aegis-query

SQL query engine for the Aegis Database Platform.

Overview

aegis-query provides a complete SQL query pipeline including parsing, semantic analysis, query planning, and execution. It supports standard SQL with extensions for time series and document queries.

Features

• SQL Parser - Full SQL syntax support via sqlparser
• Semantic Analyzer - Type checking and schema validation
• Query Planner - Cost-based optimization with rule transformations
• Query Executor - Vectorized execution engine
• Index Support - B-tree and Hash index selection for predicates
• Direct Execution API - execute_update_direct and execute_update_indexed_fn for bypassing the full pipeline
• Plan Cache - LRU-based plan cache (1024 entries) for repeated query patterns
• Query Limits - QueryLimits with configurable max_result_rows and query_timeout_secs
• Data Classification - DataClassification enum for column-level sensitivity tagging
• VACUUM Support - Reclaim storage space and rebuild indexes

Architecture

SQL Query
    │
    ▼
┌─────────────┐
│   Parser    │  SQL text → AST
└──────┬──────┘
       │
       ▼
┌─────────────┐
│  Analyzer   │  AST → Validated AST (types, schema)
└──────┬──────┘
       │
       ▼
┌─────────────┐
│  Planner    │  Validated AST → Optimized Plan
└──────┬──────┘
       │
       ▼
┌─────────────┐
│  Executor   │  Plan → Results
└─────────────┘

Modules

Usage

[dependencies]
aegis-query = { path = "../aegis-query" }

Example

use aegis_query::{QueryEngine, QueryResult};

let engine = QueryEngine::new(storage)?;

// Execute a query
let result = engine.execute("SELECT * FROM users WHERE age > 21")?;

// Iterate results
for row in result.rows() {
    println!("{:?}", row);
}

Supported SQL

DDL:

CREATE TABLE users (
    id INT PRIMARY KEY,
    name VARCHAR(100) NOT NULL,
    email VARCHAR(255) UNIQUE,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);

DROP TABLE users;
ALTER TABLE users ADD COLUMN status VARCHAR(20);

DML:

INSERT INTO users (name, email) VALUES ('Alice', 'alice@example.com');
UPDATE users SET status = 'active' WHERE id = 1;
DELETE FROM users WHERE status = 'inactive';

Queries:

SELECT u.name, COUNT(o.id) as order_count
FROM users u
LEFT JOIN orders o ON u.id = o.user_id
WHERE u.created_at > '2024-01-01'
GROUP BY u.name
HAVING COUNT(o.id) > 5
ORDER BY order_count DESC
LIMIT 10;

Query Optimization

The planner applies several optimization rules:

• Predicate Pushdown - Move filters closer to data source
• Projection Pruning - Only read needed columns
• Join Reordering - Optimize join order based on cardinality
• Index Selection - Choose optimal indexes for predicates
• Constant Folding - Evaluate constant expressions at plan time
• B-tree Index Scan - Use B-tree indexes for range and equality predicates
• Hash Index Scan - Use hash indexes for exact-match lookups

Tests

cargo test -p aegis-query

Test count: 634 tests (workspace total)

License

Apache-2.0