BUG: Allow merging on Index vectors #19073
Conversation
gfyoung
added
Regression
Codecov Report
@@ Coverage Diff @@
## master #19073 +/- ##
==========================================
+ Coverage 91.53% 91.53% +<.01%
==========================================
Files 148 148
Lines 48688 48690 +2
==========================================
+ Hits 44566 44568 +2
Misses 4122 4122
Continue to review full report at Codecov.
|
pandas/core/reshape/merge.py
Outdated
| x, (np.ndarray, Series)) and len(x) == len(left) | ||
| is_rkey = lambda x: isinstance( | ||
| x, (np.ndarray, Series)) and len(x) == len(right) | ||
| is_lkey = lambda x: isinstance(x, list_types) and len(x) == len(left) |
There was a problem hiding this comment.
can you use is_list_like here?
There was a problem hiding this comment.
I thought about that, but passing in list or tuple is not accepted (later on, we check for the dtype attribute by comparing dtypes). Those data types will pass the is_list_like check, however.
There was a problem hiding this comment.
maybe lambda x: is_list_like(x) and hasattr(x, 'dtype')?
just trying to avoid hardcoding (actually you could add that as is_array_like in dtypes/common, seems useful
There was a problem hiding this comment.
Yeah, is_array_like makes sense as an abstraction. I'll add that.
| @@ -1370,6 +1370,22 @@ def f(): | |||
| household.join(log_return, how='outer') | |||
| pytest.raises(NotImplementedError, f) | |||
|
|
|||
| def test_merge_datetime_index(self): | |||
There was a problem hiding this comment.
can you parameterize on ndarray, Series, Index (for the df.index.year arg)
gfyoung
force-pushed
the
merge-index-regression
branch
2 times, most recently
from
224ee15 to
6ce573a
Compare
pandas/core/dtypes/inference.py
Outdated
| ------- | ||
| is_array_like : bool | ||
| Whether `obj` has array-like properties. | ||
| """ |
There was a problem hiding this comment.
can you show Index, Series, ndarray (and a scalar) as examples
There was a problem hiding this comment.
as a followup we could replace code in various places to use this
This behavior used to work in v0.19.0 and is consistent with the documentation. Closes pandas-devgh-19038
Used for abstracting checks in DataFrame.merge, but the function itself can be quite useful.
gfyoung
force-pushed
the
merge-index-regression
branch
from
6ce573a to
1a553bc
Compare
|
@jreback : Comments have been addressed, and all is green. PTAL. |
|
very nice @gfyoung thanks! |
This behavior used to work in v0.19.0 and
is consistent with the documentation.
Closes #19038