PERF: .clip with datetime64[ns, UTC] Series regressed

[jreback]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this issue exists on the latest version of pandas.

• I have confirmed this issue exists on the main branch of pandas.

Reproducible Example

1.2.5

i = pd.date_range('20220101', periods=100000, freq='s', tz='UTC')
clipper = i[0:1000].repeat(100)
%timeit pd.Series(i).clip(upper=clipper)
3.47 ms ± 53 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

1.3.5

i = pd.date_range('20220101', periods=100000, freq='s', tz='UTC')
clipper = i[0:1000].repeat(100)
%timeit pd.Series(i).clip(upper=clipper)
452 ms ± 3.35 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

1.4.3

i = pd.date_range('20220101', periods=100000, freq='s', tz='UTC')
clipper = i[0:1000].repeat(100)
%timeit pd.Series(i).clip(upper=clipper)
132 ms ± 794 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

This work-around gets close to the original (you can do better by using np.where as we do in the impl of clip, but nbd).

i = pd.date_range('20220101', periods=100000, freq='s', tz='UTC')
clipper = i[0:1000].repeat(100)
%timeit pd.Series(pd.to_datetime(pd.Series(i.view('i8')).clip(upper=clipper.view('i8')), utc=True))
6.08 ms ± 62.7 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

Installed Versions

Replace this line with the output of pd.show_versions()

Prior Performance

No response

[phofl]

I am getting around 150ms on 1.4.3 too, but looks like we are back to 1.x ms on main again

%timeit ser.clip(upper=clipper)
1.34 ms ± 90.6 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

[jreback]

oh nice!

ok then
can repurpose this to make sure we have sufficient asvs for this