pola-rs/polars · error

ordering for Object dtype is not supported

Error message

ordering for Object dtype is not supported

What it means

AnyValue::partial_cmp cannot order Object scalars. Object values are opaque user-defined types registered via the PolarsObject mechanism; polars has no ordering rule for them unless the user supplies one, so the arm is unimplemented.

Source

Thrown at crates/polars-core/src/datatypes/any_value.rs:1449

                l_str.partial_cmp(r_str)
            },
            #[cfg(feature = "dtype-categorical")]
            (Enum(l_cat, l_map), Enum(r_cat, r_map)) => {
                if !Arc::ptr_eq(l_map, r_map) {
                    unimplemented!("can't order enums from different FrozenCategories")
                }
                l_cat.partial_cmp(r_cat)
            },
            (List(_), List(_)) => {
                unimplemented!("ordering for List dtype is not supported")
            },
            #[cfg(feature = "dtype-array")]
            (Array(..), Array(..)) => {
                unimplemented!("ordering for Array dtype is not supported")
            },
            #[cfg(feature = "object")]
            (Object(_), Object(_)) => {
                unimplemented!("ordering for Object dtype is not supported")
            },
            #[cfg(feature = "dtype-struct")]
            (StructOwned(_), StructOwned(_))
            | (StructOwned(_), Struct(..))
            | (Struct(..), StructOwned(_))
            | (Struct(..), Struct(..)) => {
                unimplemented!("ordering for Struct dtype is not supported")
            },
            #[cfg(feature = "dtype-decimal")]
            (Decimal(lv, _lp, ls), Decimal(rv, _rp, rs)) => Some(dec128_cmp(*lv, *ls, *rv, *rs)),

            (_, _) => {
                unimplemented!(
                    "scalar ordering for mixed dtypes {self:?} and {other:?} is not supported"
                )
            },
        }
    }

View on GitHub (pinned to df599052da)

Solutions

1. Store a comparable key alongside the object (a separate Int/Float/String column) and sort by that
2. Convert the object column to a native dtype before ordering
3. In Rust, order by downcasting the objects yourself in user code rather than via AnyValue comparisons

Example fix

# before
out = sorted(df["obj"].to_list())  # comparing objects -> panic

# after
out = df.sort("obj_sort_key")["obj"].to_list()  # order by a native key column

def orderable_dtype(dt) -> bool:
    return not isinstance(dt, (pl.List, pl.Array, pl.Struct)) and dt != pl.Object

def orderable_dtype(dt) -> bool:
    return not isinstance(dt, (pl.List, pl.Array, pl.Struct)) and dt != pl.Object

When it happens

Trigger: Comparing, sorting, or ranking scalars from an Object column (registered custom Rust type): df["obj"].get(0) < df["obj"].get(1).

Common situations: Custom Rust types stored in Object columns (e.g. geometries, decimals wrappers, intervals) that later flow into generic sort/min/max code paths.

Related errors

• comparing datetimes with different units or timezones is not
• comparing durations with different units is not supported
• can't order enums from different FrozenCategories
• ordering for List dtype is not supported
• ordering for Array dtype is not supported

AI-assisted analysis of pola-rs/polars@df599052da (2026-08-16). Data as JSON: /api/errors/ef78b4ff90991930. Report an issue: GitHub.