Struct JoinOptimizer 

pub struct JoinOptimizer {
    pub max_patterns: usize,
}

Cost-based join reordering optimizer.

Fields

max_patterns: usize

Maximum number of patterns handled by the DP algorithm before falling back to the greedy heuristic.

Implementations

impl JoinOptimizer

pub fn new(max_patterns: usize) -> Self

Create a new optimizer.

max_patterns controls the DP threshold (recommended: ≤12).

pub fn optimize(&self, patterns: Vec<JoinPattern>) -> JoinPlan

Choose the best join order.

Uses DP when patterns.len() <= max_patterns, greedy otherwise.

pub fn greedy(patterns: &[JoinPattern]) -> JoinPlan

Greedy left-deep join reordering.

At each step we pick the unplaced pattern that, when joined with the current pipeline, yields the lowest cost increment.

pub fn dynamic_programming(&self, patterns: &[JoinPattern]) -> JoinPlan

Exact optimal join order via dynamic programming (subset enumeration).

Limited to patterns.len() ≤ 12 to keep exponential blowup manageable.

pub fn cost_join(left: &JoinPattern, right: &JoinPattern) -> f64

Incremental join cost when appending right to a pipeline ending with left.

Considers:

• base sizes of both patterns
• selectivity from shared variables (each shared variable reduces cost)

pub fn build_join_graph(patterns: &[JoinPattern]) -> Vec<JoinEdge>

Build a join graph: one edge for every pair of patterns that share at least one variable.

pub fn shared_variables(a: &JoinPattern, b: &JoinPattern) -> Vec<String>

Return variables that appear in both a and b.

pub fn merge_patterns(left: &JoinPattern, right: &JoinPattern) -> JoinPattern

Merge two patterns into the combined pattern produced by their join.

Trait Implementations

impl Clone for JoinOptimizer

fn clone(&self) -> JoinOptimizer

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for JoinOptimizer

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

Formats the value using the given formatter.