Struct MergeSortExec 

pub(crate) struct MergeSortExec {
    inner: SortPreservingMergeExec,
}

An opaque physical execution node for MergeSortLogicalPlan.

It delegates execution and physical properties to DataFusion’s [SortPreservingMergeExec], but intentionally does not expose itself as a SortPreservingMergeExec. EnforceSorting is allowed to replace a bare SortPreservingMergeExec with CoalescePartitionsExec when the parent does not require ordering. MergeSortExec represents the distributed TopK merge stage itself, so later physical optimizer rules must not rewrite it into an unordered fetch.

Fields

inner: SortPreservingMergeExec

Implementations

impl MergeSortExec

pub(crate) fn new( ordering: LexOrdering, input: Arc<dyn ExecutionPlan>, fetch: Option<usize>, ) -> Self

fn input_with_fetch(&self, fetch: Option<usize>) -> Arc<dyn ExecutionPlan>

Trait Implementations

impl Clone for MergeSortExec

fn clone(&self) -> MergeSortExec

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MergeSortExec

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

Formats the value using the given formatter.

impl DisplayAs for MergeSortExec

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 MergeSortExec

fn as_any(&self) -> &dyn Any

Keeps this node intentionally opaque to DataFusion’s type-specialized optimizer rewrites.

MergeSortExec delegates most behavior to DataFusion’s SortPreservingMergeExec, but it must not expose itself as that type. DataFusion’s EnforceSorting optimizer recognizes a bare SortPreservingMergeExec via as_any().downcast_ref::<...>() and may replace it with an unordered CoalescePartitionsExec(fetch) when the parent does not require sorted output.

That rewrite is valid for an ordinary SPM used only to satisfy parent ordering, but not for GreptimeDB’s distributed TopK merge stage. In a scalar-subquery shape like ORDER BY ts DESC LIMIT 1, this node is the operator that merges region-local TopK streams into the global TopK. Replacing it with unordered coalescing can return a partial/latest row from one region instead of the global latest row.

required_input_ordering() separately tells DataFusion what ordering this node needs from its child, so EnforceSorting can insert a SortExec below MergeSortExec when MergeScanExec cannot preserve per-partition ordering. This opacity is specifically about protecting the merge stage itself from the EnforceSorting rewrite above.

fn with_preserve_order( &self, preserve_order: bool, ) -> Option<Arc<dyn ExecutionPlan>>

Forwards DataFusion’s order-preserving scan hint through this wrapper.

This mirrors SortPreservingMergeExec::with_preserve_order(): if the child can produce an order-preserving variant, rebuild the same merge stage on top of that child. The returned plan must stay a MergeSortExec, not a bare SPM, so the distributed TopK merge remains opaque to EnforceSorting’s SPM-specific rewrite.

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

Tells DataFusion that MergeSortExec requires each input partition to be ordered. This is the contract that makes EnforceSorting insert a SortExec below MergeSortExec when the input cannot preserve ordering.

The opacity of MergeSortExec::as_any, not this requirement, is what prevents DataFusion from rewriting the merge stage itself as a bare SortPreservingMergeExec.

fn supports_limit_pushdown(&self) -> bool

Intentionally keeps DataFusion’s generic limit pushdown disabled.

MergeSortExec still supports its own global fetch through with_fetch(). What we must not allow is pushing an external limit below this required distributed TopK merge. DataFusion’s limit pushdown rules know how to treat a bare SortPreservingMergeExec as a partition-combining node, but MergeSortExec is intentionally opaque to those SPM-specific downcasts. Enabling generic limit pushdown without also teaching the optimizer about this wrapper could return partition-local rows instead of the global TopK.

fn try_swapping_with_projection( &self, projection: &ProjectionExec, ) -> Result<Option<Arc<dyn ExecutionPlan>>>

Lets DataFusion push a projection below this merge when it can rewrite the ordering expressions safely.

This mirrors SortPreservingMergeExec::try_swapping_with_projection() for plan quality, but re-wraps the result as MergeSortExec so the distributed merge stage keeps its type identity and opacity.

fn name(&self) -> &str

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

fn properties(&self) -> &Arc<PlanProperties>

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

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 benefits_from_input_partitioning(&self) -> Vec<bool>

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

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

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

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 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 [RecordBatch]es.

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

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

fn partition_statistics(&self, partition: Option<usize>) -> Result<Statistics>

Returns statistics for a specific partition of this ExecutionPlan node. If statistics are not available, should return [Statistics::new_unknown] (the default), not an error. If partition is None, it returns statistics for the entire plan.

fn cardinality_effect(&self) -> CardinalityEffect

Gets the effect on cardinality, if known

fn fetch(&self) -> Option<usize>

Gets the fetch count for the operator, None means there is no fetch.

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 ‘DataSourceExec’. Like name but can be called without an instance.

fn schema(&self) -> Arc<Schema>

Get the schema for this execution plan

fn check_invariants(&self, check: InvariantLevel) -> Result<(), DataFusionError>

Returns an error if this individual node does not conform to its invariants. These invariants are typically only checked in debug mode.

fn reset_state( self: Arc<Self>, ) -> Result<Arc<dyn ExecutionPlan>, DataFusionError>

Reset any internal state within this [ExecutionPlan].

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

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

fn gather_filters_for_pushdown( &self, _phase: FilterPushdownPhase, parent_filters: Vec<Arc<dyn PhysicalExpr>>, _config: &ConfigOptions, ) -> Result<FilterDescription, DataFusionError>

Collect filters that this node can push down to its children. Filters that are being pushed down from parents are passed in, and the node may generate additional filters to push down. For example, given the plan FilterExec -> HashJoinExec -> DataSourceExec, what will happen is that we recurse down the plan calling ExecutionPlan::gather_filters_for_pushdown:

fn handle_child_pushdown_result( &self, _phase: FilterPushdownPhase, child_pushdown_result: ChildPushdownResult, _config: &ConfigOptions, ) -> Result<FilterPushdownPropagation<Arc<dyn ExecutionPlan>>, DataFusionError>

Handle the result of a child pushdown.

fn with_new_state( &self, _state: Arc<dyn Any + Send + Sync>, ) -> Option<Arc<dyn ExecutionPlan>>

Injects arbitrary run-time state into this execution plan, returning a new plan instance that incorporates that state if it is relevant to the concrete node implementation.

fn try_pushdown_sort( &self, _order: &[PhysicalSortExpr], ) -> Result<SortOrderPushdownResult<Arc<dyn ExecutionPlan>>, DataFusionError>

Try to push down sort ordering requirements to this node.