Fixed #38419 - BUG: set_index screws up the dtypes on empty DataFrames #38430
Conversation
|
Hello @jordi-crespo! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found: There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻 Comment last updated at 2020-12-13 12:05:16 UTC |
| {"a": Series(dtype="datetime64[ns]"), "b": Series(dtype="int64"), "c": []} | ||
| ) | ||
|
|
||
| if df.empty: |
There was a problem hiding this comment.
we don't need this. There's only one frame and we know it's empty.
There was a problem hiding this comment.
Although feel free to add
assert df.empty is True| df = DataFrame(columns=list(df.columns.values)) | ||
|
|
||
| expected = df.set_index(["a", "b"]) | ||
| assert (df.loc[:, ["a", "b"]].dtypes == expected.index.to_frame().dtypes).all() |
There was a problem hiding this comment.
Sorry but absolutely not, this it too complicated
Construct expected ahead of time and just do
tm.assert_frame_equal(result, expected)
There was a problem hiding this comment.
Make the construction as explicit as possible. The point here is to make it as easy as possible to see what the test does
There was a problem hiding this comment.
You want to do something like this
In [21]: import pandas as pd
...: import pandas._testing as tm
...:
...: df1 = pd.DataFrame({'a': pd.Series(dtype='datetime64[ns]'), 'b': pd.Series(dtype='int64'), 'c': []})
...: df2= df1.set_index(['a', 'b'])
...:
...: result = df2.dtypes
...: expected = df1[['c']].dtypes # double square brackets to slice to a DataFrame not to a Series
...: tm.assert_series_equal(result, expected)There was a problem hiding this comment.
Please check
jreback
added
Index
|
thanks @jordi-crespo |
Fixes #38419 - BUG: set_index screws up the dtypes on empty DataFrames