Skip to content
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

Jump to bottom

Support at/iat indexers in cudf.pandas #16177

Merged
merged 8 commits into from
Jul 8, 2024
Merged

Support at/iat indexers in cudf.pandas #16177

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
7edab79
Support at/iat indexers in cudf.pandas
mroeschke Jul 2, 2024
41be22e
Remove redundant
mroeschke Jul 2, 2024
b88bd32
Merge remote-tracking branch 'upstream/branch-24.08' into bug/cdfp/at
mroeschke Jul 3, 2024
a43454c
Add test for at.__setitem__ with empty
mroeschke Jul 3, 2024
8293be7
Merge remote-tracking branch 'upstream/branch-24.08' into bug/cdfp/at
mroeschke Jul 5, 2024
97a283c
Make alias for at/iat
mroeschke Jul 5, 2024
c444a64
Merge remote-tracking branch 'upstream/branch-24.08' into bug/cdfp/at
mroeschke Jul 5, 2024
420294a
Use inheritance and be explicit with types
mroeschke Jul 5, 2024
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
12 changes: 10 additions & 2 deletions python/cudf/cudf/core/dataframe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -462,6 +462,10 @@ def _setitem_tuple_arg(self, key, value):
self._frame[col].loc[key[0]] = value[i]


class _DataFrameAtIndexer(_DataFrameLocIndexer):
pass


class _DataFrameIlocIndexer(_DataFrameIndexer):
"""
For selection by index.
Expand Down Expand Up @@ -584,6 +588,10 @@ def _setitem_tuple_arg(self, key, value):
self._frame[col].iloc[key[0]] = value[i]


class _DataFrameiAtIndexer(_DataFrameIlocIndexer):
pass


class DataFrame(IndexedFrame, Serializable, GetAttrGetItemMixin):
"""
A GPU Dataframe object.
Expand Down Expand Up @@ -2581,14 +2589,14 @@ def iat(self):
"""
Alias for ``DataFrame.iloc``; provided for compatibility with Pandas.
"""
return self.iloc
return _DataFrameiAtIndexer(self)

@property
def at(self):
"""
Alias for ``DataFrame.loc``; provided for compatibility with Pandas.
"""
return self.loc
return _DataFrameAtIndexer(self)

@property # type: ignore
@_external_only_api(
Expand Down
12 changes: 12 additions & 0 deletions python/cudf/cudf/pandas/_wrappers/pandas.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -775,6 +775,18 @@ def Index__new__(cls, *args, **kwargs):
pd.core.indexing._LocIndexer,
)

_AtIndexer = make_intermediate_proxy_type(
"_AtIndexer",
cudf.core.dataframe._DataFrameAtIndexer,
pd.core.indexing._AtIndexer,
)

_iAtIndexer = make_intermediate_proxy_type(
"_iAtIndexer",
cudf.core.dataframe._DataFrameiAtIndexer,
pd.core.indexing._iAtIndexer,
)

FixedForwardWindowIndexer = make_final_proxy_type(
"FixedForwardWindowIndexer",
_Unusable,
Expand Down
19 changes: 19 additions & 0 deletions python/cudf/cudf_pandas_tests/test_cudf_pandas.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1566,3 +1566,22 @@ def test_arrow_string_arrays():
)

tm.assert_equal(cu_arr, pd_arr)


@pytest.mark.parametrize("indexer", ["at", "iat"])
def test_at_iat(indexer):
Copy link
Contributor

Choose a reason for hiding this comment

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

Does this cover the case of a new column that doesn't exist? I'd like to see a test that more clearly resembles the patterns in the minimal repro of #16112.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yup, added a unit test in a43454c to mirror the issue. (It appears that iat.__setitem__ in pandas doesn't support the type of behavior demonstrated in the issue)

Copy link
Contributor Author

@mroeschke mroeschke Jul 3, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

Huh, thanks for encouraging a unit test with the issue. Even though the test passed locally the CI didn't agree.

Left some investigation notes in #16112 (comment) that I may come back to later

df = xpd.DataFrame(range(3))
result = getattr(df, indexer)[0, 0]
assert result == 0

getattr(df, indexer)[0, 0] = 1
expected = pd.DataFrame([1, 1, 2])
tm.assert_frame_equal(df, expected)


def test_at_setitem_empty():
df = xpd.DataFrame({"name": []}, dtype="float64")
df.at[0, "name"] = 1.0
df.at[0, "new"] = 2.0
expected = pd.DataFrame({"name": [1.0], "new": [2.0]})
tm.assert_frame_equal(df, expected)
Loading