Struct ParallelSearchConfig 

pub struct ParallelSearchConfig {
    pub enabled: bool,
    pub max_depth: usize,
    pub max_states_per_layer: usize,
    pub pruning_threshold: f64,
    pub time_budget_ms: u64,
    pub use_time_budget: bool,
    pub verbose: bool,
}

Configuration for parallel join order search

Fields

enabled: bool

Enable parallel BFS search (vs sequential DFS)

max_depth: usize

Maximum depth to explore with parallel BFS (tables with >max_depth use DFS)

max_states_per_layer: usize

Maximum states per layer before pruning

pruning_threshold: f64

Prune states with cost > best * threshold

time_budget_ms: u64

Maximum time budget for join order search (milliseconds) Default: 1000ms for OLAP workloads

use_time_budget: bool

Whether to use time-bounded search (vs table-count cutoff)

verbose: bool

Enable verbose logging of search statistics

Implementations

impl ParallelSearchConfig

pub fn with_table_count(num_tables: usize) -> Self

Create a config with adaptive time budget based on table count

The time budget increases with query complexity:

• 1-3 tables: 500ms (simple queries)
• 4-5 tables: 1000ms (typical OLAP queries)
• 6-7 tables: 1500ms (complex multi-way joins like Q7)
• 8+ tables: 2000ms (very complex queries like Q21)

This adaptive approach gives more time to complex queries that need it while keeping simple queries fast.

Trait Implementations

impl Clone for ParallelSearchConfig

fn clone(&self) -> ParallelSearchConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ParallelSearchConfig

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

Formats the value using the given formatter.

impl Default for ParallelSearchConfig

fn default() -> Self

Returns the “default value” for a type.