TST: Extend datetime64 arith tests to array classes, fix several broken cases #23771
Conversation
|
Hello @jbrockmendel! Thanks for updating the PR.
Comment last updated on November 18, 2018 at 21:11 Hours UTC |
| if self.freq is not None: | ||
| return type(self)(-self._data, freq=-self.freq) | ||
| return type(self)(-self._data) | ||
|
|
There was a problem hiding this comment.
This will be made unnecessary by #23642.
| # RelativeDelta DateOffsets | ||
|
|
||
| # ------------------------------------------------------------- | ||
| # Non-Tick, Non-RelativeDelta DateOffsets |
There was a problem hiding this comment.
Some of the tests currently at the bottom of this file will end up in these two sections. For now leaving them in place to make the diff clearer.
jreback
added
Testing
| @@ -791,6 +795,10 @@ def __sub__(self, other): | |||
| else: # pragma: no cover | |||
| return NotImplemented | |||
|
|
|||
| if is_timedelta64_dtype(result) and isinstance(result, np.ndarray): | |||
There was a problem hiding this comment.
Not really. It would just say # wrap result in TimedeltaArray which is pretty redundant given the one-liner here
| pd.offsets.MonthEnd()]) | ||
| exp = klass([Timestamp('2001-1-1'), Timestamp('2000-2-29')]) | ||
| other = tm.box_expected(other, box_with_array) |
There was a problem hiding this comment.
this file getting kind of hefty. ok with making a subdir if it helps, e.g.
pandas/tests/arithmetic/datetime64/test_offsets....etc.
There was a problem hiding this comment.
I'm hopeful we'll be able to de-duplicate a bunch of these (many moved from series/frame/index tests) and slim it down. If not then subdirs may be the way to go.
|
looks good. ping on green. |
Travis fail is the test_apply segfault. gentle ping |
|
thanks |
Some of the DateOffset arithmetic didn't work with the EA classes; this fixed that. Other pieces of the DateOffset arithmetic will have to wait until TimedeltaArray gets a few more methods.