Skip to content

BUG: Series(category).replace to maintain order of categories #51057

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 2 commits into from
Feb 1, 2023
Merged

BUG: Series(category).replace to maintain order of categories #51057

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2fb2fdb
BUG: Series(category).replace to maintain order of categories
lukemanley Jan 29, 2023
c12d40b
fix
lukemanley Jan 30, 2023
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
11 changes: 10 additions & 1 deletion pandas/core/arrays/categorical.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2284,7 +2284,16 @@ def _replace(self, *, to_replace, value, inplace: bool = False):
ser = ser.replace(to_replace=to_replace, value=value)

all_values = Index(ser)
new_categories = Index(ser.drop_duplicates(keep="first"))

# GH51016: maintain order of existing categories
idxr = cat.categories.get_indexer_for(all_values)
Copy link
Member

Choose a reason for hiding this comment

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

Do we need to do all this if ordered=False?

Copy link
Member Author

Choose a reason for hiding this comment

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

1.5x maintains the existing order of categories regardless of ordered so I'd lean towards keeping that behavior. The new ordering logic is barely measurable perf-wise relative to the whole replace operation.

locs = np.arange(len(ser))
locs = np.where(idxr == -1, locs, idxr)
locs = locs.argsort()

new_categories = ser.take(locs)
new_categories = new_categories.drop_duplicates(keep="first")
new_categories = Index(new_categories)
new_codes = recode_for_categories(
cat._codes, all_values, new_categories, copy=False
)
Expand Down
10 changes: 10 additions & 0 deletions pandas/tests/arrays/categorical/test_replace.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,3 +78,13 @@ def test_replace_categorical_ea_dtype():
result = pd.Series(cat).replace(["a", "b"], ["c", pd.NA])._values
expected = Categorical(pd.array(["c", pd.NA], dtype="string"))
tm.assert_categorical_equal(result, expected)


def test_replace_maintain_ordering():
# GH51016
dtype = pd.CategoricalDtype([0, 1, 2], ordered=True)
ser = pd.Series([0, 1, 2], dtype=dtype)
result = ser.replace(0, 2)
Copy link
Member

Choose a reason for hiding this comment

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

Do we have a test where a missing value is replaced?

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 don't see any. Series(category).replace doesn't actually work for replacing missing values - see below. I checked 1.5 and it doesn't replace in 1.5 either.

In [13]: import pandas as pd

In [14]: import numpy as np

In [15]: dtype = pd.CategoricalDtype(["A", "B"])

In [16]: ser = pd.Series(["A", None], dtype=dtype)

In [17]: ser
Out[17]: 
0      A
1    NaN
dtype: category
Categories (2, object): ['A', 'B']

In [18]: ser.replace(np.nan, "C")           # <- this doesn't actually replace the missing value
Out[18]: 
0      A
1    NaN
dtype: category
Categories (2, object): ['A', 'B']

expected_dtype = pd.CategoricalDtype([1, 2], ordered=True)
expected = pd.Series([2, 1, 2], dtype=expected_dtype)
tm.assert_series_equal(expected, result, check_category_order=True)