DateTimeIndex.__iter__().next() rounds time to microseconds, when timezone aware

[enritoomey]

Code Sample

>> import pandas as pd
>> datetimeindex = pd.DatetimeIndex(["2018-02-08 15:00:00.168456358"])
>> datetimeindex
DatetimeIndex(['2018-02-08 15:00:00.168456358'], dtype='datetime64[ns]', freq=None)
>> datetimeindex = datetimeindex.tz_localize(datetime.timezone.utc)
>> datetimeindex
DatetimeIndex(['2018-02-08 15:00:00.168456358+00:00'], dtype='datetime64[ns, UTC+00:00]', freq=None)
>> datetimeindex.__getitem__(0)
Timestamp('2018-02-08 15:00:00.168456358+0000', tz='UTC+00:00')
>> datetimeindex.__iter__().__next__()
Timestamp('2018-02-08 15:00:00.168456+0000', tz='UTC+00:00')

Problem description

When using localize DateTimeIndex with nanosecond precision, getitem behavious differs from iter().next behaviour, as when iterating thought the DateTimeIndex the date is round to microseconds. This doen not happends if the DatetimeIndex has no timezone.

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.4.2.final.0
python-bits: 64
OS: Linux
OS-release: 4.9.0-0.bpo.2-amd64
machine: x86_64
processor:
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.22.0
pytest: None
pip: 9.0.1
setuptools: 36.5.0
Cython: None
numpy: 1.14.0
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.2.1
sphinx: None
patsy: None
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.0.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: None
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[TomAugspurger]

It seems like this is just for UTC?

In [81]: list(datetimeindex)[0] == datetimeindex[0]
Out[81]: False

In [82]: list(datetimeindex.tz_convert("US/Central"))[0] == datetimeindex.tz_convert("US/Central")[0]
Out[82]: True

You're welcome to take a look @enritoomey! Presumably in ints_to_pydatetime in pandas/_libs/tslib.pyx