Struct DefaultPhysicalPlanner

pub struct DefaultPhysicalPlanner { /* private fields */ }

Default single node physical query planner that converts a LogicalPlan to an ExecutionPlan suitable for execution.

This planner will first flatten the LogicalPlan tree via a depth first approach, which allows it to identify the leaves of the tree.

Tasks are spawned from these leaves and traverse back up the tree towards the root, converting each LogicalPlan node it reaches into their equivalent ExecutionPlan node. When these tasks reach a common node, they will terminate until the last task reaches the node which will then continue building up the tree.

Up to planning_concurrency tasks are buffered at once to execute concurrently.

Implementations

impl DefaultPhysicalPlanner

pub fn with_extension_planners( extension_planners: Vec<Arc<dyn ExtensionPlanner + Send + Sync>>, ) -> Self

Create a physical planner that uses extension_planners to plan user-defined logical nodes LogicalPlan::Extension. The planner uses the first ExtensionPlanner to return a non-None plan.

impl DefaultPhysicalPlanner

pub fn optimize_physical_plan<F>( &self, plan: Arc<dyn ExecutionPlan>, session_state: &SessionState, observer: F, ) -> Result<Arc<dyn ExecutionPlan>>

where F: FnMut(&dyn ExecutionPlan, &dyn PhysicalOptimizerRule),

Optimize a physical plan by applying each physical optimizer, calling observer(plan, optimizer after each one)

Trait Implementations

impl Default for DefaultPhysicalPlanner

fn default() -> DefaultPhysicalPlanner

Returns the “default value” for a type.

impl PhysicalPlanner for DefaultPhysicalPlanner

fn create_physical_plan<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, logical_plan: &'life1 LogicalPlan, session_state: &'life2 SessionState, ) -> Pin<Box<dyn Future<Output = Result<Arc<dyn ExecutionPlan>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Create a physical plan from a logical plan

fn create_physical_expr( &self, expr: &Expr, input_dfschema: &DFSchema, session_state: &SessionState, ) -> Result<Arc<dyn PhysicalExpr>>

Create a physical expression from a logical expression suitable for evaluation

e: the expression to convert

input_dfschema: the logical plan schema for evaluating e