Skip to content

DOC/TYP: index.take return val #40521

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 15 commits into from
May 23, 2021
Merged

DOC/TYP: index.take return val #40521

Changes from 4 commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
5253d39
DOC/TYP: index.take
mzeitlin11 Mar 19, 2021
67b6299
Fix default
mzeitlin11 Mar 19, 2021
b609767
Update pandas/core/indexes/base.py
mzeitlin11 Mar 22, 2021
8b917de
Use AnyArrayLike instead of ArrayLike
mzeitlin11 Mar 22, 2021
eb6016a
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Mar 22, 2021
2ff3ea5
Merge branch 'doc/index_take_meth' of github.com:/mzeitlin11/pandas i…
mzeitlin11 Mar 22, 2021
a53cb32
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Mar 31, 2021
d7ec12c
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Apr 2, 2021
c37be8c
Remove complicated typing, doc fixes
mzeitlin11 Apr 2, 2021
b23c43d
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Apr 6, 2021
a16b108
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Apr 8, 2021
74ca93a
Merge remote-tracking branch 'origin/master' into doc/index_take_meth
mzeitlin11 Apr 12, 2021
9c34ff3
Merge remote-tracking branch 'origin' into doc/index_take_meth
mzeitlin11 Apr 23, 2021
abd8e7c
Merge remote-tracking branch 'upstream/master' into doc/index_take_meth
mzeitlin11 Apr 30, 2021
a14be0b
Merge remote-tracking branch 'upstream/master' into doc/index_take_meth
mzeitlin11 May 11, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 17 additions & 7 deletions pandas/core/indexes/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -886,7 +886,7 @@ def astype(self, dtype, copy=True):

Parameters
----------
indices : list
indices : array-like
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i think Sequence[int] would be more helpful

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Happy to change, my only question would be if we want to be using type hints for user facing API. The numpy documentation for take just uses array_like

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i think Sequence[int] would be more helpful

we shouldn't use typing notation for docstrings. https://pandas.pydata.org/pandas-docs/dev/development/contributing_docstring.html#parameter-types

and from that guide...

If the exact type is not relevant, but must be compatible with a NumPy array, array-like can be specified. If Any type that can be iterated is accepted, iterable can be used:

but the guide doesn't mention sequence.

maybe sequence of int is allowed, but I think matching numpy is fine.

Indices to be taken.
axis : int, optional
The axis over which to select values, always 0.
Expand All @@ -897,8 +897,8 @@ def astype(self, dtype, copy=True):

Returns
-------
numpy.ndarray
Elements of given indices.
Index
An index formed of elements at the given indices.

See Also
--------
Expand All @@ -907,16 +907,26 @@ def astype(self, dtype, copy=True):
"""

@Appender(_index_shared_docs["take"] % _index_doc_kwargs)
def take(self, indices, axis=0, allow_fill=True, fill_value=None, **kwargs):
def take(
self,
indices: Union[AnyArrayLike, Sequence[int]],
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

isnt Sequence[int] going to cover the AnyArrayLike case?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think even with AnyArrayLike in place of ArrayLike we run into the issue mentioned here #40521 (comment)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is too complicated and specific an annotation to put here, leave this untyped for now, or put it in a typing alias, but then we should enforce it, alt we shouldn't allow non-array-likes here

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Left untyped for now, now mostly just a doc fix

axis: int = 0,
allow_fill: bool = True,
fill_value=None,
**kwargs,
):
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Any advice on typing this return value? Can almost use the TypeVar _IndexT, but the problem is that RangeIndex.take returns a superclass (Int64Index).

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think in the base class use Index, and can then use a typevar in the subclasses that do return the same type.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I tried something along those lines originally - the problem I was running into is how to handle subclasses which rely on parent implementations. For example, for ExtensionIndex would like to be able to write a stub like

def take(
        self: _ExtensionIndexT,
        indices,
        axis: int = 0,
        allow_fill: bool = True,
        fill_value=None,
        **kwargs,
    ) -> _ExtensionIndexT: ...

to essentially type the child while still relying on the inherited implementation. But couldn't figure out a way to achieve this.

if kwargs:
nv.validate_take((), kwargs)
indices = ensure_platform_int(indices)
allow_fill = self._maybe_disallow_fill(allow_fill, fill_value, indices)
indices_as_array = ensure_platform_int(indices)
allow_fill = self._maybe_disallow_fill(allow_fill, fill_value, indices_as_array)

# Note: we discard fill_value and use self._na_value, only relevant
# in the case where allow_fill is True and fill_value is not None
taken = algos.take(
self._values, indices, allow_fill=allow_fill, fill_value=self._na_value
self._values,
indices_as_array,
allow_fill=allow_fill,
fill_value=self._na_value,
)
return type(self)._simple_new(taken, name=self.name)

Expand Down