Sql 在多个查询中使用结果集

我想在相当大的基础上创建一个小型备份（这样开发者可以下载~1-2gb而不是15gb）

为了做到这一点，我复制了基，并运行了一些脚本来截断一些表（日志等），我想删除一些用户之外的数据（总是相同的用户）

现在我有这个：

-- delete order details not in test accounts
DELETE FROM order_details WHERE Album_ID NOT IN (
    SELECT Album_ID FROM albums WHERE User_ID IN (
        SELECT User_ID FROM users
            WHERE Email_ID LIKE '%@xxx.com'
            OR Email_ID LIKE '%@yyy.com'
            OR Email_ID LIKE '%@zzz.com'
    )
)
DELETE FROM orders WHERE User_ID NOT IN (
    SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'
)

-- delete albums not in test accounts
DELETE FROM albums WHERE User_ID NOT IN (
    SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'
)
-- snip a few more of the same

如您所见，我总是在许多地方使用相同的

从用户中选择用户ID，这些用户的电子邮件ID如“%@xxx.com”或电子邮件ID如“%@yyy.com”或…

子查询

你会怎么做，这样你就不必重复了

---

谢谢大家!

将结果存储在临时表中

   -- create temporary table to keep track of users
   CREATE TABLE #users
   (
      user_id BIGINT
   )

INSERT INTO #users (user_id)
SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'

IF OBJECT_ID('tempdb.dbo.#tmpUserID') IS NOT NULL DROP TABLE dbo.#tmpUserID
SELECT User_ID 
INTO dbo.#tmpUserID
FROM users
WHERE Email_ID LIKE '%@xxx.com'
  OR Email_ID LIKE '%@yyy.com'
  OR Email_ID LIKE '%@zzz.com'

-- delete order details not in test accounts
DELETE FROM order_details 
WHERE Album_ID NOT IN (
                       SELECT Album_ID 
                       FROM albums 
                       WHERE User_ID IN (
                                         SELECT User_ID FROM #tmpUserID
                                         )
                       )      
DELETE FROM orders 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

-- delete albums not in test accounts
DELETE FROM albums 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

加载临时表

   -- create temporary table to keep track of users
   CREATE TABLE #users
   (
      user_id BIGINT
   )

INSERT INTO #users (user_id)
SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'

IF OBJECT_ID('tempdb.dbo.#tmpUserID') IS NOT NULL DROP TABLE dbo.#tmpUserID
SELECT User_ID 
INTO dbo.#tmpUserID
FROM users
WHERE Email_ID LIKE '%@xxx.com'
  OR Email_ID LIKE '%@yyy.com'
  OR Email_ID LIKE '%@zzz.com'

-- delete order details not in test accounts
DELETE FROM order_details 
WHERE Album_ID NOT IN (
                       SELECT Album_ID 
                       FROM albums 
                       WHERE User_ID IN (
                                         SELECT User_ID FROM #tmpUserID
                                         )
                       )      
DELETE FROM orders 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

-- delete albums not in test accounts
DELETE FROM albums 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

然后可以在其他查询中使用临时表

-- delete order details not in test accounts
DELETE FROM order_details WHERE Album_ID NOT IN (
    SELECT Album_ID FROM albums WHERE User_ID IN (
        SELECT User_ID FROM #users
    )
)
DELETE FROM orders WHERE User_ID NOT IN (
    SELECT User_ID FROM #users
)

-- delete albums not in test accounts
DELETE FROM albums WHERE User_ID NOT IN (
    SELECT User_ID FROM #users
)

---

这项技术允许您遵循，因为您只在一个地方定义了查询。如果需要更改业务逻辑，只需更改一个查询，而不必更改原始代码中的3个查询。

这是一个非常好的问题。SQL不能提供很好的功能来组合SQL查询，这是不幸的

我能想到的最好的方法是视图或表值函数。您可以在脚本开始时创建它，然后在脚本结束时删除它

---

这是一个有点丑陋的黑客行为，在这种情况下，它似乎比动态构建SQL字符串要好。

创建一个表变量或临时表并存储Select查询数据

 Declare @User table
 (User_Id int)

 Insert into @User
  SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com' 

DELETE FROM order_details WHERE Album_ID NOT IN (
SELECT Album_ID FROM albums WHERE User_ID IN (
   Select User_Id from @User)
)

---

尝试将值插入临时变量，并在所有where条件下使用相同的值

DECLARE @table as TABLE(User_ID  Nvarchar(50))

INSERT INTO @table (User_ID)
SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'

---

使用临时表

   -- create temporary table to keep track of users
   CREATE TABLE #users
   (
      user_id BIGINT
   )

INSERT INTO #users (user_id)
SELECT User_ID FROM users
        WHERE Email_ID LIKE '%@xxx.com'
        OR Email_ID LIKE '%@yyy.com'
        OR Email_ID LIKE '%@zzz.com'

IF OBJECT_ID('tempdb.dbo.#tmpUserID') IS NOT NULL DROP TABLE dbo.#tmpUserID
SELECT User_ID 
INTO dbo.#tmpUserID
FROM users
WHERE Email_ID LIKE '%@xxx.com'
  OR Email_ID LIKE '%@yyy.com'
  OR Email_ID LIKE '%@zzz.com'

-- delete order details not in test accounts
DELETE FROM order_details 
WHERE Album_ID NOT IN (
                       SELECT Album_ID 
                       FROM albums 
                       WHERE User_ID IN (
                                         SELECT User_ID FROM #tmpUserID
                                         )
                       )      
DELETE FROM orders 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

-- delete albums not in test accounts
DELETE FROM albums 
WHERE User_ID NOT IN (
                      SELECT User_ID FROM #tmpUserID
                      )

---

如果在表上设置了外键，则可以将它们设置为级联删除。如果操作正确，只需从users表中删除，这将级联到所有其他表

---

关于cascade delete有大量文档，因此我不在这里重复，但它确实值得一看。

在这种情况下，我并不认为OP希望从Users表中删除，但cascade delete可以节省一些删除。在上面的例子中-订单详细信息可以与相册一起删除。是的。。。这是一个遗留数据库，几百个表有3个外键。很明显，添加FK会花费很长时间来清理数据库。好主意，但不可能执行：(