TST(string dtype): Fix xfails in test_block_internals.py #60765
Conversation
| def test_construction_with_mixed(self, float_string_frame, using_infer_string): | ||
| # test construction edge cases with mixed types |
There was a problem hiding this comment.
I assume something got lost in refactor, but the code removed here is not actually being tested. Both the result and expected variables are getting shadowed before they are compared
| def test_construction_with_conversions(self): | ||
| # convert from a numpy array of non-ns timedelta64; as of 2.0 this does | ||
| # *not* convert | ||
| arr = np.array([1, 2, 3], dtype="timedelta64[s]") | ||
| df = DataFrame(index=range(3)) | ||
| df["A"] = arr | ||
| df = DataFrame({"A": arr}) |
There was a problem hiding this comment.
The failures in this module go back to the discussion in #60338
I don't think it is important for these tests to use that construction pattern
| @@ -401,14 +383,16 @@ def test_update_inplace_sets_valid_block_values(): | |||
| assert isinstance(df._mgr.blocks[0].values, Categorical) | |||
|
|
|||
|
|
|||
| @pytest.mark.xfail(using_string_dtype(), reason="TODO(infer_string)") | |||
| def test_nonconsolidated_item_cache_take(): | |||
| # https://github.com/pandas-dev/pandas/issues/35521 | |||
|
|
|||
| # create non-consolidated dataframe with object dtype columns | |||
There was a problem hiding this comment.
Is this still non-consolidated?
There was a problem hiding this comment.
Ah nice catch - let me fix that
| # check dtypes | ||
| result = df.dtypes | ||
| expected = Series({"datetime64[us]": 3}) | ||
|
|
There was a problem hiding this comment.
Would adding a tm.assertsimething pass here?
There was a problem hiding this comment.
No - adding tm.assert_series_equal(result, expected) doesn't work on main irrespective of the string data type work.
The data provided is 3 dimensional so I'm not sure why the expected Series is just one element, unless it meant to do list multiplication instead of creating a dictionary
There was a problem hiding this comment.
I'm happy to make that change if we want the list multiplication. I just can't figure out why its here in the first place though, so figured a good removal candidate
There was a problem hiding this comment.
The data provided is 3 dimensional
The data is 2 dimensional, so it creates the dataframe just fine?
In [12]: data = [
...: [datetime(2001, 1, 5), np.nan, datetime(2001, 1, 2)],
...: [datetime(2000, 1, 2), datetime(2000, 1, 3), datetime(2000, 1, 1)],
...: ]
...: df = DataFrame(data)
In [13]: df
Out[13]:
0 1 2
0 2001-01-05 NaT 2001-01-02
1 2000-01-02 2000-01-03 2000-01-01
In [14]: df.dtypes
Out[14]:
0 datetime64[us]
1 datetime64[us]
2 datetime64[us]
dtype: object
But indeed the series should have three elements.
And agreed that it looks like out of place to being tested here, I am fine with removing this
|
Owee, I'm MrMeeseeks, Look at me. There seem to be a conflict, please backport manually. Here are approximate instructions:
And apply the correct labels and milestones. Congratulations — you did some good work! Hopefully your backport PR will be tested by the continuous integration and merged soon! Remember to remove the If these instructions are inaccurate, feel free to suggest an improvement. |
|
Manual backport -> #60781 |
…s.py (#60765) (#60781) TST(string dtype): Fix xfails in test_block_internals.py (#60765) (cherry picked from commit d38706a) Co-authored-by: William Ayd <[email protected]>
No description provided.