Add timetz attribute to DatetimeIndex

[jquinon]

• closes ENH: Add timetz attribute to DatetimeIndex and Timestamp #21358
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

I only added timetz to DatetimeIndex because there was already a timetz method in Timestamp that did essentially the same thing. If it is still worth adding to Timestamp though, I can certainly do that.

[mroeschke]

1. I don't think you need the ~ before ~DatetimeIndex.timetz
2. Can you clarify that this returns local time with timezone information

[mroeschke]

Instead could you use the tz_naive_fixture?

pandas/pandas/conftest.py

Line 290 in 26b3e7d

def tz_naive_fixture(request):

[mroeschke]

Looks good overall!

1. Could you add the timetz attribute to this list: http://pandas.pydata.org/pandas-docs/stable/timeseries.html?#time-date-components

2. Could you check if we have some tests for Timestamp.timetz and add one if we don't have any?

[jquinon]

I didn't add a timetz attribute to Timestamp yet because it inherits a timetz method directly from standard library datetime. So should we add an attribute on top of that? And I assume we don't have any tests right now because it is a standard library thing

[mroeschke]

I don't think you have to add a timetz property to Timestamp, but please go ahead and add some timetz tests for Timestamp if we don't have any so we can make sure the behavior is consistent with datetime.

[codecov]

Codecov Report

Merging #22132 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #22132      +/-   ##
==========================================
+ Coverage   92.06%   92.06%   +<.01%     
==========================================
  Files         169      169              
  Lines       50694    50696       +2     
==========================================
+ Hits        46671    46673       +2     
  Misses       4023     4023

Flag	Coverage Δ
#multiple	90.47% <100%> (ø)	⬆️
#single	42.32% <66.66%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/arrays/datetimes.py	95.52% <100%> (+0.02%)	⬆️
pandas/core/indexes/datetimes.py	95.54% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 776fed3...6cfa9a1. Read the comment docs.

[jquinon]

Ok, gotcha

[jreback]

can we test for timetz on Timestamp (only in PY3 I think)?

[jreback]

can add in the api section for Timestamp though

[jquinon]

The timetz method was already there for Timestamp, and I added the property for DatetimeIndex

[mroeschke]

Sorry forgot one more thing; could you add a timetz test for with the dt accessor (e.g. Series.dt.timetz)? They can be found here.

https://github.com/pandas-dev/pandas/blob/master/pandas/tests/series/test_datetime_values.py

After that one test LGTM.

[jreback]

I find an attribute for this is very confusing, as compared to the function call similar to .time(). would a function call be better here? (of course this would be a break with datetime.datetime)

[mroeschke]

For the PR, I think we should leave it as a property for now since we break this property vs. method for other DatetimeIndex attributes (date, time, weekday), and then in another PR consider converting these to method calls (maybe even before v0.24).

(Although I don't even know why these are method calls for datetime in the first place...)

[jquinon]

@mroeschke how does it look?

[mroeschke]

Any other comments @jreback?

[jreback]

thanks @jquinon nice patch!