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

[REVIEW] Preserve the correct ListDtype while creating an identical empty column #10151

Merged
merged 3 commits into from
Jan 31, 2022
Merged

[REVIEW] Preserve the correct ListDtype while creating an identical empty column #10151

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
02f87a4
handle list dtype in column_empty
galipremsagar Jan 27, 2022
78d763a
add test
galipremsagar Jan 27, 2022
10bc21d
copyright
galipremsagar Jan 27, 2022
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
6 changes: 6 additions & 0 deletions python/cudf/cudf/core/column/column.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1316,6 +1316,12 @@ def column_empty(
column_empty(row_count, field_dtype)
for field_dtype in dtype.fields.values()
)
elif is_list_dtype(dtype):
data = None
children = (
full(row_count + 1, 0, dtype="int32"),
column_empty(row_count, dtype=dtype.element_type),
)
elif is_categorical_dtype(dtype):
data = None
children = (
Expand Down
13 changes: 12 additions & 1 deletion python/cudf/cudf/tests/test_list.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
# Copyright (c) 2020-2021, NVIDIA CORPORATION.
# Copyright (c) 2020-2022, NVIDIA CORPORATION.

import functools
import operator

Expand Down Expand Up @@ -586,3 +587,13 @@ def test_listcol_setitem_error_cases(data, item, error):
sr = cudf.Series(data)
with pytest.raises(BaseException, match=error):
sr[1] = item


def test_listcol_setitem_retain_dtype():
df = cudf.DataFrame(
{"a": cudf.Series([["a", "b"], []]), "b": [1, 2], "c": [123, 321]}
)
df1 = df[df.b.isna()]
df1["b"] = df1["c"]
Copy link
Contributor

Choose a reason for hiding this comment

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

This needs a comment to explain the logic of the test, and particularly why this line is necessary. On the surface, this line appears to do nothing (it assigns an empty Series "b" to have the data of another empty Series "c", but both are int64 types). In the failing case, this triggers a change in the dtype of df1["a"] because it is empty and goes from ListDtype(object) (list of strings) to ListDtype(int8). That side-effect is not clear because this line doesn't reference "a" at all.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Addressed these comments in https://github.com/rapidsai/cudf/pull/10176/files

df2 = df1.drop(["c"], axis=1)
assert df2.a.dtype == df.a.dtype
Copy link
Contributor

Choose a reason for hiding this comment

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

I would prefer to use the same square-bracket indexing here as above, rather than switch to attribute-based access:

Suggested change
assert df2.a.dtype == df.a.dtype
assert df2["a"].dtype == df["a"].dtype