Join GitHub today
GitHub is home to over 20 million developers working together to host and review code, manage projects, and build software together.
squeeze works on 0 length arrays #11230
Conversation
jreback
commented on an outdated diff
Oct 3, 2015
| @@ -1717,6 +1717,10 @@ def test_squeeze(self): | ||
| p4d = tm.makePanel4D().reindex(labels=['label1'],items=['ItemA']) | ||
| tm.assert_frame_equal(p4d.squeeze(),p4d.ix['label1','ItemA']) | ||
| + # don't fail with 0 length dimensions | ||
| + empty=pd.DataFrame(pd.np.empty(shape=(0,1))) |
jreback
Contributor
|
sinhrks
added Bug Reshaping
labels
Oct 3, 2015
|
@jreback tests added |
|
What should the behavior be here: empty_series=pd.Series([], name='five').squeeze()numpy returns itself: In [18]: pd.np.empty(0)
Out[18]: array([], dtype=float64)
In [19]: pd.np.empty(0).squeeze()
Out[19]: array([], dtype=float64)...which is the current pandas implementation too. |
|
no squeeze does not return a scalar always a Series (or high dim) |
|
OK @jreback, test added for that too |
|
Also closes pydata#8999 |
jreback
added this to the
0.17.1
milestone
Oct 3, 2015
|
their is a whatsnew for 0.17.1 pls add a release note. |
|
@jreback what's new added, commits squashed, ready to go from this end |
jreback
commented on an outdated diff
Oct 5, 2015
| @@ -42,3 +42,5 @@ Performance Improvements | ||
| Bug Fixes | ||
| ~~~~~~~~~ | ||
| + | ||
| +- Bug in `squeeze` with zero length arrays (:issue:`11230`) & (:issue:`8999`) |
jreback
Contributor
|
|
@jreback done! |
|
merged via aa94ae4 thanks! |
MaximilianR commentedOct 3, 2015
fixes: pydata#11229.
fixes: pydata#8999.
Also a better implementation that avoids
ixShould I add to What's new for 0.17? Or is that closed now?