DOC: clarfiy sum of empty Series case in 0.21.0 whatsnew

[jorisvandenbossche]

Attempt to better highlight the change of empty Series, instead of mainly focussing now on all-NaN (given the recent issues that were raised were often about the empty case)

[jreback]

this is ok, but are you intending to actually update the current docs?

[codecov]

Codecov Report

Merging #18204 into master will decrease coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #18204      +/-   ##
==========================================
- Coverage    91.4%   91.38%   -0.02%     
==========================================
  Files         163      163              
  Lines       50064    50064              
==========================================
- Hits        45759    45751       -8     
- Misses       4305     4313       +8

Flag	Coverage Δ
#multiple	89.19% <ø> (ø)	⬆️
#single	40.36% <ø> (-0.06%)	⬇️

Impacted Files	Coverage Δ
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.8% <0%> (-0.1%)	⬇️
pandas/core/indexes/datetimes.py	95.48% <0%> (+0.09%)	⬆️

---

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 ca737ac...8556b24. Read the comment docs.

[jorisvandenbossche]

this is ok, but are you intending to actually update the current docs?

Well, that would also be nice. But at least backporting it to 0.21.1 will make sure it is in the 0.21 docs by the time that is released, and there will still be upgrading lots of people to 0.21 by that time.

@TomAugspurger to update the current docs, I don't think it is easy to make a new build with those changes? (as then the commit tag is not correctly 0.21.0 etc). Maybe I can just edit the source html a bit if I really want it there?

[TomAugspurger]

Uhm, I can just hardcode __version__ in pandas/__init__.py before building.

@jorisvandenbossche could you also expand the "matplotlib no longer imported"? Last item here: http://pandas.pydata.org/pandas-docs/stable/whatsnew.html#other-api-changes

Something like

Pandas no registers our datetime, datetime64, and Period converters with
matplotlib on import. Matplotlib plot methods (``plt.plot``, ``ax.plot``, ...),
will not nicely format the x-axis for ``DatetiemIndex`` or ``PeriodIndex``
values. You must explicitly register these methods:

.. ipython:: python

   from pandas.tseries import converter
   converter.register()

   fig, ax = plt.subplots()
   plt.plot(pd.date_range('2017', periods=12), range(12))

Pandas built-in ``Series.plot`` and ``DataFrame.plot`` *will* register these
on first-use (:issue:17710).

[TomAugspurger]

Actually, I'll just push that to this branch.