DEPR: Warn on empty Series dtype inference #20802
Conversation
FutureWarning from float to object
TomAugspurger
added
Dtype Conversions
|
FYI, d80e1c7 has the changes. Everything else is catching / fixing warning internal to pandas. This ended up being a decent-sized change. |
| @@ -862,6 +862,7 @@ Other API Changes | |||
| Deprecations | |||
| ~~~~~~~~~~~~ | |||
|
|
|||
| - The inferred dtype for an empty Series will change from ``float`` to ``object`` (:issue:`17261`). | |||
There was a problem hiding this comment.
double-backticks on Series
| @@ -954,7 +954,11 @@ def _map_values(self, mapper, na_action=None): | |||
| # we specify the keys here to handle the | |||
| # possibility that they are tuples | |||
| from pandas import Series | |||
| mapper = Series(mapper) | |||
| if len(mapper) == 0: | |||
There was a problem hiding this comment.
I think this might be worthwhile to make a function in pandas.core.dtypes.cast.maybe_infer_1d_array_dtype
| @@ -5399,7 +5399,11 @@ def fillna(self, value=None, method=None, axis=None, inplace=False, | |||
| if self.ndim == 1: | |||
| if isinstance(value, (dict, ABCSeries)): | |||
| from pandas import Series | |||
| value = Series(value) | |||
| if len(value) == 0: | |||
| series_dtype = 'float' | |||
| name(s) to set | ||
| level : int, level name, or sequence of int/level names (default None) | ||
| If the index is a MultiIndex (hierarchical), level(s) to set (None | ||
| for all levels). Otherwise level must be None |
There was a problem hiding this comment.
looks like you have some extra commits?
| @@ -388,10 +388,14 @@ def test_map(self): | |||
|
|
|||
| @pytest.mark.parametrize("index", tm.all_index_generator(10)) | |||
| def test_map_empty(self, index): | |||
| s = Series(index) | |||
| if len(index) == 0: | |||
|
Sorry about the empty commits, will fix that up. Still going through and squashing warnings. There are a few cases internally where we construct an empty series. e.g. In [1]: import pandas as pd
In [2]: pd.Series([], dtype='object').groupby([]).apply(lambda x: 1)
Out[2]: Series([], dtype: float64)How should we handle that?
The problem w/ 2 is that there's nothing the user can do to silence the warning. |
|
@TomAugspurger can you close or update |
|
@jreback @jorisvandenbossche do you think we want to change this before 1.0? |
|
@TomAugspurger can you merge master |
|
There's a decent amount of work to do here to avoid the warnings
internally. Will try to make time.
…On Sun, Nov 18, 2018 at 4:07 PM Jeff Reback ***@***.***> wrote:
@TomAugspurger <https://github.com/TomAugspurger> can you merge master
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#20802 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ABQHIlxE1_nFXT33ZyVrGy5GKVNKLTv7ks5uwdopgaJpZM4Tgluf>
.
|
|
@TomAugspurger keep this one open? |
|
Probably not :/ I suppose we can just live with this. A little artifact of the time Series used to be an ndarray. |
We're changing
Series([])to beobjectinstead offloat. This does the warninggit diff upstream/master -u -- "*.py" | flake8 --diff