Subclassed reshape #15564
Subclassed reshape #15564
Conversation
Codecov Report
@@ Coverage Diff @@
## master #15564 +/- ##
==========================================
- Coverage 91.06% 91.04% -0.02%
==========================================
Files 137 136 -1
Lines 49307 49171 -136
==========================================
- Hits 44899 44769 -130
+ Misses 4408 4402 -6
Continue to review full report at Codecov.
|
jreback
added
the
Compat
|
can you also have a look over: http://pandas-docs.github.io/pandas-docs-travis/internals.html#subclassing-pandas-data-structures and see if anything needs amendment |
doc/source/whatsnew/v0.9.1.txt
Outdated
| @@ -105,6 +105,8 @@ New features | |||
| - DataFrame.drop now supports non-unique indexes (:issue:`2101`) | |||
| - Panel.shift now supports negative periods (:issue:`2164`) | |||
| - DataFrame now support unary ~ operator (:issue:`2110`) | |||
| - `stack`, `unstack`, and `pivot` operations now preserve subclass family | |||
| @@ -410,13 +413,24 @@ def unstack(obj, level, fill_value=None): | |||
|
|
|||
| if isinstance(obj, DataFrame): | |||
| if isinstance(obj.index, MultiIndex): | |||
| return _unstack_frame(obj, level, fill_value=fill_value) | |||
| unstacked = _unstack_frame(obj, level, fill_value=fill_value) | |||
| else: | |||
There was a problem hiding this comment.
hmm I think this should be done inside the .stack and _unstack_frame functions no?
| else: | ||
| unstacker = _Unstacker(obj.values, obj.index, level=level, | ||
| fill_value=fill_value) | ||
| return unstacker.get_result() | ||
| unstacked = unstacker.get_result() | ||
|
|
There was a problem hiding this comment.
same here, this now has logic in 2 places, the creation itself of the unstacked result, then the return class. should be a part of get_result (it certainly could be another method that is called by get_result) though
| @@ -515,7 +529,7 @@ def factorize(index): | |||
| mask = notnull(new_values) | |||
| new_values = new_values[mask] | |||
| new_index = new_index[mask] | |||
| return Series(new_values, index=new_index) | |||
| return frame._constructor_sliced(new_values, index=new_index) | |||
pandas/tests/frame/test_subclass.py
Outdated
| @@ -125,6 +125,144 @@ def test_indexing_sliced(self): | |||
| tm.assert_series_equal(res, exp) | |||
| tm.assertIsInstance(res, tm.SubclassedSeries) | |||
|
|
|||
| def test_subclass_stack(self): | |||
| df = tm.SubclassedDataFrame([[1, 2, 3], [4, 5, 6], [7, 8, 9]], | |||
There was a problem hiding this comment.
can you add the issue number as a comment
|
thanks @delgadom reshape could certainly use some tlc! |
|
Thanks @jreback - I'm working on your suggestions. Keeping this clean is a bit tough, especially because unstacking is sometimes handled by To me this class seems really nasty and should be handled by functions rather than a class (would you ever use Happy to take a stab if you agree. |
|
@delgadom the reshaping code is certainly in need of refactoring to make things easier :> certainly take a stab at it. Yes the You might want to do a refactor before this PR to make it easier to test / debug. Though if its relatively simple go ahead here. |
|
Yeah I think I'll refactor and then create a new PR. Thanks for the help! |
|
ok! |
git diff upstream/master | flake8 --diffReshape operations in pandas/core/reshape.py modified to use
obj._constructor,obj._constructor_sliced, andobj._constructor_expanddimfor Series and DataFrame. This preserves a subclass family onstack,unstack, andpivotoperations.