Struct PlanRoot

pub struct PlanRoot {
    plan: PlanRef,
    phase: PlanPhase,
    required_dist: RequiredDist,
    required_order: Order,
    out_fields: FixedBitSet,
    out_names: Vec<String>,
}

PlanRoot is used to describe a plan. planner will construct a PlanRoot with LogicalNode. and required distribution and order. And PlanRoot can generate corresponding streaming or batch plan with optimization. the required Order and Distribution columns might be more than the output columns. for example:

   select v1 from t order by id;

the plan will return two columns (id, v1), and the required order column is id. the id column is required in optimization, but the final generated plan will remove the unnecessary column in the result.

Fields

plan: PlanRef

phase: PlanPhase

required_dist: RequiredDist

required_order: Order

out_fields: FixedBitSet

out_names: Vec<String>

Implementations

impl PlanRoot

pub fn new_with_logical_plan( plan: PlanRef, required_dist: RequiredDist, required_order: Order, out_fields: FixedBitSet, out_names: Vec<String>, ) -> Self

pub fn new_with_batch_plan( plan: PlanRef, required_dist: RequiredDist, required_order: Order, out_fields: FixedBitSet, out_names: Vec<String>, ) -> Self

fn new_inner( plan: PlanRef, phase: PlanPhase, required_dist: RequiredDist, required_order: Order, out_fields: FixedBitSet, out_names: Vec<String>, ) -> Self

pub fn set_out_names(&mut self, out_names: Vec<String>) -> Result<(), RwError>

Set customized names of the output fields, used for CREATE [MATERIALIZED VIEW | SINK] r(a, b, ..).

If the number of names does not match the number of output fields, an error is returned.

pub fn schema(&self) -> Schema

Get the plan root’s schema, only including the fields to be output.

pub fn into_unordered_subplan(self) -> PlanRef

Transform the PlanRoot back to a PlanRef suitable to be used as a subplan, for example as insert source or subquery. This ignores Order but retains post-Order pruning (out_fields).

pub fn into_array_agg(self) -> Result<PlanRef, RwError>

Transform the PlanRoot wrapped in an array-construction subquery to a PlanRef supported by ARRAY_AGG. Similar to the unordered version, this abstracts away internal self.plan which is further modified by self.required_order then self.out_fields.

pub fn gen_optimized_logical_plan_for_stream( &mut self, ) -> Result<PlanRef, RwError>

Apply logical optimization to the plan for stream.

pub fn gen_optimized_logical_plan_for_batch( &mut self, ) -> Result<PlanRef, RwError>

Apply logical optimization to the plan for batch.

pub fn gen_batch_plan(&mut self) -> Result<PlanRef, RwError>

Optimize and generate a singleton batch physical plan without exchange nodes.

pub fn gen_batch_distributed_plan(self) -> Result<PlanRef, RwError>

Optimize and generate a batch query plan for distributed execution.

pub fn gen_batch_local_plan(self) -> Result<PlanRef, RwError>

Optimize and generate a batch query plan for local execution.

fn gen_optimized_stream_plan( &mut self, emit_on_window_close: bool, allow_snapshot_backfill: bool, ) -> Result<PlanRef, RwError>

Generate optimized stream plan

fn gen_optimized_stream_plan_inner( &mut self, emit_on_window_close: bool, stream_scan_type: StreamScanType, ) -> Result<PlanRef, RwError>

fn gen_stream_plan( &mut self, emit_on_window_close: bool, stream_scan_type: StreamScanType, ) -> Result<PlanRef, RwError>

Generate create index or create materialize view plan.

fn compute_cardinality(&self) -> Cardinality

Visit the plan root and compute the cardinality.

Panics if not called on a logical plan.

pub fn gen_table_plan( self, context: OptimizerContextRef, table_name: String, columns: Vec<ColumnCatalog>, definition: String, pk_column_ids: Vec<ColumnId>, row_id_index: Option<usize>, append_only: bool, on_conflict: Option<OnConflict>, with_version_column: Option<String>, watermark_descs: Vec<WatermarkDesc>, version: Option<TableVersion>, with_external_source: bool, retention_seconds: Option<NonZeroU32>, cdc_table_id: Option<String>, ) -> Result<StreamMaterialize, RwError>

Optimize and generate a create table plan.

pub fn gen_materialize_plan( self, mv_name: String, definition: String, emit_on_window_close: bool, ) -> Result<StreamMaterialize, RwError>

Optimize and generate a create materialized view plan.

pub fn gen_index_plan( self, index_name: String, definition: String, retention_seconds: Option<NonZeroU32>, ) -> Result<StreamMaterialize, RwError>

Optimize and generate a create index plan.

pub fn gen_sink_plan( &mut self, sink_name: String, definition: String, properties: WithOptionsSecResolved, emit_on_window_close: bool, db_name: String, sink_from_table_name: String, format_desc: Option<SinkFormatDesc>, without_backfill: bool, target_table: Option<Arc<TableCatalog>>, partition_info: Option<PartitionComputeInfo>, user_specified_columns: bool, ) -> Result<StreamSink, RwError>

Optimize and generate a create sink plan.

pub fn should_use_arrangement_backfill(&self) -> bool

pub fn should_use_snapshot_backfill(&self) -> bool

pub fn target_columns_to_plan_mapping( &self, tar_cols: &[ColumnCatalog], user_specified_columns: bool, ) -> Vec<Option<usize>>

used when the plan has a target relation such as DML and sink into table, return the mapping from table’s columns to the plan’s schema

Trait Implementations

impl Clone for PlanRoot

fn clone(&self) -> PlanRoot

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for PlanRoot

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

Formats the value using the given formatter.