Struct mz_adapter::explain::optimizer_trace::OptimizerTrace

pub struct OptimizerTrace(Dispatch);

Provides functionality for tracing plans generated by the execution of an optimization pipeline.

Internally, this will create a layered tracing::subscriber::Subscriber consisting of one layer for each supported plan type T and wrap it into a dispatcher::Dispatch instance.

Use OptimizerTrace::as_guard to activate the dispatcher::Dispatch and collect a trace.

Use OptimizerTrace::into_rows or OptimizerTrace::into_plan_insights to cleanly destroy the OptimizerTrace instance and obtain the tracing result.

Tuple Fields

0: Dispatch

Implementations

impl OptimizerTrace

pub fn new(filter: Option<SmallVec<[NamedPlan; 4]>>) -> OptimizerTrace

Create a new OptimizerTrace.

The instance will only accumulate TraceEntry instances along the prefix of the given path if path is present, or it will accumulate all TraceEntry instances otherwise.

pub fn as_guard<'s>(&'s self) -> DispatchGuard<'s>

Enter this OptimizerTrace’s tracing dispatcher::Dispatch, returning a guard.

Linked to this OptimizerTrace with a lifetime to ensure OptimizerTrace::into_rows isn’t called until the guard is dropped.

pub async fn into_rows( self, format: ExplainFormat, config: &ExplainConfig, features: &OptimizerFeatures, humanizer: &dyn ExprHumanizer, row_set_finishing: Option<RowSetFinishing>, target_cluster: Option<&Cluster>, dataflow_metainfo: DataflowMetainfo, stage: ExplainStage, stmt_kind: ExplaineeStatementKind, insights_ctx: Option<PlanInsightsContext>, ) -> Result<Vec<Row>, AdapterError>

Convert the optimizer trace into a vector or rows that can be returned to the client.

pub async fn into_plan_insights( self, features: &OptimizerFeatures, humanizer: &dyn ExprHumanizer, row_set_finishing: Option<RowSetFinishing>, target_cluster: Option<&Cluster>, dataflow_metainfo: DataflowMetainfo, insights_ctx: Option<PlanInsightsContext>, ) -> Result<String, AdapterError>

Collect a insights::PlanInsights with insights about the the optimized plans rendered as a JSON String.

fn collect_all( &self, format: ExplainFormat, config: &ExplainConfig, features: &OptimizerFeatures, humanizer: &dyn ExprHumanizer, row_set_finishing: Option<RowSetFinishing>, target_cluster: Option<&str>, dataflow_metainfo: DataflowMetainfo, ) -> Result<TraceEntries<String>, ExplainError>

Collect all traced plans for all plan types T that are available in the wrapped dispatcher::Dispatch.

fn collect_global_plan( &self, ) -> Option<DataflowDescription<OptimizedMirRelationExpr>>

Collects the global optimized plan from the trace, if it exists.

fn collect_fast_path_plan(&self) -> Option<FastPathPlan>

Collects the fast path plan from the trace, if it exists.

fn collect_explainable_entries<T>( &self, format: &ExplainFormat, context: &mut ExplainContext<'_>, ) -> Result<Vec<TraceEntry<String>>, ExplainError>

where T: Clone + Debug + 'static, for<'a> Explainable<'a, T>: Explain<'a, Context = ExplainContext<'a>>,

Collect all trace entries of a plan type T that implements Explainable.

fn collect_scalar_entries<T>(&self) -> Vec<TraceEntry<String>>

where T: Clone + Debug + 'static + Display,

Collect all trace entries of a plan type T.

fn collect_string_entries(&self) -> Vec<TraceEntry<String>>

Collect all trace entries with plans of type String.

Trait Implementations

impl Debug for OptimizerTrace

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

Formats the value using the given formatter.