BUG: Fix tz handling in date_range when start arg has tz, #2926

[Moisan]

This commit should fix issue 2926.

[ghost]

tests please.

[Moisan]

I tested with nosetests pandas and everything was OK. Do you want some sort of output?

[ghost]

Sorry, I meant you've fixed a bug which isn't caught by the current test suite,
so it's important to add a test that fails before the fix and demonstrates that your
change indeed corrects the bug. It'll also prevent future regressions if someone
modifies that part of the code.

The tests covering date_range live at pandas/tseries/tests/test_daterange.py,
please add a new test there.

[xdong]

Hi Moisan,

I am using the stable release 10.1. It's not easy for me to test the latest commit right now. Could you check if your patch fixes the following:

(1) issue #2906 date_range bug?

(2) I found a related bug when using concat:

In [17]: from pandas import *

In [18]: dates_a = [Timestamp('20130219 13:00', tz='US/Eastern'), Timestamp('20130220 13:00', tz='US/Eastern')]

In [19]: dates_b = [Timestamp('20130222 13:00', tz='US/Eastern'), Timestamp('20130223 13:00', tz='US/Eastern')]

In [20]: a = Series(range(2), dates_a)

In [21]: b = Series(range(2), dates_b)

In [22]: concat([a, b])
Out[22]:
2013-02-19 18:00:00-05:00 0
2013-02-20 18:00:00-05:00 1
2013-02-22 18:00:00-05:00 0
2013-02-23 18:00:00-05:00 1

Note that the index of Out[22] is wrong (time shifted by 5 hours).

[Moisan]

Hi xdong,

No my patch does not fix either of these issues. However I do believe they are related to each other. I tried to force the timezone in issue 2906 and found another bug and opened issue 2926. I'll go back to issue 2906 soon :).

[wesm]

Cherry-picked, thank you!