Ts pivot hourly

[timmie]

Now I figured out.

Here is the PR for #2153

[wesm]

@changhiskhan do you have time to review this for 0.9.1?

[changhiskhan]

yeah, doesn't look like too many items left. I'll look at it before release.

[changhiskhan]

Tests don't pass. I'm pushing this off til the next release.

[changhiskhan]

@timmie were you able to get the tests to pass on your machine?

[timmie]

When is 0.92 due?

were you able to get the tests to pass on your machine?

I tried them line by line in the IPython. They worked.
What command do you use to run them?

It was the first time I write dedicated unit tests for a module in my life.
I am learning a lot with you guys and hope to get as serious with TDD.

I will recheck on Monday.

[changhiskhan]

@timmie, thanks for the fast response!
0.10 is going to be next release (fingers crossed we don't have to do a 0.92), which will probably be out maybe mid-late December?

generally you can run the tests by doing something like python pandas/tseries/tests/test_util.py from the root pandas directory. And you can run the entire test suite using test_fast.sh (excludes plotting tests) or test.sh.
You need nosetests and coverage installed.

When you run it in IPython your environment already has a bunch of stuff in it so that's probably why there's a discrepancy. The tests have to pass from clean slate.

After you get the unit tests to pass, if you really want to make the code clean and nice, try to merge the methods in pivot and util. Your method should be called by the method in util.pivot_annual (i.e., I would add a check in the cascading ifs in util.pivot_annual to handle hourly frequency and call your function).

Let me know if you have any other questions. As always, we're grateful for your contributions!

[timmie]

generally you can run the tests by doing something like

Thanks for the patient instructions.

try to merge the methods in pivot and util. Your method should be called by the method in util.pivot_annual

actually this is another point were I would like to receive your feedback.
The result given by my function is transposed in comparison to existing privot_annual

I think that my version is more convenient when viewing results later in Excel.
That's why it was not included but put into a separate file.

[timmie]

Please see PR 2183 instead