PERF: PeriodArray._format_native_types

[jbrockmendel]

• closes CLN: Refactor PeriodArray._format_native_types to use DatetimeArray #23979 (Replace xxxx with the GitHub issue number)
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

#23979 suggests implementing PeriodArray._format_native_types with self.to_timestamp()._format_native_types. That is faster than this, but this avoids copies, is robust to weekly freqs, and should further improve with #51459.

pi = pd.period_range("2016-01-01", periods=10**5, freq='H')
pa = pi._data
pa[4] = pd.NaT

%timeit pa._format_native_types()
246 ms ± 5.61 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)   # <- PR
411 ms ± 13.4 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)   # <- main

%timeit pa._format_native_types(date_format="%Y-%m-%d")
274 ms ± 8.41 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)   # <- PR
420 ms ± 13.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)   # <- main

In [18]: %timeit pa.to_timestamp()._format_native_types()
164 ms ± 3.43 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

[mroeschke]

This can't be a str?

[jbrockmendel]

In _format_native_types it is annotated as str | float

[jbrockmendel]

woops, thought you were referring to na_rep. this can't be str bc na_rep isn't str

[mroeschke]

Darn, (odd na_rep can be a float but okay)

[mroeschke]

Thanks @jbrockmendel