DataFrame.dropna bug fix #8366
DataFrame.dropna bug fix #8366
Conversation
|
need a test for this |
…columns in subset arg
| @@ -2552,7 +2552,11 @@ def dropna(self, axis=0, how='any', thresh=None, subset=None, | |||
| agg_obj = self | |||
| if subset is not None: | |||
| ax = self._get_axis(agg_axis) | |||
| agg_obj = self.take(ax.get_indexer_for(subset),axis=agg_axis) | |||
| indices = ax.get_indexer_for(subset) | |||
| if -1 in indices : | |||
There was a problem hiding this comment.
do something more like this (keeping with pandas coding style and more efficient)
indexer = ax.get_indexer_for(subset)
check = indexer == -1
if check.any():
raise KeyError(subset[check])
agg_object = self.take(indicies,axis=agg_axis
jreback
added
Bug
Can't Repro
Missing-data
|
pls add a release note in v0.15.0 Bug Fix section as well (and squash when you are done). ping when green. Git references are here: https://github.com/pydata/pandas/wiki/Using-Git |
|
I added the release note and changed the fix to use the more natural pandas/numpy syntax. I submitted a new pull request with all of this in one commit. (I did something bad to my 8303 branch when I tried to rebase and I couldn't push from it - Hopefully I get the hang of rebasing soon.) On Sep 23, 2014, at 7:23 AM, jreback [email protected] wrote:
|
Bug fix for #8303. Now raise KeyError if any non-existent columns are passed in the subset argument.