Sql expert

SQL Expert

Specialized expertise in SQL and relational database design, providing deep guidance on query optimization, indexing strategies, schema design, and database-specific features across PostgreSQL, MySQL, and SQL Server.

When to Use This Skill

Use this skill for:

• Writing complex SQL queries
• Optimizing slow queries
• Designing database schemas
• Understanding execution plans
• Creating effective indexes
• Database-specific features
• Performance tuning

Trigger phrases: "SQL", "database query", "optimize query", "index", "execution plan", "schema design", "PostgreSQL", "MySQL", "SQL Server"

What This Skill Does

Provides SQL expertise including:

• Query Optimization: Rewriting queries for performance
• Index Strategy: Designing effective indexes
• Schema Design: Normalization, data modeling
• Execution Plans: Reading and understanding plans
• Database Features: DB-specific capabilities
• Performance Tuning: Identifying and fixing bottlenecks

Instructions

Step 1: Analyze Query Performance

Understanding Execution Plans:

-- PostgreSQL
EXPLAIN ANALYZE
SELECT u.name, COUNT(o.id) as order_count
FROM users u
LEFT JOIN orders o ON o.user_id = u.id
WHERE u.created_at > '2024-01-01'
GROUP BY u.id, u.name;

-- MySQL
EXPLAIN FORMAT=JSON
SELECT ...

-- SQL Server
SET STATISTICS IO ON;
SET STATISTICS TIME ON;
SELECT ...

Reading Execution Plans:

Operation	Good Sign	Bad Sign
Index Scan	Using covering index	Full table scan when index exists
Seq Scan	Small table	Large table with selective WHERE
Nested Loop	Small inner table	Large tables on both sides
Hash Join	Large tables, equality join	Not enough memory
Sort	Indexed order	Large sort without index

Step 2: Optimize Common Query Patterns

**Avoid SELECT ***:

-- Bad: Fetches all columns
SELECT * FROM users WHERE status = 'active';

-- Good: Only needed columns
SELECT id, name, email FROM users WHERE status = 'active';

Use EXISTS Instead of IN for Subqueries:

-- Potentially slow with large subquery result
SELECT * FROM products
WHERE category_id IN (SELECT id FROM categories WHERE active = true);

-- Better: EXISTS stops at first match
SELECT * FROM products p
WHERE EXISTS (
    SELECT 1 FROM categories c
    WHERE c.id = p.category_id AND c.active = true
);

Optimize JOINs:

-- Bad: Joining on non-indexed columns
SELECT o.*, p.name
FROM orders o
JOIN products p ON p.sku = o.product_sku;

-- Good: Join on indexed foreign key
SELECT o.*, p.name
FROM orders o
JOIN products p ON p.id = o.product_id;

-- Use appropriate join type
-- INNER JOIN: Need matches in both tables
-- LEFT JOIN: Need all from left, matches from right
-- Don't use LEFT JOIN if you actually need INNER JOIN

Avoid Functions on Indexed Columns:

-- Bad: Function prevents index usage
SELECT * FROM users WHERE YEAR(created_at) = 2024;

-- Good: Range condition uses index
SELECT * FROM users
WHERE created_at >= '2024-01-01' AND created_at < '2025-01-01';

-- Bad: Function on column
SELECT * FROM users WHERE LOWER(email) = '[email protected]';

-- Good: Use functional index (PostgreSQL) or store lowercase
CREATE INDEX idx_users_email_lower ON users (LOWER(email));
-- Or store email in lowercase

Step 3: Design Effective Indexes

Index Types:

-- B-tree (default): Equality, range queries
CREATE INDEX idx_users_email ON users(email);

-- Composite: Multiple columns (order matters!)
CREATE INDEX idx_orders_user_date ON orders(user_id, created_at);
-- Good for: WHERE user_id = ? AND created_at > ?
-- Good for: WHERE user_id = ?
-- Bad for: WHERE created_at > ? (user_id not in query)

-- Partial index: Subset of rows
CREATE INDEX idx_active_users ON users(email) WHERE status = 'active';

-- Covering index: Includes all needed columns
CREATE INDEX idx_orders_covering ON orders(user_id, created_at)
INCLUDE (total, status);

-- GIN/Full-text: Text search, arrays, JSONB
CREATE INDEX idx_products_tags ON products USING GIN(tags);
CREATE INDEX idx_products_search ON products USING GIN(to_tsvector('english', name || ' ' || description));

Index Strategy Guidelines:

Query Pattern	Index Strategy
WHERE col = ?	Single column B-tree
WHERE col1 = ? AND col2 = ?	Composite (col1, col2)
WHERE col1 = ? ORDER BY col2	Composite (col1, col2)
WHERE col IN (?, ?, ?)	Single column B-tree
WHERE col LIKE 'prefix%'	B-tree (prefix only)
WHERE col LIKE '%substring%'	GIN trigram
WHERE col @> '{"key": "value"}'	GIN on JSONB

Step 4: Optimize Specific Patterns

Pagination:

-- Bad: OFFSET is slow for large offsets
SELECT * FROM products ORDER BY id LIMIT 20 OFFSET 10000;

-- Good: Keyset pagination (cursor-based)
SELECT * FROM products
WHERE id > 10000
ORDER BY id
LIMIT 20;

-- For complex ordering, use indexed columns
SELECT * FROM products
WHERE (created_at, id) > ('2024-01-15', 5000)
ORDER BY created_at, id
LIMIT 20;

Aggregations:

-- Slow: Aggregating large table
SELECT category_id, COUNT(*) FROM products GROUP BY category_id;

-- Better: Maintain materialized view or counter table
CREATE MATERIALIZED VIEW category_counts AS
SELECT category_id, COUNT(*) as product_count
FROM products GROUP BY category_id;

-- Or use approximate counts
SELECT reltuples::bigint AS estimate
FROM pg_class WHERE relname = 'products';

Batch Operations:

-- Bad: Many individual inserts
INSERT INTO logs (message) VALUES ('log1');
INSERT INTO logs (message) VALUES ('log2');
-- ... thousands more

-- Good: Batch insert
INSERT INTO logs (message) VALUES
('log1'), ('log2'), ('log3'), ... ;

-- Better: Use COPY (PostgreSQL) or LOAD DATA (MySQL)
COPY logs (message) FROM '/path/to/data.csv' CSV;

-- For updates, batch with CTEs
WITH batch AS (
    SELECT id FROM products WHERE needs_update = true LIMIT 1000
)
UPDATE products SET updated_at = NOW()
WHERE id IN (SELECT id FROM batch);

Step 5: Schema Design Best Practices

Normalization Levels:

-- 1NF: Atomic values, no repeating groups
-- Bad: Repeating columns
CREATE TABLE orders (
    id INT,
    product1_id INT, product1_qty INT,
    product2_id INT, product2_qty INT
);

-- Good: Separate table
CREATE TABLE orders (id INT PRIMARY KEY);
CREATE TABLE order_items (
    order_id INT REFERENCES orders(id),
    product_id INT,
    quantity INT
);

-- 3NF: No transitive dependencies
-- Bad: city depends on zip, not directly on id
CREATE TABLE customers (
    id INT,
    name VARCHAR(100),
    zip VARCHAR(10),
    city VARCHAR(100)  -- Derived from zip
);

-- Good: Separate lookup
CREATE TABLE zipcodes (zip VARCHAR(10) PRIMARY KEY, city VARCHAR(100));
CREATE TABLE customers (
    id INT,
    name VARCHAR(100),
    zip VARCHAR(10) REFERENCES zipcodes(zip)
);

Strategic Denormalization:

-- When to denormalize:
-- 1. Read-heavy with complex joins
-- 2. Aggregations computed frequently
-- 3. Performance critical paths

-- Example: Storing computed total
CREATE TABLE orders (
    id INT PRIMARY KEY,
    user_id INT,
    total DECIMAL(10,2),  -- Denormalized sum of items
    created_at TIMESTAMP
);

-- Keep consistent with trigger
CREATE FUNCTION update_order_total() RETURNS TRIGGER AS $$
BEGIN
    UPDATE orders SET total = (
        SELECT SUM(quantity * price)
        FROM order_items WHERE order_id = NEW.order_id
    ) WHERE id = NEW.order_id;
    RETURN NEW;
END;
$$ LANGUAGE plpgsql;

Step 6: Database-Specific Features

PostgreSQL:

-- JSONB for semi-structured data
CREATE TABLE events (
    id SERIAL PRIMARY KEY,
    data JSONB NOT NULL
);
CREATE INDEX idx_events_data ON events USING GIN(data);

-- Query JSONB
SELECT * FROM events WHERE data @> '{"type": "click"}';
SELECT data->>'user_id' as user_id FROM events;

-- Window functions
SELECT
    name,
    department,
    salary,
    AVG(salary) OVER (PARTITION BY department) as dept_avg,
    RANK() OVER (ORDER BY salary DESC) as salary_rank
FROM employees;

-- CTEs and recursive queries
WITH RECURSIVE org_tree AS (
    SELECT id, name, manager_id, 0 as depth
    FROM employees WHERE manager_id IS NULL

    UNION ALL

    SELECT e.id, e.name, e.manager_id, t.depth + 1
    FROM employees e
    JOIN org_tree t ON e.manager_id = t.id
)
SELECT * FROM org_tree;

-- Full-text search
SELECT * FROM products
WHERE to_tsvector('english', name || ' ' || description)
      @@ to_tsquery('english', 'wireless & headphones');

MySQL:

-- Generated columns
CREATE TABLE products (
    id INT PRIMARY KEY,
    price DECIMAL(10,2),
    tax_rate DECIMAL(4,2),
    total_price DECIMAL(10,2) GENERATED ALWAYS AS (price * (1 + tax_rate)) STORED
);

-- JSON functions
SELECT
    id,
    JSON_EXTRACT(metadata, '$.category') as category
FROM products
WHERE JSON_CONTAINS(metadata, '"electronics"', '$.tags');

-- Full-text search
ALTER TABLE products ADD FULLTEXT(name, description);
SELECT * FROM products
WHERE MATCH(name, description) AGAINST('wireless headphones' IN NATURAL LANGUAGE MODE);

SQL Server:

-- Window functions with ROWS/RANGE
SELECT
    date,
    amount,
    SUM(amount) OVER (
        ORDER BY date
        ROWS BETWEEN 6 PRECEDING AND CURRENT ROW
    ) as rolling_7day
FROM transactions;

-- JSON support
SELECT
    id,
    JSON_VALUE(data, '$.name') as name,
    JSON_QUERY(data, '$.addresses') as addresses
FROM customers
WHERE JSON_VALUE(data, '$.status') = 'active';

-- Temporal tables (system-versioned)
CREATE TABLE products (
    id INT PRIMARY KEY,
    name NVARCHAR(100),
    price DECIMAL(10,2),
    valid_from DATETIME2 GENERATED ALWAYS AS ROW START,
    valid_to DATETIME2 GENERATED ALWAYS AS ROW END,
    PERIOD FOR SYSTEM_TIME (valid_from, valid_to)
) WITH (SYSTEM_VERSIONING = ON);

-- Query historical data
SELECT * FROM products FOR SYSTEM_TIME AS OF '2024-01-01';

Best Practices

• Profile before optimizing - Use EXPLAIN ANALYZE
• Index selectively - Indexes have write overhead
• **Avoid SELECT *** - Specify needed columns
• Use appropriate data types - Smaller is faster
• Batch large operations - Don't update 1M rows at once
• Monitor slow queries - Log and review regularly
• Test with production-like data - Volume matters
• Keep statistics updated - ANALYZE/UPDATE STATISTICS

Common Patterns

Pattern 1: Upsert (Insert or Update)

-- PostgreSQL
INSERT INTO products (id, name, price)
VALUES (1, 'Widget', 9.99)
ON CONFLICT (id) DO UPDATE
SET name = EXCLUDED.name, price = EXCLUDED.price;

-- MySQL
INSERT INTO products (id, name, price)
VALUES (1, 'Widget', 9.99)
ON DUPLICATE KEY UPDATE
name = VALUES(name), price = VALUES(price);

-- SQL Server
MERGE INTO products AS target
USING (VALUES (1, 'Widget', 9.99)) AS source (id, name, price)
ON target.id = source.id
WHEN MATCHED THEN UPDATE SET name = source.name, price = source.price
WHEN NOT MATCHED THEN INSERT (id, name, price) VALUES (source.id, source.name, source.price);

Pattern 2: Running Totals

SELECT
    date,
    amount,
    SUM(amount) OVER (ORDER BY date) as running_total
FROM transactions;

Pattern 3: Gap and Island Detection

-- Find consecutive date ranges
WITH numbered AS (
    SELECT
        date,
        date - (ROW_NUMBER() OVER (ORDER BY date) * INTERVAL '1 day') as grp
    FROM events
)
SELECT
    MIN(date) as start_date,
    MAX(date) as end_date,
    COUNT(*) as consecutive_days
FROM numbered
GROUP BY grp
ORDER BY start_date;

Quality Checklist

• Query uses indexes effectively (EXPLAIN verified)
• No SELECT * in production queries
• JOINs are on indexed columns
• No functions on indexed columns in WHERE
• Large operations are batched
• Pagination uses keyset method
• Appropriate indexes created
• Query tested with production data volume

Related Skills

• performance-review - Database performance assessment
• code-quality - SQL code standards
• security-review - SQL injection prevention
• terraform-specialist - Database infrastructure

---

Version: 1.0.0 Last Updated: January 2026 Based on: Use the Index, Luke; awesome-claude-code-subagents patterns

Iterative Refinement Strategy

This skill is optimized for an iterative approach:

1. Execute: Perform the core steps defined above.
2. Review: Critically analyze the output (coverage, quality, completeness).
3. Refine: If targets aren't met, repeat the specific implementation steps with improved context.
4. Loop: Continue until the definition of done is satisfied.