Trait IntoDataFrameSplit

Source

pub trait IntoDataFrameSplit: Sized {
    // Required method
    fn into_dataframe_split(self) -> DataFrameShape;
}

Expand description

Enum rows → variant-partitioned R data.frames. The split-representation sibling of IntoDataFrame.

Where IntoDataFrame produces one aligned data.frame (field-name union, NA fill for columns a variant doesn’t carry), this verb partitions the rows by variant: each partition is a data.frame with only that variant’s own columns (non-optional types — no NA fill from sibling variants), keyed by the snake_case variant name.

Implemented by #[derive(DataFrameRow)] on an enum row type (for Vec<Row>, via the hidden DataFrameRowSplit bridge below). Struct rows have no variants to partition, so the struct derive does not provide it. Call it on your data: rows.into_dataframe_split().

§Return shape

The returned DataFrameShape says which R value the split produced:

DataFrameShape::Bare — a single-variant enum: one data.frame.
DataFrameShape::PerVariantList — a multi-variant enum: one (snake_case_name, data.frame) entry per declared variant, in declaration order, which converts to a named R list of data.frames.

The derive never produces DataFrameShape::Split (that variant belongs to the serde result_to_dataframe helper). Variants absent from the input still appear as 0-row data.frames with that variant’s column shape; unit variants produce a 0-column data.frame with the correct row count. DataFrameShape is IntoR, so a #[miniextendr] function returns it directly.

§GC discipline (rooted by construction, #1750)

Every partition is an owned, GC-rooted BuiltDataFrame, so the result can be held across R allocations (building another frame, calling back into R) before it is returned or converted; there is no convert-immediately contract. The IntoR conversion assembles the named list and releases the per-partition roots.