Crate df_derive_core 

Shared runtime trait identity for df-derive.

What this crate provides

df-derive-core owns the default dataframe traits used by the user-facing df-derive facade. Sharing these traits across crates lets derived models compose as nested ToDataFrame types without each crate inventing a local runtime identity.

The dataframe module exposes:

• dataframe::ToDataFrame — the per-instance API the derive populates.
• dataframe::Columnar — the columnar batch API the derive populates.
• dataframe::ToDataFrameVec — the slice extension trait that routes [T]::to_dataframe() through Columnar or empty_dataframe.
• dataframe::Decimal128Encode — the contract for encoding a decimal value as an i128 mantissa rescaled to a target scale. The reference rust_decimal::Decimal impl is gated behind the rust_decimal feature (enabled by default).
• impl ToDataFrame for () and impl Columnar for () — the zero-column payload behavior used by generic Wrapper<()> shapes.

When to use this crate

Most users get this crate through df-derive. Depend on this crate directly when you want the shared traits without the facade, or when you use df-derive-macros directly and still want the default runtime identity.

[dependencies]
df-derive-core = "0.3"
df-derive-macros = "0.3"
polars = "0.53"

Default-runtime generated code uses hidden dependency re-exports from this crate, so direct polars-arrow dependencies are not required unless you use a custom runtime.

use df_derive_core::dataframe::{ToDataFrame as _, ToDataFrameVec as _};
use df_derive_macros::ToDataFrame;

#[derive(ToDataFrame)]
struct Trade { symbol: String, price: f64, size: u64 }

Validating a custom decimal backend

The Decimal128Encode contract requires round-half-to-even (banker’s rounding) on scale-down. The reference rust_decimal::Decimal impl in this crate honours that contract and is checked against Polars’ decimal string-cast behavior in this crate’s integration tests.

Modules

dataframe