saygoweb / anorm / 18485748311

84%

push

github

Implement comprehensive join optimization system to solve N+1 query problem (#42)

* ai: optimize join plan

* Phase 1: Core Infrastructure Setup - Batch Loading Interfaces and Strategy Components

- Created BatchLoaderInterface for batch relationship loading
- Created QueryStrategyInterface for strategy selection
- Extended abstract Relationship class with batch loading methods
- Implemented DataSizeEstimator for data transfer optimization
- Implemented QueryStrategySelector with decision logic
- Implemented FieldSelectionParser for field selection syntax
- Updated plan document to track progress

This establishes the foundation for optimizing N+1 query problems
in relationship loading through intelligent batch loading strategies.

* Phase 2: IN Clause Batch Loading Implementation

- Created OneHasManyBatchLoader for hasMany relationships
- Created ManyHasOneBatchLoader for belongsTo relationships
- Created ManyHasManyBatchLoader for many-to-many relationships
- Extended all relationship classes with batch loading methods
- Added estimateDataSize() and getCardinality() methods
- Fixed RelationshipManager duplicate method issues
- All tests passing - backward compatibility maintained

This implements the core batch loading functionality that reduces
N+1 queries to single IN clause queries for each relationship type.

* Phase 4: QueryBuilder Integration - Batch Loading Orchestration

- Created BatchLoadingOrchestrator for coordinating batch loading
- Modified QueryBuilder.some() to use batch loading when relationships are specified
- Added configuration options for batch loading behavior
- Implemented strategy selection and fallback mechanisms
- Added comprehensive batch loading tests
- All 157 tests passing with backward compatibility maintained

The QueryBuilder now automatically uses batch loading for eager relationships,
reducing N+1 queries to O(1) queries per relationship type. The system
intelligently selects between batch loading and individual... (continued)

848 of 963 new or added lines in 19 files covered. (88.06%)

1636 of 1939 relevant lines covered (84.37%)

15.71 hits per line