REGR: Fix construction of PeriodIndex from strings

[dsaxton]

• closes Constructing PeriodIndex with string data #26109
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[simonjayhawkins]

Thanks @dsaxton for the PR

[simonjayhawkins]

can you do data = ...

so that result = statement is the bit being tested. (i.e. we are not testing the klass constructor)

can you also use the data from the OP, since that was not monotonic

[simonjayhawkins]

can you add the issue number.

[simonjayhawkins]

could maybe add these cases to test_period_array_ok instead/as well.

[dsaxton]

Thanks for the feedback @simonjayhawkins . Do you know by chance what's happening in the CI? The error is cryptic and seems unrelated: https://dev.azure.com/pandas-dev/e01c9671-576d-40f3-a1a1-7bef78604087/_apis/build/builds/32518/logs/72

[jorisvandenbossche]

The failurs actually seem related to me. Eg on one of the builds (https://dev.azure.com/pandas-dev/pandas/_build/results?buildId=32640&view=logs&jobId=a67b4c4c-cd2e-5e3c-a361-de73ac9c05f9&j=a67b4c4c-cd2e-5e3c-a361-de73ac9c05f9&t=33d2fdd0-c376-5f94-e6d3-957bdd23a3b8) I see

==================================== ERRORS ==================================== ERRORS ====================================
______________ ERROR collecting pandas/tests/indexes/test_base.py ______________
pandas/tests/indexes/test_base.py:46: in <module>
    class TestIndex(Base):
        tm.assert_index_equal(PeriodIndex(idx.values), idx)
        tm.assert_index_equal(PeriodIndex(list(idx.values)), idx)
    
        msg = "freq not specified and cannot be inferred"
        with pytest.raises(ValueError, match=msg):
            PeriodIndex(idx.asi8)
        with pytest.raises(ValueError, match=msg):
            PeriodIndex(list(idx.asi8))
    
        msg = "'Period' object is not iterable"
        with pytest.raises(TypeError, match=msg):
            PeriodIndex(data=Period("2007", freq="A"))
    
>       result = PeriodIndex(iter(idx))

pandas/tests/indexes/period/test_constructors.py:158: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
pandas/core/indexes/period.py:215: in __new__
    data = period_array(data=data, freq=freq)
pandas/core/arrays/period.py:855: in period_array
    return PeriodArray._from_sequence(data, dtype=dtype)
pandas/core/arrays/period.py:197: in _from_sequence
    freq = freq or libperiod.extract_freq(periods)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

>   def extract_freq(ndarray[object] values):
E   ValueError: Buffer has wrong number of dimensions (expected 1, got 0)

pandas/_libs/tslibs/period.pyx:1469: ValueError

which is at least period related.

[jreback]

thanks @dsaxton