在 MySQL 中进行重复密钥更新

Question

假设我有 2 tables

CREATE TABLE shipping(id int(10) NOT NULL PRIMARY KEY AUTO_INCREMENT, origin INT(10),
order_id int(10), createdAt date NOT NULL, product_id varchar(100) NOT NULL,
amount decimal(10,0));

CREATE TABLE reprint (id int(10) NOT NULL PRIMARY KEY AUTO_INCREMENT,
product_id int(50), reprintAt varchar(255), status tinyint(4));


 



INSERT INTO shipping (origin, order_id, createdAt, product_id, amount)
VALUES (1, 11, NOW(), 234, 2000),
(1, 11, NOW(), 235, 3000),
(1, 11, NOW(), 236, 4000),
(2, 12, NOW(), 236, 3000),
(2, 12, NOW(), 235, 2100),
(3, 13, NOW(), 236, 2200),
(3, 13, NOW(), 239, 3400),
(4, 14, NOW(), 237, 2300),
(4, 14, NOW(), 233, 4000);


INSERT INTO reprint (product_id, reprintAt, status)
VALUES (234, NOW() + INTERVAL 1 DAY, 1),
(235, NOW() + INTERVAL 2 DAY, 1),
(236, NOW(), 1),
(237, NOW(), 1),
(238, NOW(), 1),
(239, NOW(), 1),
(240, NOW(), 1),
(233, NOW(), 1);

table 运费

id	origin	order_id	createdAt	product_id	amount
1	1	11	2020-12-22	234	2000
2	1	11	2020-12-22	235	3000
3	1	11	2020-12-22	236	4000
4	2	12	2020-12-22	236	3000
5	2	12	2020-12-22	235	2100
6	3	13	2020-12-22	236	2200
7	3	13	2020-12-22	239	3400
8	4	14	2020-12-22	237	2300
9	4	14	2020-12-22	233	4000

table转载

id	product_id	reprintAt	status
1	234	2020-12-23 10:38:02	1
2	235	2020-12-24 10:38:02	1
3	236	2020-12-22 10:38:02	1
4	237	2020-12-22 10:38:02	1
5	238	2020-12-22 10:38:02	1
6	239	2020-12-22 10:38:02	1
7	240	2020-12-22 10:38:02	1
8	233	2020-12-22 10:38:02	1

所以有2个table，同一个product_id。如果转载和发货的product_id个数不一样，则将一些包含缺失product_id的记录放入发货。在这种情况下，product_Id '238' 和 '240' 不包含在 table 运费中，因此我必须在运费中插入包含 product_id '238' 和 '240' 的新记录table.

对于这些要求，预期的结果应该是这样的（请查看 product_id 238 和 240 table）

id	origin	order_id	createdAt	product_id	amount
1	1	11	2020-12-22	234	2000
2	1	11	2020-12-22	235	3000
3	1	11	2020-12-22	236	4000
4	2	12	2020-12-22	236	3000
5	2	12	2020-12-22	235	2100
6	3	13	2020-12-22	236	2200
7	3	13	2020-12-22	239	3400
8	4	14	2020-12-22	237	2300
9	4	14	2020-12-22	233	4000
10	5	16	2020-12-22	238	4000
11	7	16	2020-12-22	240	4000

我正在尝试这个查询

DECLARE test_1 INT DEFAULT 0;
DECLARE test_2 INT DEFAULT 0;

SELECT count(distinct product_id) INTO test_1 FROM shipping;

SELECT count(distinct product_id) INTO test_2 FROM reprint;

IF test_1 != test_2 THEN

INSERT INTO shipping (id, origin, order_id, createdAt, product_id, amount) 
SELECT LAST_INSERT_ID(), reprint.id, 16, reprint.reprintAt, reprint.product_id, 4000 
FROM
    reprint

    UNION 
       SELECT * FROM shipping
    
ON duplicate key update product_id = values(product_id);

但是它说

You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'DECLARE test_1 INT DEFAULT 0;
DECLARE test_2 INT DEFAULT 0;

SELECT count(distin' at line 1

这是fiddlehttps://dbfiddle.uk/?rdbms=mysql_5.7&fiddle=c489fdb3c6c07bb597bf23928ab97397

Answer 1

如果我没看错的话，你可以直接用not exists:

insert into shipping (origin, order_id, createdat, product_id, amount) 
select id, 16, reprintat, product_id, 4000
from reprint r
where not exists (select 1 from shipping s where s.product_id = r.product_id)

这会为转载中的每一行创建一个新的运输行，其产品在目标中尚不存在，数量和订单 ID 不变。

我认为 on duplicate key 在这种情况下没有意义，因为产品不是唯一的列。

Answer 2

我认为您的问题可以通过简单的 SQL 查询使用 NOT EXISTS 语句解决：

INSERT INTO shipping (origin, order_id, createdAt, product_id, amount) 
SELECT reprint.id, 16, reprint.reprintAt, reprint.product_id, 4000 
FROM reprint
WHERE NOT EXISTS (
    SELECT product_id FROM shipping WHERE shipping.product_id = reprint.product_id
);

这里是fiddle

在 MySQL 中进行重复密钥更新

put on duplicate key update in MySQL

mysql

sql

store

subquery

sql-insert