Struct datafusion_physical_plan::coalesce_batches::CoalesceBatchesExec

pub struct CoalesceBatchesExec { /* private fields */ }

CoalesceBatchesExec combines small batches into larger batches for more efficient use of vectorized processing by later operators.

The operator buffers batches until it collects target_batch_size rows and then emits a single concatenated batch. When only a limited number of rows are necessary (specified by the fetch parameter), the operator will stop buffering and returns the final batch once the number of collected rows reaches the fetch value.

Background

Generally speaking, larger RecordBatches are more efficient to process than smaller record batches (until the CPU cache is exceeded) because there is fixed processing overhead per batch. This code concatenates multiple small record batches into larger ones to amortize this overhead.

┌────────────────────┐
│    RecordBatch     │
│   num_rows = 23    │
└────────────────────┘                 ┌────────────────────┐
                                       │                    │
┌────────────────────┐     Coalesce    │                    │
│                    │      Batches    │                    │
│    RecordBatch     │                 │                    │
│   num_rows = 50    │  ─ ─ ─ ─ ─ ─ ▶  │                    │
│                    │                 │    RecordBatch     │
│                    │                 │   num_rows = 106   │
└────────────────────┘                 │                    │
                                       │                    │
┌────────────────────┐                 │                    │
│                    │                 │                    │
│    RecordBatch     │                 │                    │
│   num_rows = 33    │                 └────────────────────┘
│                    │
└────────────────────┘

Implementations

impl CoalesceBatchesExec

pub fn new(input: Arc<dyn ExecutionPlan>, target_batch_size: usize) -> Self

Create a new CoalesceBatchesExec

pub fn with_fetch(self, fetch: Option<usize>) -> Self

Update fetch with the argument

pub fn input(&self) -> &Arc<dyn ExecutionPlan>

The input plan

pub fn target_batch_size(&self) -> usize

Minimum number of rows for coalesces batches

Trait Implementations

impl Debug for CoalesceBatchesExec

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

Formats the value using the given formatter.

impl DisplayAs for CoalesceBatchesExec

fn fmt_as(&self, t: DisplayFormatType, f: &mut Formatter<'_>) -> Result

Format according to DisplayFormatType, used when verbose representation looks different from the default one

impl ExecutionPlan for CoalesceBatchesExec

fn as_any(&self) -> &dyn Any

Return a reference to Any that can be used for downcasting

fn name(&self) -> &'static str

Short name for the ExecutionPlan, such as ‘ParquetExec’.

fn properties(&self) -> &PlanProperties

Return properties of the output of the ExecutionPlan, such as output ordering(s), partitioning information etc.

fn children(&self) -> Vec<&Arc<dyn ExecutionPlan>>

Get a list of children ExecutionPlans that act as inputs to this plan. The returned list will be empty for leaf nodes such as scans, will contain a single value for unary nodes, or two values for binary nodes (such as joins).

fn maintains_input_order(&self) -> Vec<bool>

Returns false if this ExecutionPlan’s implementation may reorder rows within or between partitions.

fn benefits_from_input_partitioning(&self) -> Vec<bool>

Specifies whether the ExecutionPlan benefits from increased parallelization at its input for each child.

fn with_new_children( self: Arc<Self>, children: Vec<Arc<dyn ExecutionPlan>>, ) -> Result<Arc<dyn ExecutionPlan>>

Returns a new ExecutionPlan where all existing children were replaced by the children, in order

fn execute( &self, partition: usize, context: Arc<TaskContext>, ) -> Result<SendableRecordBatchStream>

Begin execution of partition, returning a Stream of RecordBatches.

fn metrics(&self) -> Option<MetricsSet>

Return a snapshot of the set of Metrics for this ExecutionPlan. If no Metrics are available, return None.

fn statistics(&self) -> Result<Statistics>

Returns statistics for this ExecutionPlan node. If statistics are not available, should return Statistics::new_unknown (the default), not an error.

fn with_fetch(&self, limit: Option<usize>) -> Option<Arc<dyn ExecutionPlan>>

Returns a fetching variant of this ExecutionPlan node, if it supports fetch limits. Returns None otherwise.

fn static_name() -> &'static str

where Self: Sized,

Short name for the ExecutionPlan, such as ‘ParquetExec’. Like name but can be called without an instance.

fn schema(&self) -> SchemaRef

Get the schema for this execution plan

fn required_input_distribution(&self) -> Vec<Distribution>

Specifies the data distribution requirements for all the children for this ExecutionPlan, By default it’s [Distribution::UnspecifiedDistribution] for each child,

fn required_input_ordering(&self) -> Vec<Option<Vec<PhysicalSortRequirement>>>

Specifies the ordering required for all of the children of this ExecutionPlan.

fn repartitioned( &self, _target_partitions: usize, _config: &ConfigOptions, ) -> Result<Option<Arc<dyn ExecutionPlan>>>

If supported, attempt to increase the partitioning of this ExecutionPlan to produce target_partitions partitions.

fn supports_limit_pushdown(&self) -> bool

Returns true if a limit can be safely pushed down through this ExecutionPlan node.