database（数据库）

Database Development

Schema design, optimization, and management best practices.

Schema Design

Normalization

-- 1NF: Atomic values, no repeating groups
-- BAD
CREATE TABLE orders (
    id INT,
    products VARCHAR(255)  -- "shirt,pants,shoes" - NOT atomic
);

-- GOOD
CREATE TABLE orders (id INT PRIMARY KEY);
CREATE TABLE order_items (
    order_id INT REFERENCES orders(id),
    product_id INT REFERENCES products(id),
    quantity INT
);

-- 2NF: No partial dependencies (all non-key columns depend on entire PK)
-- 3NF: No transitive dependencies (non-key columns don't depend on other non-key columns)

Data Types

-- Use appropriate types
CREATE TABLE users (
    id BIGINT UNSIGNED AUTO_INCREMENT PRIMARY KEY,  -- BIGINT for large tables
    uuid CHAR(36) NOT NULL UNIQUE,                  -- Fixed-length UUID
    email VARCHAR(255) NOT NULL,                    -- Variable length
    status ENUM('active', 'inactive', 'banned'),    -- Constrained values
    balance DECIMAL(10,2) NOT NULL DEFAULT 0,       -- Exact precision for money
    metadata JSON,                                   -- Flexible schema
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
);

-- PostgreSQL specific
CREATE TABLE events (
    id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
    data JSONB NOT NULL,                            -- Binary JSON, indexable
    tags TEXT[] NOT NULL DEFAULT '{}',              -- Array type
    tsv TSVECTOR,                                   -- Full-text search
    created_at TIMESTAMPTZ DEFAULT NOW()            -- Timezone-aware
);

Relationships

-- One-to-Many
CREATE TABLE posts (
    id SERIAL PRIMARY KEY,
    user_id INT NOT NULL REFERENCES users(id) ON DELETE CASCADE,
    title VARCHAR(255) NOT NULL
);

-- Many-to-Many with pivot table
CREATE TABLE post_tags (
    post_id INT REFERENCES posts(id) ON DELETE CASCADE,
    tag_id INT REFERENCES tags(id) ON DELETE CASCADE,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    PRIMARY KEY (post_id, tag_id)
);

-- One-to-One
CREATE TABLE user_profiles (
    user_id INT PRIMARY KEY REFERENCES users(id) ON DELETE CASCADE,
    bio TEXT,
    avatar_url VARCHAR(255)
);

Indexing

Index Types

-- B-Tree (default, most common)
CREATE INDEX idx_users_email ON users(email);

-- Composite index (order matters!)
CREATE INDEX idx_orders_user_status ON orders(user_id, status);
-- Good for: WHERE user_id = ? AND status = ?
-- Good for: WHERE user_id = ?
-- NOT good for: WHERE status = ?  (leftmost prefix rule)

-- Unique index
CREATE UNIQUE INDEX idx_users_email_unique ON users(email);

-- Partial index (PostgreSQL)
CREATE INDEX idx_active_users ON users(email) WHERE status = 'active';

-- Expression index
CREATE INDEX idx_users_email_lower ON users(LOWER(email));

-- Full-text index (MySQL)
CREATE FULLTEXT INDEX idx_posts_content ON posts(title, content);

-- GIN index for JSONB (PostgreSQL)
CREATE INDEX idx_events_data ON events USING GIN(data);

Index Strategy

-- Index columns used in:
-- 1. WHERE clauses
-- 2. JOIN conditions
-- 3. ORDER BY (if used frequently)
-- 4. Foreign keys

-- Check existing indexes
SHOW INDEX FROM orders;  -- MySQL
\d orders               -- PostgreSQL

-- Analyze query execution
EXPLAIN ANALYZE SELECT * FROM orders WHERE user_id = 123;

Query Optimization

EXPLAIN Analysis

-- MySQL
EXPLAIN SELECT * FROM orders
WHERE user_id = 123
AND created_at > '2024-01-01';

-- Look for:
-- type: "ref" or "range" (good), "ALL" (table scan, bad)
-- key: Which index is used (NULL = no index)
-- rows: Estimated rows examined

-- PostgreSQL
EXPLAIN (ANALYZE, BUFFERS, FORMAT TEXT)
SELECT * FROM orders WHERE user_id = 123;

Common Optimizations

-- BAD: SELECT *
SELECT * FROM users WHERE id = 1;

-- GOOD: Select only needed columns
SELECT id, name, email FROM users WHERE id = 1;

-- BAD: OR can prevent index usage
SELECT * FROM users WHERE email = 'a@b.com' OR name = 'John';

-- GOOD: Use UNION for OR conditions
SELECT * FROM users WHERE email = 'a@b.com'
UNION ALL
SELECT * FROM users WHERE name = 'John' AND email != 'a@b.com';

-- BAD: Functions on indexed columns
SELECT * FROM users WHERE YEAR(created_at) = 2024;

-- GOOD: Use range
SELECT * FROM users
WHERE created_at >= '2024-01-01'
AND created_at < '2025-01-01';

-- BAD: LIKE with leading wildcard
SELECT * FROM products WHERE name LIKE '%shirt%';

-- GOOD: Full-text search
SELECT * FROM products
WHERE MATCH(name) AGAINST('shirt' IN BOOLEAN MODE);

N+1 Problem

-- BAD: N+1 queries
-- Query 1: SELECT * FROM posts LIMIT 10
-- Query 2-11: SELECT * FROM users WHERE id = ?  (for each post)

-- GOOD: JOIN
SELECT p.*, u.name as author_name
FROM posts p
JOIN users u ON p.user_id = u.id
LIMIT 10;

-- GOOD: Subquery with IN
SELECT * FROM users
WHERE id IN (SELECT DISTINCT user_id FROM posts WHERE ...);

Migrations

Migration Best Practices

-- Always wrap in transactions
BEGIN;

-- Add column (non-locking in PostgreSQL)
ALTER TABLE users ADD COLUMN phone VARCHAR(20);

-- Add index concurrently (PostgreSQL, non-locking)
CREATE INDEX CONCURRENTLY idx_users_phone ON users(phone);

-- Rename column safely
ALTER TABLE users RENAME COLUMN phone TO phone_number;

COMMIT;

-- Rollback script
BEGIN;
ALTER TABLE users DROP COLUMN phone_number;
DROP INDEX idx_users_phone;
COMMIT;

Safe Migration Patterns

-- Adding NOT NULL column with default
-- Step 1: Add nullable column
ALTER TABLE users ADD COLUMN status VARCHAR(20);

-- Step 2: Backfill data
UPDATE users SET status = 'active' WHERE status IS NULL;

-- Step 3: Add NOT NULL constraint
ALTER TABLE users ALTER COLUMN status SET NOT NULL;
ALTER TABLE users ALTER COLUMN status SET DEFAULT 'active';

-- Renaming tables (zero downtime)
-- Step 1: Create new table
CREATE TABLE accounts (LIKE users INCLUDING ALL);

-- Step 2: Copy data
INSERT INTO accounts SELECT * FROM users;

-- Step 3: Create triggers for sync
-- Step 4: Switch application
-- Step 5: Drop old table

Performance

Connection Pooling

// Node.js with pg-pool
const { Pool } = require('pg');

const pool = new Pool({
    host: 'localhost',
    database: 'myapp',
    max: 20,                    // Max connections
    idleTimeoutMillis: 30000,   // Close idle connections
    connectionTimeoutMillis: 2000
});

// Always use pool, not direct connections
const result = await pool.query('SELECT * FROM users WHERE id = $1', [userId]);

Pagination

-- BAD: OFFSET for large datasets
SELECT * FROM posts ORDER BY created_at DESC LIMIT 20 OFFSET 10000;
-- Gets slower as offset increases

-- GOOD: Cursor-based pagination
SELECT * FROM posts
WHERE created_at < '2024-01-15 10:30:00'
ORDER BY created_at DESC
LIMIT 20;

-- GOOD: Keyset pagination with ID
SELECT * FROM posts
WHERE (created_at, id) < ('2024-01-15 10:30:00', 12345)
ORDER BY created_at DESC, id DESC
LIMIT 20;

Batch Operations

-- BAD: Many individual inserts
INSERT INTO logs (message) VALUES ('log1');
INSERT INTO logs (message) VALUES ('log2');
-- ... 1000 more

-- GOOD: Batch insert
INSERT INTO logs (message) VALUES
    ('log1'),
    ('log2'),
    ('log3');
    -- Up to ~1000 at a time

-- GOOD: COPY for bulk loading (PostgreSQL)
COPY logs (message) FROM '/path/to/file.csv' WITH CSV;

Transactions

ACID Properties

-- Atomicity: All or nothing
BEGIN;
UPDATE accounts SET balance = balance - 100 WHERE id = 1;
UPDATE accounts SET balance = balance + 100 WHERE id = 2;
-- If any fails, ROLLBACK
COMMIT;

-- Isolation levels
SET TRANSACTION ISOLATION LEVEL READ COMMITTED;  -- Default
SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;   -- Strictest

-- Deadlock prevention: Always lock in same order
BEGIN;
SELECT * FROM accounts WHERE id = 1 FOR UPDATE;  -- Lock row
SELECT * FROM accounts WHERE id = 2 FOR UPDATE;
-- Do work
COMMIT;

NoSQL Patterns

Document Database (MongoDB)

// Schema design: Embed vs Reference
// Embed: Data accessed together, 1:few relationships
{
    _id: ObjectId("..."),
    title: "Blog Post",
    author: {                    // Embedded
        name: "John",
        email: "john@example.com"
    },
    comments: [                  // Embedded array
        { text: "Great!", user: "Jane" }
    ]
}

// Reference: Large documents, many relationships
{
    _id: ObjectId("..."),
    title: "Blog Post",
    author_id: ObjectId("...")   // Reference to users collection
}

// Indexes
db.posts.createIndex({ "author_id": 1 });
db.posts.createIndex({ "title": "text", "content": "text" });  // Text search

Key-Value (Redis)

# Caching pattern
SET user:123 '{"name":"John"}' EX 3600  # Expires in 1 hour
GET user:123

# Counter
INCR page:views:homepage
GET page:views:homepage

# Rate limiting
INCR rate:ip:192.168.1.1
EXPIRE rate:ip:192.168.1.1 60  # Reset every minute

Backup & Recovery

# MySQL
mysqldump -u root -p database > backup.sql
mysql -u root -p database < backup.sql

# PostgreSQL
pg_dump -Fc database > backup.dump
pg_restore -d database backup.dump

# Point-in-time recovery (PostgreSQL)
# Requires WAL archiving configured
pg_basebackup -D /backup/base -Fp -Xs -P

Monitoring Queries

-- MySQL slow query log
SET GLOBAL slow_query_log = 'ON';
SET GLOBAL long_query_time = 1;  -- Log queries > 1 second

-- PostgreSQL: Currently running queries
SELECT pid, now() - pg_stat_activity.query_start AS duration, query
FROM pg_stat_activity
WHERE state = 'active'
ORDER BY duration DESC;

-- Table sizes
SELECT
    table_name,
    pg_size_pretty(pg_total_relation_size(quote_ident(table_name)))
FROM information_schema.tables
WHERE table_schema = 'public'
ORDER BY pg_total_relation_size(quote_ident(table_name)) DESC;