LLM-test/test/sql/2-sql测试-1/deepseek-v3.2-web.md

225 lines
7.7 KiB
Markdown
Raw Normal View History

我来设计一个具有复杂关联关系和迷惑项的电商数据库模型:
## 数据库E-Commerce System
### 表设计(包含迷惑项)
#### 1. 用户相关表
```sql
-- 用户主表(与客户表名称相似,容易混淆)
CREATE TABLE users (
user_id INT PRIMARY KEY AUTO_INCREMENT,
username VARCHAR(50) UNIQUE NOT NULL,
email VARCHAR(100) UNIQUE NOT NULL,
phone_number VARCHAR(20),
created_date TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
updated_date TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
status ENUM('active', 'inactive', 'suspended') DEFAULT 'active'
);
-- 客户信息表(与用户表相似,但有不同用途)
CREATE TABLE customers (
customer_id INT PRIMARY KEY AUTO_INCREMENT,
user_id INT NOT NULL, -- 外键但名称与users表主键不完全匹配
full_name VARCHAR(100),
birth_date DATE,
registration_date DATE,
customer_type ENUM('individual', 'business', 'vip') DEFAULT 'individual',
FOREIGN KEY (user_id) REFERENCES users(id) -- 故意写错应该是user_id REFERENCES users(user_id)
);
-- 用户地址表
CREATE TABLE user_addresses (
address_id INT PRIMARY KEY AUTO_INCREMENT,
user_id INT NOT NULL,
customer_id INT, -- 冗余字段,增加复杂度
address_line1 VARCHAR(200),
address_line2 VARCHAR(200),
city_id INT,
postal_code VARCHAR(20),
is_default BOOLEAN DEFAULT FALSE,
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
FOREIGN KEY (user_id) REFERENCES users(user_id),
FOREIGN KEY (customer_id) REFERENCES customers(customer_id)
);
```
#### 2. 商品相关表
```sql
-- 商品表
CREATE TABLE products (
product_id INT PRIMARY KEY AUTO_INCREMENT,
product_code VARCHAR(50) UNIQUE NOT NULL, -- 与sku相似
product_name VARCHAR(200) NOT NULL,
description TEXT,
price DECIMAL(10,2) NOT NULL,
cost_price DECIMAL(10,2), -- 与price相似
category_id INT,
supplier_id INT,
stock_quantity INT DEFAULT 0,
reorder_level INT DEFAULT 10,
created_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
modified_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
);
-- 产品表名称与products相似但内容不同
CREATE TABLE items (
item_id INT PRIMARY KEY AUTO_INCREMENT,
product_id INT NOT NULL,
sku VARCHAR(50) UNIQUE, -- 与product_code相似
variant_name VARCHAR(100),
additional_price DECIMAL(10,2) DEFAULT 0,
inventory_count INT DEFAULT 0,
FOREIGN KEY (product_id) REFERENCES products(product_id)
);
-- 商品分类表
CREATE TABLE categories (
category_id INT PRIMARY KEY AUTO_INCREMENT,
category_name VARCHAR(100) NOT NULL,
parent_category_id INT, -- 自引用,实现层级分类
category_path VARCHAR(500),
FOREIGN KEY (parent_category_id) REFERENCES categories(category_id)
);
```
#### 3. 订单相关表
```sql
-- 订单主表
CREATE TABLE orders (
order_id INT PRIMARY KEY AUTO_INCREMENT,
order_number VARCHAR(50) UNIQUE NOT NULL,
customer_id INT NOT NULL, -- 注意这里引用customers表不是users表
user_id INT, -- 冗余字段
order_date DATETIME DEFAULT CURRENT_TIMESTAMP,
order_status ENUM('pending', 'processing', 'shipped', 'delivered', 'cancelled') DEFAULT 'pending',
total_amount DECIMAL(10,2),
net_amount DECIMAL(10,2), -- 与total_amount相似
shipping_address_id INT,
billing_address_id INT,
payment_method VARCHAR(50)
);
-- 订单项表
CREATE TABLE order_items (
order_item_id INT PRIMARY KEY AUTO_INCREMENT,
order_id INT NOT NULL,
item_id INT, -- 引用items表
product_id INT, -- 也引用products表增加复杂度
quantity INT NOT NULL,
unit_price DECIMAL(10,2) NOT NULL,
discount DECIMAL(10,2) DEFAULT 0,
subtotal DECIMAL(10,2),
FOREIGN KEY (order_id) REFERENCES orders(order_id),
FOREIGN KEY (item_id) REFERENCES items(item_id),
FOREIGN KEY (product_id) REFERENCES products(product_id)
);
-- 发货表
CREATE TABLE shipments (
shipment_id INT PRIMARY KEY AUTO_INCREMENT,
order_id INT NOT NULL,
shipment_date DATE,
estimated_delivery DATE, -- 与delivery_date相似
actual_delivery DATE, -- 与estimated_delivery相似
carrier VARCHAR(100),
tracking_number VARCHAR(100),
status ENUM('preparing', 'shipped', 'in_transit', 'delivered') DEFAULT 'preparing'
);
```
#### 4. 多对多关系表
```sql
-- 商品-供应商关系(多对多)
CREATE TABLE product_suppliers (
product_id INT NOT NULL,
supplier_id INT NOT NULL,
is_primary BOOLEAN DEFAULT FALSE,
purchase_price DECIMAL(10,2),
PRIMARY KEY (product_id, supplier_id),
FOREIGN KEY (product_id) REFERENCES products(product_id)
-- 故意不创建supplier_id的外键约束增加测试难度
);
-- 用户-收藏夹关系(多对多)
CREATE TABLE user_wishlists (
user_id INT NOT NULL,
product_id INT NOT NULL,
added_date TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
priority INT DEFAULT 1,
PRIMARY KEY (user_id, product_id),
FOREIGN KEY (user_id) REFERENCES users(user_id),
FOREIGN KEY (product_id) REFERENCES products(product_id)
);
```
#### 5. 其他相关表(增加迷惑性)
```sql
-- 支付表
CREATE TABLE payments (
payment_id INT PRIMARY KEY AUTO_INCREMENT,
order_id INT NOT NULL,
payment_date DATETIME DEFAULT CURRENT_TIMESTAMP,
amount DECIMAL(10,2),
payment_type VARCHAR(50), -- 与payment_method相似
transaction_id VARCHAR(100),
payment_status ENUM('pending', 'completed', 'failed', 'refunded')
);
-- 评价表
CREATE TABLE reviews (
review_id INT PRIMARY KEY AUTO_INCREMENT,
product_id INT NOT NULL,
user_id INT NOT NULL,
rating INT CHECK (rating >= 1 AND rating <= 5),
comment TEXT,
review_date DATE, -- 与created_at相似但不同
helpful_count INT DEFAULT 0
);
-- 促销表
CREATE TABLE promotions (
promo_id INT PRIMARY KEY AUTO_INCREMENT,
promo_code VARCHAR(50) UNIQUE,
promo_name VARCHAR(100),
discount_type ENUM('percentage', 'fixed', 'buy_one_get_one'),
discount_value DECIMAL(10,2),
start_date DATE,
end_date DATE,
minimum_purchase DECIMAL(10,2)
);
```
### 迷惑项总结:
1. **相似表名**products/items, users/customers
2. **相似字段名**price/cost_price, total_amount/net_amount, estimated_delivery/actual_delivery
3. **含义相近字段**created_date/created_at/created_time, payment_type/payment_method
4. **冗余字段**order_items表中同时有item_id和product_id
5. **不一致的命名**user_id vs customer_id在不同表的引用
6. **缺失的外键**product_suppliers表中的supplier_id没有外键约束
7. **相似的用途不同名**product_code vs sku
8. **同义词字段**stock_quantity vs inventory_count
### 测试查询示例:
```sql
-- 复杂查询1找出每个用户最常购买的商品类别
-- 需要连接users → customers → orders → order_items → products → categories
-- 复杂查询2计算每个供应商的商品在最近一个月的销售额
-- 需要连接product_suppliers → products → order_items → orders
-- 复杂查询3找出有相同商品的用户对社交推荐
-- 需要多对多关系分析users → user_wishlists → products ← user_wishlists ← users
-- 复杂查询4层级分类的商品统计递归查询
-- 需要使用WITH RECURSIVE处理categories表的层级关系
```
这个设计包含了:
- 一对多关系(用户-地址、订单-订单项)
- 多对多关系(用户-收藏夹、商品-供应商)
- 自引用关系(分类层级)
- 复杂的业务逻辑和数据冗余
- 各种迷惑项和陷阱
你可以用这个数据库模型来测试LLM对复杂SQL查询的理解和生成能力。