pkg/cdc/hnsw_sinker_test.go

• Comprehensive test suite for HNSW CDC sinker functionality with 692
lines of test code
• Mock implementations for SQL executors and error
handling scenarios
• Test cases covering sinker creation, execution,
error handling, and data processing
• Tests for snapshot and atomic
batch processing with vector data

pkg/vectorindex/hnsw/sync_test.go

• Test suite for HNSW CDC synchronization with various operation
scenarios
• Tests covering upsert, delete, insert operations with
single and multiple models
• Mock implementations for SQL execution
and streaming operations
• Shuffle testing for concurrent operation
handling

pkg/vectorindex/hnsw/search_test.go

• Added mock functions for catalog SQL operations and multi-file
scenarios
• New test helper functions for creating metadata and index
batches
• Enhanced test coverage for search operations with multiple
index files

pkg/vectorindex/hnsw/model_test.go

• Added comprehensive test suite for HnswModel functionality
• Tests
cover search operations, loading/unloading, add/remove operations, and
SQL generation
• Includes edge case testing for nil model scenarios
•
Uses mock SQL functions for testing database interactions

pkg/sql/plan/function/func_hnsw_test.go

• Added test cases for hnswCdcUpdate function
• Tests various error
conditions including null arguments and invalid JSON
• Validates
function parameter validation and error handling

pkg/vectorindex/hnsw/build_test.go

• Updated test code to use HnswModel instead of HnswSearchIndex
•
Changed function call from NewHnswBuildIndex to NewHnswModelForBuild
•
Updated struct initialization to use new model type

pkg/vectorindex/types_test.go

• Added test cases for CDC functionality
• Tests Insert, Delete,
Upsert operations and JSON serialization
• Validates CDC data
structure behavior and state management

pkg/cdc/sinker_test.go

• Updated test calls to NewSinker to include the new cnUUID parameter

• Maintains test compatibility with updated function signature

pkg/frontend/cdc_test.go

• Updated mock NewSinker stub to include cnUUID parameter
• Maintains
test compatibility with updated function signature

pkg/sql/plan/function/function_id_test.go

• Updated predefined function IDs to include HNSW_CDC_UPDATE
•
Incremented FUNCTION_END_NUMBER to maintain test consistency

test/distributed/cases/vector/vector_hnsw_sync.result

• Added test results for HNSW CDC synchronization scenarios
• Covers
empty data, bulk load, and incremental update test cases
• Validates
vector search functionality after CDC operations

test/distributed/cases/vector/vector_hnsw_sync.sql

• Added comprehensive test cases for HNSW CDC synchronization
• Tests
PITR and CDC task creation, data operations, and vector searches
•
Includes scenarios for empty tables, bulk loads, and incremental
updates

pkg/vectorindex/hnsw/sync.go

• New CDC synchronization functionality for HNSW index updates via SQL
function hnsw_cdc_update()
• HnswSync struct for managing CDC
operations with insert, update, delete operations
• Multi-threaded
processing support with concurrent model loading and vector operations

• SQL generation for metadata and storage table updates

pkg/cdc/hnsw_sinker.go

• New hnswSyncSinker implementation for updating HNSW indexes via CDC
changes
• Support for both float32 and float64 vector types with
JSON-based CDC updates
• Transaction-based SQL execution with error
handling and rollback support
• Processing of snapshot and tail data
with atomic batch operations

pkg/sql/plan/function/func_hnsw.go

• Implemented hnswCdcUpdate function for processing CDC updates
•
Validates input parameters (database name, table name, dimension, CDC
JSON)
• Calls hnsw.CdcSync to perform the actual synchronization
•
Includes comprehensive error handling and logging

pkg/sql/compile/util.go

• Added genCdcHnswIndex function to generate CDC task creation SQL
•
Creates PITR and CDC task SQL statements for HNSW index
synchronization
• Includes placeholder logic for future CDC task
registration

pkg/sql/plan/function/list_builtIn.go

• Added HNSW_CDC_UPDATE function definition to built-in functions list

• Configured function signature with varchar and int32 parameters
returning uint64

pkg/sql/compile/ddl_index_algo.go

• Added call to genCdcHnswIndex in vector HNSW index handling
•
Executes generated CDC SQL statements during index creation

pkg/vectorindex/hnsw/model.go

• New HnswModel struct replacing HnswBuildIndex with enhanced CDC
support
• Added dirty tracking, atomic length counters, and view mode
support
• Enhanced file operations with checksum validation and
streaming SQL loading
• Methods for concurrent vector operations and
model lifecycle management

pkg/vectorindex/types.go

• Added CDC-related constants (CDC_INSERT, CDC_UPSERT, CDC_DELETE)
•
Introduced VectorIndexCdc and VectorIndexCdcEntry structs for CDC
operations
• Added HnswCdcParam struct for CDC parameters
•
Implemented CDC data manipulation methods (Insert, Upsert, Delete,
ToJson)

pkg/cdc/sinker.go

• Added support for CDCSinkType_HnswSync sink type
• Updated NewSinker
function signature to include cnUUID parameter
• Fixed potential nil
pointer dereference in error handling

pkg/sql/compile/ddl.go

• Added TODO comments for CDC task cleanup in DropIndex and DropTable
methods
• Placeholder logic for cleaning up CDC tasks when dropping
vector/fulltext indexes

pkg/vectorindex/sqlexec/sqlexec.go

• Added RunTxn function for executing SQL operations within
transactions
• Provides transaction-aware SQL execution with proper
context and options setup

pkg/sql/plan/function/func_cast.go

• Enhanced array casting with dimension validation
• Added bypass for
max dimension check when width equals MaxArrayDimension
• Improved
error handling for dimension mismatches

pkg/sql/plan/hnsw.go

• Commented out table scan validation in buildHnswCreate
• Relaxed
constraints on child node type checking

pkg/frontend/cdc_options.go

• Added support for CDCSinkType_HnswSync in CDC options validation
•
Extended sink type validation to include HNSW sync type

pkg/frontend/cdc_exector.go

• Updated NewSinker call to include cnUUID parameter
• Passes
executor's CN UUID to sinker creation

pkg/vectorindex/hnsw/build.go

• Refactored to use HnswModel instead of HnswBuildIndex for
consistency
• Removed duplicate HnswBuildIndex struct and related
methods
• Updated build operations to work with the new unified model
structure

pkg/vectorindex/hnsw/search.go

• Removed HnswSearchIndex struct and related methods (loadChunk,
LoadIndex, Search)
• Replaced HnswSearchIndex with HnswModel in the
HnswSearch struct
• Refactored LoadMetadata to be a standalone
function and updated field mappings
• Updated LoadIndex method to use
new HnswModel.LoadIndex signature

pkg/sql/plan/function/function_id.go

• Added HNSW_CDC_UPDATE function ID constant
• Updated
FUNCTION_END_NUMBER and function registry mapping

pkg/cdc/types.go

• Added CDCSinkType_HnswSync constant for HNSW synchronization sink
type

test/distributed/cases/vector/vector_hnsw.result

• Updated error message format for dimension mismatch from "vector ops
between different dimensions" to "expected vector dimension X !=
actual dimension Y"

test/distributed/cases/vector/vector_index.result

• Updated error message format for dimension mismatch to use new
standardized format

test/distributed/cases/array/array.result

• Updated error messages for array dimension mismatches to use new
standardized format