Fix AttributeError when groupby as_index=False on empty DataFrame #35324
Conversation
simonjayhawkins
added
Groupby
Regression
| # select everything except for the last level, which is the one | ||
| # containing the name of the function(s), see GH 32040 | ||
| result.columns = result.columns.rename( | ||
| [self._selected_obj.columns.name] * result.columns.nlevels | ||
| ).droplevel(-1) | ||
|
|
||
| except ValueError as err: |
There was a problem hiding this comment.
if you leave this unchanged, does it work to put another try/except in the else
There was a problem hiding this comment.
- I was trying to avoid that because nesting exception handling increase complexity.
- But you are right. There is a chance that the
self._aggregate_multiple_funcsalso raiseAttributeErrorand this will change the behavior in that case. - Will update the code as your suggestion. Thanks!
There was a problem hiding this comment.
I agree that nesting exception handling is not ideal, maybe an isinstance check for a Series instead.
There was a problem hiding this comment.
between nesting exception handling and explicit type-checking (also nesting inside another exception handling), I think the first one is less evil :). what do you think?
There was a problem hiding this comment.
from a static typing perspective, the isinstance is probably better. In this case, an AttributeError should be easily avoided. hasattr is not ideal from a static typing/mypy perspective either. However, leave it as a try/except for now and see what others think.
| df = DataFrame(columns=["A", "B", "C"]) | ||
| left = df.groupby(by="A", as_index=False)["B"].sum() | ||
| assert type(left) is DataFrame | ||
| assert left.to_dict() == {"A": {}, "B": {}} |
There was a problem hiding this comment.
can you change this test to be more like (and closer to the issue op)
df = pd.DataFrame(columns=["a", "b"])
grp = df.groupby(by="a", as_index=False)["b"]
result = grp.sum()
expected = pd.DataFrame(index=pd.Int64Index([], dtype="int64"), columns=["a", "b"])
expected["b"] = expected["b"].astype("float64")
tm.assert_frame_equal(result, expected)
There was a problem hiding this comment.
this is great. I was struggling with IndexType
|
Thanks @salem3358 for the PR. Can you add a release note to 1.1.0. even if this doesn't make the RC, since this fixes a regression we will be backporting if necessary. cc @TomAugspurger |
|
Added a releases note. Merging on green. |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diff