Struct DataChunk 

pub struct DataChunk { /* private fields */ }

A batch of rows stored column-wise for vectorized processing.

Instead of storing rows like [(a1,b1), (a2,b2), ...], we store columns like [a1,a2,...], [b1,b2,...]. This is cache-friendly for analytical queries that touch few columns but many rows.

The optional SelectionVector lets you filter rows without copying data - just mark which row indices are “selected”.

Example

use grafeo_core::execution::DataChunk;
use grafeo_core::execution::ValueVector;
use grafeo_common::types::Value;

// Create columns
let names = ValueVector::from_values(&[Value::from("Alix"), Value::from("Gus")]);
let ages = ValueVector::from_values(&[Value::from(30i64), Value::from(25i64)]);

// Bundle into a chunk
let chunk = DataChunk::new(vec![names, ages]);
assert_eq!(chunk.len(), 2);

Implementations

impl DataChunk

pub fn empty() -> Self

Creates an empty data chunk with no columns.

pub fn new(columns: Vec<ValueVector>) -> Self

Creates a new data chunk from existing vectors.

pub fn with_schema(column_types: &[LogicalType]) -> Self

Creates a new empty data chunk with the given schema.

pub fn with_capacity(column_types: &[LogicalType], capacity: usize) -> Self

Creates a new data chunk with the given schema and capacity.

pub fn column_count(&self) -> usize

Returns the number of columns.

pub fn row_count(&self) -> usize

Returns the number of rows (considering selection).

pub fn len(&self) -> usize

Alias for row_count().

pub fn columns(&self) -> &[ValueVector]

Returns all columns.

pub fn total_row_count(&self) -> usize

Returns the total number of rows (ignoring selection).

pub fn is_empty(&self) -> bool

Returns true if the chunk is empty.

pub fn capacity(&self) -> usize

Returns the capacity of this chunk.

pub fn is_full(&self) -> bool

Returns true if the chunk is full.

pub fn column(&self, index: usize) -> Option<&ValueVector>

Gets a column by index.

pub fn column_mut(&mut self, index: usize) -> Option<&mut ValueVector>

Gets a mutable column by index.

pub fn selection(&self) -> Option<&SelectionVector>

Returns the selection vector.

pub fn set_selection(&mut self, selection: SelectionVector)

Sets the selection vector.

pub fn clear_selection(&mut self)

Clears the selection vector (selects all rows).

pub fn set_zone_hints(&mut self, hints: ChunkZoneHints)

Sets zone map hints for this chunk.

Zone map hints enable the filter operator to skip entire chunks when predicates can’t possibly match based on min/max statistics.

pub fn zone_hints(&self) -> Option<&ChunkZoneHints>

Returns zone map hints if available.

Used by the filter operator for chunk-level predicate pruning.

pub fn clear_zone_hints(&mut self)

Clears zone map hints.

pub fn set_count(&mut self, count: usize)

Sets the row count.

pub fn reset(&mut self)

Resets the chunk for reuse.

pub fn flatten(&mut self)

Flattens the selection by copying only selected rows.

After this operation, selection is None and count equals the previously selected row count.

pub fn selected_indices(&self) -> Box<dyn Iterator<Item = usize> + '_>

Returns an iterator over selected row indices.

pub fn concat(chunks: &[DataChunk]) -> DataChunk

Concatenates multiple chunks into a single chunk.

All chunks must have the same schema (same number and types of columns).

pub fn filter(&self, predicate: &SelectionVector) -> DataChunk

Applies a filter predicate and returns a new chunk with selected rows.

pub fn slice(&self, offset: usize, count: usize) -> DataChunk

Returns a slice of this chunk.

Returns a new DataChunk containing rows [offset, offset + count).

pub fn num_columns(&self) -> usize

Returns the number of columns.

Trait Implementations

impl Clone for DataChunk

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for DataChunk

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl From<DataChunk> for ChunkVariant

fn from(chunk: DataChunk) -> Self

Converts to this type from the input type.