REF: Refactor Date/TimeLikeOps #24038
Conversation
Moves the implementation / inheritence from indexes to arrays.
|
Hello @TomAugspurger! Thanks for submitting the PR.
|
Codecov Report
@@ Coverage Diff @@
## master #24038 +/- ##
==========================================
- Coverage 42.46% 42.45% -0.01%
==========================================
Files 161 161
Lines 51557 51557
==========================================
- Hits 21892 21889 -3
- Misses 29665 29668 +3
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #24038 +/- ##
==========================================
- Coverage 42.46% 42.45% -0.01%
==========================================
Files 161 161
Lines 51557 51556 -1
==========================================
- Hits 21892 21888 -4
- Misses 29665 29668 +3
Continue to review full report at Codecov.
|
| @@ -80,6 +82,189 @@ def _get_attributes_dict(self): | |||
| return {k: getattr(self, k, None) for k in self._attributes} | |||
|
|
|||
|
|
|||
| class DatelikeOps(object): | |||
| """ | |||
| Common ops for DatetimeIndex/PeriodIndex, but not TimedeltaIndex. | |||
| from pandas import Index | ||
| return Index(self.format(date_format=date_format), | ||
| dtype=compat.text_type) | ||
| strftime.__doc__ = """ |
There was a problem hiding this comment.
Not a big deal, but is there a reason not to put this docstring in the normal docstring place and then accomplish the formatting with @Substitution(...)?
| Common ops for TimedeltaIndex/DatetimeIndex, but not PeriodIndex. | ||
| """ | ||
|
|
||
| _round_doc = ( |
There was a problem hiding this comment.
Can we avoid having these in the TimedeltaArray/Index namespace?
|
|
||
| Returns | ||
| ------- | ||
| DatetimeIndex, TimedeltaIndex, or Series |
| if 'tz' in attribs: | ||
| attribs['tz'] = None | ||
| return self._ensure_localized( | ||
| self._shallow_copy(result, **attribs), ambiguous, nonexistent |
There was a problem hiding this comment.
The array classes don't have _shallow_copy anymore. Use _simple_new directly?
| ------- | ||
| i8 1d array | ||
| """ | ||
| from pandas import Index |
There was a problem hiding this comment.
Is there any way to avoid this? We've pretty assiduously kept the EA subclasses Index-ignorant so far
There was a problem hiding this comment.
That's what the old version did. I haven't looked at which is more appropriate here.
| attribs['freq'] = None | ||
| if 'tz' in attribs: | ||
| attribs['tz'] = None | ||
| return self._ensure_localized( |
There was a problem hiding this comment.
Do the array classes have _ensure_localized?
|
+1 on the scope. Needs tests for the methods added to DTA/TDA/PA |
|
In response to So in response to
That can't be done yet (at least I don't see an easy way), until the indexes have been split from the arrays. Things like |
|
ok with going ahead with this to avoid blocking and making diffs easier. @jbrockmendel |
|
OK. I'll see if I can implement tests in a follow-up. |
|
thanks @TomAugspurger |
No real functional changes, just an inheritance reorganization to make the diff at #24024 smaller.
Changes: