Skip to content

[Python][Interchange] from_dataframe drops the categorical ordered flag #51046

Open

Description

@Kayvan-Zahiri

Describe the bug, including details regarding any error messages, version, and platform.

pyarrow.interchange.from_dataframe builds the dictionary array without the is_ordered flag the producer reports, so an ordered categorical column comes back unordered. A pyarrow table whose column is dictionary(int32, string, ordered=True) does not survive its own round trip, and an ordered pandas Categorical loses its ordering on the way in.

import pyarrow as pa
from pyarrow.interchange import from_dataframe

t = pa.table({"c": pa.array(["a", "b", "a"]).dictionary_encode().cast(
    pa.dictionary(pa.int32(), pa.string(), ordered=True))})
print(t.schema.field("c").type.ordered)                    # True
print(from_dataframe(t).schema.field("c").type.ordered)    # False

The interchange protocol exposes the flag: a categorical column's describe_categorical returns is_ordered, and the pandas consumer passes it through. The pyarrow consumer reads the mapping but drops is_ordered.

Component(s)

Python

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions