BasicStatsPlanVisitor — Computing Statistics for Cost-Based Optimization-spark技术分享

BasicStatsPlanVisitor — Computing Statistics for Cost-Based Optimization

BasicStatsPlanVisitor is a LogicalPlanVisitor that computes the statistics of a logical query plan for cost-based optimization (i.e. when cost-based optimization is enabled).

Note	Cost-based optimization is enabled when spark.sql.cbo.enabled configuration property is on, i.e. `true`, and is disabled by default.

BasicStatsPlanVisitor comes with custom handlers for a few logical operators and falls back to SizeInBytesOnlyStatsPlanVisitor for the others.

Table 1. BasicStatsPlanVisitor’s Visitor Handlers
Logical Operator	Handler	Behaviour
Aggregate	visitAggregate	Requests `AggregateEstimation` for statistics estimates and query hints or falls back to SizeInBytesOnlyStatsPlanVisitor
`Filter`	visitFilter	Requests `FilterEstimation` for statistics estimates and query hints or falls back to SizeInBytesOnlyStatsPlanVisitor
Join	visitJoin	Requests `JoinEstimation` for statistics estimates and query hints or falls back to SizeInBytesOnlyStatsPlanVisitor
Project	visitProject	Requests `ProjectEstimation` for statistics estimates and query hints or falls back to SizeInBytesOnlyStatsPlanVisitor