Struct JoinSelectivity 

pub struct JoinSelectivity {
    pub left_rel: RelId,
    pub right_rel: RelId,
    pub left_keys: Vec<usize>,
    pub right_keys: Vec<usize>,
    pub selectivity: f64,
    pub is_pk_fk: bool,
    /* private fields */
}

Join selectivity model for estimating join output cardinality.

Tracks information about joins between two relations, including the join keys and estimated selectivity. This is crucial for the optimizer to choose between nested-loop, hash, and sort-merge join strategies.

Fields

left_rel: RelId

Left relation in the join

right_rel: RelId

Right relation in the join

left_keys: Vec<usize>

Column indices used as join keys on the left relation

right_keys: Vec<usize>

Column indices used as join keys on the right relation

selectivity: f64

Estimated selectivity factor (0.0 to 1.0)

is_pk_fk: bool

Whether this is a primary key to foreign key join

Implementations

impl JoinSelectivity

pub fn new(left_rel: RelId, right_rel: RelId) -> Self

Creates a new join selectivity model between two relations.

Initializes with default selectivity of 1.0 (cross product).

Arguments

• left_rel - The left relation’s ID
• right_rel - The right relation’s ID

Returns

A new JoinSelectivity with default values.

pub fn set_keys(&mut self, left_keys: Vec<usize>, right_keys: Vec<usize>)

Sets the join keys for both relations.

Arguments

• left_keys - Column indices on the left relation
• right_keys - Column indices on the right relation

pub fn set_selectivity(&mut self, selectivity: f64)

Sets the selectivity factor.

Arguments

• selectivity - The selectivity factor (0.0 to 1.0)

pub fn mark_pk_fk(&mut self)

Marks this as a primary key to foreign key join.

PK-FK joins have special selectivity characteristics: the output cardinality equals the FK side’s cardinality.

pub fn estimate_output_rows(&self, left_rows: u64, right_rows: u64) -> u64

Estimates the output row count for this join.

For PK-FK joins, returns the cardinality of the FK side. For other joins, returns: left_rows * right_rows * selectivity

Arguments

• left_rows - Cardinality of the left relation
• right_rows - Cardinality of the right relation

Returns

The estimated output cardinality (minimum of 1)

pub fn estimate_selectivity_from_stats( left_distinct: u64, right_distinct: u64, ) -> f64

Estimates selectivity from column statistics.

Uses the “independence assumption” and distinct value counts: selectivity = 1 / max(distinct_left, distinct_right)

Arguments

• left_distinct - Distinct value count for left join key
• right_distinct - Distinct value count for right join key

Returns

The estimated selectivity

pub fn update_from_observation( &mut self, left_rows: u64, right_rows: u64, output_rows: u64, )

Updates selectivity based on observed join statistics.

This can be called after query execution to improve future estimates.

Arguments

• left_rows - Actual left cardinality
• right_rows - Actual right cardinality
• output_rows - Actual output cardinality

Trait Implementations

impl Clone for JoinSelectivity

fn clone(&self) -> JoinSelectivity

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for JoinSelectivity

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

Formats the value using the given formatter.