Expand description
Compatibility facade over the framework and domain-owned optimizer families.
Modules§
- context
- cost
- graph
- logical
- Logical optimizer operator metadata.
- memo
- operator
- Compatibility re-exports for optimizer operator metadata.
- predicate
- properties
- relational
- relational_
join - Relational inner-join graph contracts and validation.
- relational_
join_ hypergraph - Hypergraph csg-cmp enumeration for bound relational join operators.
- relational_
join_ rewrite - CD-C legality analysis and memo-backed join rewrite enumeration.
- relational_
planning - relational_
profile - relational_
sargability - Extracts index-compatible constraints without binding values or accessing storage.
- rule
- search
- stage
- trace
- vector
Structs§
- Adaptive
Vector Backend Decision - Adaptive
Vector Backend Input - Adaptive
Vector Backend Policy - Applied
Rule - Cascades
Optimizer - Fast
Path Physical Phase - GroupId
- Logical
Phase - Logical
Plan Root - Lowering
Ready Logical Plan Root - Lowering
Ready Phase - Memo
- Memo
Group - Operator
Cardinality Estimate - The optimizer’s output-cardinality estimate for one physical operator.
- Optimization
Pipeline - Optimization
Search Report - Optimization
Stage - Optimizer
Catalog - Optimizer
Catalog Indexes - Optimizer
Catalog Statistics - Optimizer
Config - Optimizer
Context - Optimizer
Index Statistics - Optimizer
Trace - Physical
Operator Id - A plan-local physical operator identity assigned in pre-order.
- Physical
Phase - Physical
Properties - Pipeline
Execution - Plan
Cost - Plan
Cost Breakdown - Cardinality and unweighted logical work, plus the weighted scalar total.
- Planned
Vector Search - Relational
Access Cost Context - Immutable structural work for the row source of one relation.
- Relational
Access Path Descriptor - Relational
CsgCmp Alternative - Relational
CsgCmp Enumeration - Relational
CsgCmp Join Implementation - A pre-bound implementation supported by the caller’s two-relation kernel. It competes inside memo selection, not after choosing a join order. Predicate IDs name the complete operator, including residual predicates.
- Relational
CsgCmp Plan - Relational
Inner Join Enumeration - Relational
Join Access Path - Relational
Join Conflict Analysis - Relational
Join Conflict Descriptor - Relational
Join Conflict Rule - Relational
Join Cost Contexts - Per-binding row-source costs for one immutable join enumeration.
- Relational
Join Enumeration Config - Relational
Join Graph - Relational
Join Operator - Relational
Join Operator Id - Relational
Join Plan - Relational
Join Planning Attempt - Relational
Join Planning Budget - Relational
Join Planning Cost - Relational
Join Planning Outcome - Relational
Join Predicate - Relational
Join Predicate Id - Relational
Join Relation - Relational
Join Rewrite Enumeration - Relational
Join Rewrite Plan - Relational
Join Rewrite Problem - Relational
Join Rewrite Step - Relational
Join Step - Relational
Nested Loop Join Cost - Conservative row-work estimate for an index nested-loop join.
- Relational
Operator Cardinality Profile - Estimated and observed output cardinality at one relational operator boundary.
- Relational
Operator Id - Stable, plan-local identifier for a physical relational access or join operator.
- Required
Properties - Resource
Hints - Rule
Application - Rule
Batch - Rule
Event - RuleId
- Rule
Promise - Rule
Stage - Search
Field Ref - Search
Predicate - Search
Predicate Parse Error - Search
Predicate Pushdown - Search
Predicate Set - Search
Scan Predicate Support - Selected
Plan Trace - Stage
Rule Batch - Stage
Stats - Stage
Trace - Vector
Plan Properties
Enums§
- Adaptive
Vector Backend - Apply
Order - Distribution
- Explain
Mode - Logical
Plan Class - Logical
Plan Kind - Memory
Budget Class - Optimizer
Search Directive - Optimizer
Search Directive Error - Physical
Plan Class - Physical
Plan Kind - Plan
Children - Plan
Phase Kind - Query
Family - Relational
Access Path Kind - Relational
CsgCmp Plan Node - Relational
CsgCmp Right Input Policy - Execution capability for a CSG-CMP join’s non-singleton right input.
- Relational
Equi Join Algorithm - Relational
Join Access Applicability - Relational
Join Cardinality - Relational
Join Enumeration Error - Relational
Join Operator Kind - Relational
Join Planning Directive - Relational
Join Planning Fallback Class - Relational
Join Planning Reason - Relational
Join Planning Status - Relational
Join Planning Strategy - Relational
Join Rewrite Error - Relational
Join Right Input - Relational
Join Selectivity - Relational
Join Tree - Relational
Operator Kind - Rule
Kind - Rule
Outcome - Scan
Pruning Support - Search
Mode - Search
Predicate Op - Search
Scalar Value - Statement
Class - Trace
Sink - Vector
Compression Preference - Vector
Plan Error - Vector
Precision
Traits§
Functions§
- analyze_
relational_ join_ conflicts - apply_
rule_ batch - enumerate_
relational_ csg_ cmp_ joins - enumerate_
relational_ csg_ cmp_ joins_ with_ cost_ contexts - enumerate_
relational_ csg_ cmp_ joins_ with_ implementations - enumerate_
relational_ csg_ cmp_ joins_ with_ right_ input_ policy - enumerate_
relational_ inner_ joins - enumerate_
relational_ inner_ joins_ with_ cost_ contexts - enumerate_
relational_ join_ rewrites - enumerate_
relational_ join_ rewrites_ with_ cost_ contexts - estimate_
relational_ access_ cost - estimate_
relational_ access_ path_ cost - Estimates logical work for a validated access descriptor.
- estimate_
relational_ access_ path_ cost_ with_ context - Uses the same raw-component contract with relation-specific structural context. Row-root setup CPU and metadata access patterns retain separate raw components; cardinality and output stay independent. The final constructor weights every component once.
- estimate_
relational_ join_ cost - estimate_
relational_ join_ cost_ with_ contexts - Extends join composition with the original row sources of its two inputs.
- estimate_
relational_ nested_ loop_ join_ cost - Estimates row visits for an index nested-loop join.
- estimate_
relational_ probe_ join_ cost - Extends a left-deep plan with one probe join using the canonical relational cost model.
- normalize_
search_ enum_ value - plan_
class_ counts - plan_
operator_ counts - plan_
vector_ search - push_
search_ predicates - search_
field_ is_ enum_ like - select_
adaptive_ vector_ backend - select_
relational_ access_ path - select_
relational_ access_ path_ with_ context - Selects an access after context-aware skyline pruning. Candidate costs are comparable only when this context belongs to their common row source.
- skyline_
prune_ relational_ access_ paths - Retains the Pareto frontier of relational access paths.
- skyline_
prune_ relational_ access_ paths_ with_ context - Computes the frontier with the same immutable relation context used for final selection. The legacy entrypoint retains descriptor-only costs.
- validate_
vector_ pipeline - visit_
plan - visit_
plan_ with_ ids - Visits a physical plan in pre-order and assigns each operator its plan-local identity.