BUG: Fix first_last_valid_index, now preserves the frequency. #20569
Conversation
pandas/core/generic.py
Outdated
| @@ -8763,6 +8763,51 @@ def transform(self, func, *args, **kwargs): | |||
| scalar : type of index | |||
| """ | |||
|
|
|||
| def _find_first_valid(self, reversed=False): | |||
There was a problem hiding this comment.
can you rename to _find_valid_index and make the param (required), and call it how='first'|'last'
There was a problem hiding this comment.
cool, looks better.
pandas/core/generic.py
Outdated
| i = is_valid.idxmax() | ||
| if not is_valid[i]: | ||
| return None | ||
| else: |
There was a problem hiding this comment.
don't need the elses (just return)
There was a problem hiding this comment.
hmmmm, I'm not sure, I was thinking about the case when there is no valid values (all nans), In that case this function returns the first index when should be None.
pandas/core/generic.py
Outdated
| return None | ||
| else: | ||
| return i | ||
| else: |
pandas/core/generic.py
Outdated
| if not is_valid.iat[len(self) - i - 1]: | ||
| return None | ||
| else: | ||
| return self.index[len(self) - i - 1] |
jreback
added
Bug
Indexing
|
can you add a whatsnew note as well |
|
@jreback of course, also I forgot asking you where I should add the whatsnew note. |
Removed old methods from Series and DF Added new methods into NDFrame Created new convenient method _find_first_valid
Codecov Report
@@ Coverage Diff @@
## master #20569 +/- ##
==========================================
- Coverage 91.84% 91.84% -0.01%
==========================================
Files 152 152
Lines 49270 49262 -8
==========================================
- Hits 45252 45243 -9
- Misses 4018 4019 +1
Continue to review full report at Codecov.
|
|
@jreback I added whatsnew note under "Bug fixes: Indexing" section. |
|
thanks @mmngreco |
Checklist
git diff upstream/master -u -- "*.py" | flake8 --diffI'm not sure if I should add a test somewhere else. Notice that I've added a new function
_find_first_valid.