maybe s_str can help more. Just an idea 🙂
There was a problem hiding this comment.
I have remained all the variables, clearer indeed.
tests/series/test_series.py
Outdated
| s9 = ( | ||
| pd.Series(pd.period_range(start="2017-01-01", end="2017-02-01", freq="1D")) | ||
| .diff() | ||
| .iloc[1:] | ||
| ) |
There was a problem hiding this comment.
Maybe
| s9 = ( | |
| pd.Series(pd.period_range(start="2017-01-01", end="2017-02-01", freq="1D")) | |
| .diff() | |
| .iloc[1:] | |
| ) | |
| s9 = pd.Series([Day(1)]).to_numpy() |
tests/series/test_series.py
Outdated
| ) | ||
|
|
||
| s11 = pd.Series( | ||
| [datetime.datetime.now().date(), datetime.datetime.now().date()] |
There was a problem hiding this comment.
datetime.now will eventually be caught by DTZ005 in the future. But it is fine for now.
cmp0xff
added
Compat
2 participants
to_numpydoes not reflectS1inpd.Series[S1]#1409assert_type()to assert the type of any return valueShould address the most general case for each type, happy to improve with the dtype parameter, from experience this is not used a lot but let me know, good to restrict some of those types.