database-optimization

Database Optimization Skill

When to Use

Use this skill when:

• Optimizing slow SQL queries
• Analyzing missing or redundant indexes
• Reviewing Alembic migrations for safety
• Identifying N+1 query problems
• Designing database schemas
• Planning data migrations
• Reviewing query execution plans

Optimization Checklist

1. Index Analysis

-- Check missing indexes
SELECT schemaname, tablename, indexname, indexdef
FROM pg_indexes
WHERE tablename = 'your_table';

-- Identify unused indexes
SELECT * FROM pg_stat_user_indexes
WHERE idx_scan = 0;

Index Recommendations:

✅ Always index:
- Primary keys (automatic)
- Foreign keys
- Columns in WHERE clauses
- Columns in ORDER BY
- Columns in JOIN conditions

❌ Avoid indexing:
- Low cardinality columns (boolean, status)
- Frequently updated columns
- Small tables (<1000 rows)

2. N+1 Query Detection

# ❌ N+1 Problem
users = await session.execute(select(User))
for user in users:
    # This triggers N additional queries!
    orders = await session.execute(
        select(Order).where(Order.user_id == user.id)
    )

# ✅ Solution: Eager loading
from sqlalchemy.orm import selectinload

users = await session.execute(
    select(User).options(selectinload(User.orders))
)

3. Query Optimization Patterns

# ✅ Use pagination
from sqlalchemy import select
from app.models import Product

async def get_products(skip: int = 0, limit: int = 20):
    query = select(Product).offset(skip).limit(limit)
    result = await session.execute(query)
    return result.scalars().all()

# ✅ Use specific columns (not SELECT *)
query = select(Product.id, Product.name, Product.price)

# ✅ Use exists() for existence checks
from sqlalchemy import exists
query = select(exists().where(User.email == email))

4. Migration Safety

# ✅ Safe migration patterns
def upgrade():
    # Add column with default (no table lock)
    op.add_column('users', sa.Column('status', sa.String(20), 
                                      server_default='active'))

# ❌ Dangerous patterns
def upgrade():
    # Avoid: Rename column (breaks app)
    op.alter_column('users', 'name', new_column_name='full_name')
    
    # Avoid: Change column type (data loss risk)
    op.alter_column('users', 'age', type_=sa.String())