Fix period range

[hoxbro]

Fixes #5392

[jlstevens]

Thanks for the fix!

Looks good to me though the new unit test is currently failing with:

    def test_dataset_range_for_pd_period_range(self):
        period_range = pd.period_range("1990", "1991", freq="M")
        expected_range = (pd.Timestamp('1990-01-01 00:00:00'), pd.Timestamp('1991-01-01 00:00:00'))
        curve = Curve((period_range, range(13)))
>       assert curve.range("x") == expected_range
E       AssertionError: assert (Period('1990...991-01', 'M')) == (Timestamp('1...01 00:00:00'))
E         At index 0 diff: Period('1990-01', 'M') != Timestamp('1990-01-01 00:00:00')
E         Full diff:
E         - (Timestamp('1990-01-01 00:00:00'), Timestamp('1991-01-01 00:00:00'))
E         + (Period('1990-01', 'M'), Period('1991-01', 'M'))

Looks like you just need to use Period objects instead of pd.Timestamp here....

Happy to merge once the tests are green.

[hoxbro]

After thinking about it, I think the correct fix is in max_range and not in the PandasInterface.range.

Before my last commit, it would return (np.nan, np.nan) because this raises a TypeError pd.to_datetime(pd.Period("1990", freq="M"))

TypeError                                 Traceback (most recent call last)
Input In [3], in <cell line: 1>()
----> 1 pd.to_datetime(pd.Period("1990", freq="M"))

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/core/tools/datetimes.py:1078, in to_datetime(arg, errors, dayfirst, yearfirst, utc, format, exact, unit, infer_datetime_format, origin, cache)
   1076         result = convert_listlike(arg, format)
   1077 else:
-> 1078     result = convert_listlike(np.array([arg]), format)[0]
   1080 #  error: Incompatible return value type (got "Union[Timestamp, NaTType,
   1081 # Series, Index]", expected "Union[DatetimeIndex, Series, float, str,
   1082 # NaTType, None]")
   1083 return result

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/core/tools/datetimes.py:402, in _convert_listlike_datetimes(arg, format, name, tz, unit, errors, infer_datetime_format, dayfirst, yearfirst, exact)
    400 assert format is None or infer_datetime_format
    401 utc = tz == "utc"
--> 402 result, tz_parsed = objects_to_datetime64ns(
    403     arg,
    404     dayfirst=dayfirst,
    405     yearfirst=yearfirst,
    406     utc=utc,
    407     errors=errors,
    408     require_iso8601=require_iso8601,
    409     allow_object=True,
    410 )
    412 if tz_parsed is not None:
    413     # We can take a shortcut since the datetime64 numpy array
    414     # is in UTC
    415     dta = DatetimeArray(result, dtype=tz_to_dtype(tz_parsed))

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/core/arrays/datetimes.py:2224, in objects_to_datetime64ns(data, dayfirst, yearfirst, utc, errors, require_iso8601, allow_object, allow_mixed)
   2222 order: Literal["F", "C"] = "F" if flags.f_contiguous else "C"
   2223 try:
-> 2224     result, tz_parsed = tslib.array_to_datetime(
   2225         data.ravel("K"),
   2226         errors=errors,
   2227         utc=utc,
   2228         dayfirst=dayfirst,
   2229         yearfirst=yearfirst,
   2230         require_iso8601=require_iso8601,
   2231         allow_mixed=allow_mixed,
   2232     )
   2233     result = result.reshape(data.shape, order=order)
   2234 except ValueError as err:

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/_libs/tslib.pyx:381, in pandas._libs.tslib.array_to_datetime()

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/_libs/tslib.pyx:613, in pandas._libs.tslib.array_to_datetime()

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/_libs/tslib.pyx:755, in pandas._libs.tslib._array_to_datetime_object()

File ~/miniconda3/envs/holoviz/lib/python3.9/site-packages/pandas/_libs/tslib.pyx:586, in pandas._libs.tslib.array_to_datetime()

TypeError: <class 'pandas._libs.tslibs.period.Period'> is not convertible to datetime

[codecov-commenter]

Codecov Report

Merging #5393 (f861ad2) into master (8b2ee78) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #5393   +/-   ##
=======================================
  Coverage   88.07%   88.07%           
=======================================
  Files         301      301           
  Lines       61969    61972    +3     
=======================================
+ Hits        54579    54582    +3     
  Misses       7390     7390           

Impacted Files	Coverage Δ
holoviews/core/data/pandas.py	93.30% <ø> (-0.10%)	⬇️
holoviews/core/util.py	86.06% <100.00%> (+0.03%)	⬆️
holoviews/tests/core/test_utils.py	99.20% <100.00%> (+<0.01%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[philippjfr]

Looks good to me!

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.