像Excel一样使用SQL进行数据分析
- 2019 年 11 月 22 日
- 笔记
Excel是数据分析中最常用的工具 ，利用Excel可以完成数据清洗，预处理，以及最常见的数据分类，数据筛选，分类汇总，以及数据透视等操作，而这些操作用SQL一样可以实现。SQL不仅可以从数据库中读取数据，还能通过不同的SQL函数语句直接返回所需要的结果，从而大大提高了自己在客户端应用程序中计算的效率。
1 重复数据处理
- 查找重复记录
SELECT * FROM user Where (nick_name,password) in ( SELECT nick_name,password FROM user group by nick_name,password having count(nick_name)>1 );
- 查找去重记录
查找id最大的记录
SELECT * FROM user WHERE id in (SELECT max(id) FROM user group by nick_name,password having count(nick_name)>1 );
- 删除重复记录
只保留id值最小的记录
DELETE c1 FROM customer c1,customer c2 WHERE c1.cust_email=c2.cust_email AND c1.id>c2.id;
DELETE FROM user Where (nick_name,password) in (SELECT nick_name,password FROM (SELECT nick_name,password FROM user group by nick_name,password having count(nick_name)>1) as tmp1 ) and id not in (SELECT id FROM (SELECT min(id) id FROM user group by nick_name,password having count(nick_name)>1) as tmp2 );
2 缺失值处理
- 查找缺失值记录
SELECT * FROM customer WHERE cust_email IS NULL;
- 更新列填充空值
UPDATE sale set city = "未知" WHERE city IS NULL; UPDATE orderitems set price_new=IFNULL(price_new,5.74);
- 查询并填充空值列
SELECT AVG(price_new) FROM orderitems; SELECT IFNULL(price_new,5.74) AS bus_ifnull FROM orderitems;
3 计算列
- 更新表添加计算列
ALTER TABLE orderitems ADD price_new DECIMAL(8,2) NOT NULL; UPDATE orderitems set price_new= item_price*count;
- 查询计算列
SELECT item_price*count as sales FROM orderitems;
4 排序
- 多列排序
SELECT * FROM orderitems ORDER BY price_new DESC,quantity;
- 查询排名前几的记录
SELECT * FROM orderitems ORDER BY price_new DESC LIMIT 5;
- 查询第10大的值
SELECT DISTINCT price_new FROM orderitems ORDER BY price_new DESC LIMIT 9,1;
- 排名
数值相同的排名相同且排名连续
SELECT prod_price, (SELECT COUNT(DISTINCT prod_price) FROM products WHERE prod_price>=a.prod_price ) AS rank FROM products AS a ORDER BY rank ;
5 字符串处理
- 字符串替换
UPDATE data1 SET city=REPLACE(city,'SH','shanghai'); SELECT city FROM data1;
- 按位置字符串截取
字符串截取可用于数据分列 MySQL 字符串截取函数：left(), right(), substring(), substring_index()
SELECT left('example.com', 3);
从字符串的第 4 个字符位置开始取，直到结束
SELECT substring('example.com', 4);
从字符串的第 4 个字符位置开始取，只取 2 个字符
SELECT substring('example.com', 4, 2);
- 按关键字截取字符串
取第一个分隔符之前的所有字符，结果是www
SELECT substring_index('www.google.com','.',1);
取倒数第二个分隔符之后的所有字符，结果是google.com;
SELECT substring_index('www.google.com','.',-2);
6 筛选
- 通过操作符实现高级筛选
使用 AND OR IN NOT 等操作符实现高级筛选过滤
SELECT prod_name，prod_price FROM Products WHERE vend_id IN('DLL01','BRS01'); SELECT prod_name FROM Products WHERE NOT vend_id='DLL01';
- 通配符筛选
常用通配符有% _ [] ^
SELECT * from customers WHERE country LIKE "CH%";
7 表联结
SQL表连接可以实现类似于Excel中的Vlookup函数的功能
SELECT vend_id,prod_name,prod_price FROM Vendors INNER JOIN Products ON Vendors.vend_id=Products.vend_id; SELECT prod_name,vend_name,prod_price,quantity FROM OderItems,Products,Vendors WHERE Products.vend_id=Vendors.vend_id AND OrderItems.prod_id=Products.prod_id AND order_num=20007;
自联结 在一条SELECT语句中多次使用相同的表
SELECT c1.cust_od,c1.cust_name,c1.cust_contact FROM Customers as c1,Customers as c2 WHERE c1.cust_name=c2.cust_name AND c2.cust_contact='Jim Jones';
8 数据透视
数据分组可以实现Excel中数据透视表的功能
- 数据分组
group by 用于数据分组 having 用于分组后数据的过滤
SELECT order_num,COUNT(*) as items FROM OrderItems GROUP BY order_num HAVING COUNT(*)>=3;
- 交叉表
通过CASE WHEN函数实现
SELECT data1.city, CASE WHEN colour = "A" THEN price END AS A, CASE WHEN colour = "B" THEN price END AS B, CASE WHEN colour = "C" THEN price END AS C, CASE WHEN colour = "F" THEN price END AS F FROM data1
