str.extractall with no match returns appropriate MultIndex #19075
Conversation
Codecov Report
@@ Coverage Diff @@
## master #19075 +/- ##
==========================================
- Coverage 91.51% 91.51% -0.01%
==========================================
Files 148 148
Lines 48805 48688 -117
==========================================
- Hits 44665 44555 -110
+ Misses 4140 4133 -7
Continue to review full report at Codecov.
|
pandas/tests/test_strings.py
Outdated
| @@ -1074,26 +1074,28 @@ def test_extractall_single_group_with_quantifier(self): | |||
|
|
|||
| def test_extractall_no_matches(self): | |||
There was a problem hiding this comment.
can you parameterize this with both a 0-len index and a 0-len MI.
| (['a3', 'b3', 'd4c2'], (None, 'i2')), | ||
| (['a3', 'b3', 'd4c2'], ('i1', 'i2')), | ||
| ]) | ||
| def test_extractall_no_matches(self, data, names): |
There was a problem hiding this comment.
By test comments, @jreback meant a comment with the issue number, so like # GH-19075 just below the def test_
|
Ah, yes, sorry; didn't get this step. I've added it. |
pandas/tests/test_strings.py
Outdated
| tm.assert_frame_equal(r, e) | ||
| # one named group. | ||
| r = s.str.extractall('(?P<first>z)') | ||
| e = DataFrame(columns=["first"]) | ||
| e = DataFrame(columns=["first"], index=ei) | ||
| tm.assert_frame_equal(r, e) | ||
| # two named groups. |
There was a problem hiding this comment.
can you put a blank line before each comment; easier to read the tests then
|
@jreback I've pushed the comment newline fix. |
|
thanks @jsnowacki |
Relates and fixes issue #19034. It's a small fix which creates empty MultiIndex instead of Index, with appropriate names, when
str.extractalldoesn't find any matches. Also, a related test inpandas.tests.test_stringshas been updated to reflect the expected result.