Struct JITDataFrame 

pub struct JITDataFrame {
    pub columns: Vec<Vec<f64>>,
    pub column_ptrs: Vec<*const f64>,
    pub timestamps: Vec<i64>,
    pub row_count: usize,
}

Generic DataFrame storage for JIT execution. Stores data as an array of columns, matching the generic column_ptrs design in JITContext.

Column order MUST match the DataFrameSchema used during compilation.

Fields

columns: Vec<Vec<f64>>

Column data arrays (each Vec is one column) Columns are ordered by index as defined in DataFrameSchema

column_ptrs: Vec<*const f64>

Pointers to column data (for JITContext.column_ptrs)

timestamps: Vec<i64>

Timestamps (always present, column 0 equivalent)

row_count: usize

Number of rows

Implementations

impl JITDataFrame

pub fn new() -> Self

Create an empty JITDataFrame

pub fn from_execution_context( ctx: &ExecutionContext, schema: &DataFrameSchema, ) -> Self

Create from ExecutionContext using a schema mapping. The schema determines which columns to extract and their order.

pub fn populate_context(&self, ctx: &mut JITContext)

Populate a JITContext with generic DataFrame pointers. This sets column_ptrs, column_count, row_count, and timestamps_ptr.

pub fn len(&self) -> usize

Get the number of rows

pub fn is_empty(&self) -> bool

Check if empty

pub fn column_count(&self) -> usize

Get number of columns

pub fn from_datatable(dt: &DataTable) -> Self

Create from a DataTable by extracting f64 columns and an optional timestamp column.

All f64 columns are copied into SIMD-aligned buffers. If a column named “timestamp” (or typed as Timestamp) exists, it is extracted as i64.

Trait Implementations

impl Default for JITDataFrame

fn default() -> Self

Returns the “default value” for a type.