Question

从table1删除元组的最聪明的方法是什么，在第二个表中，如果第二个表不是初始数据库的一部分，而是一些非常大的查询结果？

table1           *this table is a result of some query
-------------    -------------
| id1 | id2 |    | id1 | id2 |
-------------    -------------
| 1      2  |    | 5      6  |
| 3      4  |    | 1      2  |
| 5      6  |    | 11     12 |
| 7      8  |    -------------
| 9      10 |    
| 11     12 |
| 13     14 |
-------------

我想出了

delete from table1
where id1 in (select id1 from (         really long query to get a second table))
  and id2 in (select id2 from (the same really long query to get a second table));

它有效，但我觉得我做得太错了，并没有保持查询DRY。

如果table1有一个额外的列，你建议的工作方式是否相同，例如“somecol”？

Answer 1

IMO，您可以使用EXISTS这样的语句：

DELETE FROM table1
WHERE EXISTS (
    SELECT 1
    FROM (<your long query>) AS dt
    WHERE table1.id1 = dt.id1
      AND table1.id2 = dt.id2);

<强> [SQL Fiddle Sample]

Answer 2

使用EXISTS的相关子查询允许匹配多个列：

delete 
from table1
where exists 
 ( select * from 
    (
      "really long query"
    ) as t2
   where table1.id1 = t2.id1 -- correlating inner and outer table
     and table1.id2 = t2.id2 -- similar to a join-condition
 )

Answer 3

一种方法是使用delete，exists和with secondtable as ( <your query here> ) delete from table1 where exists (select 1 from secondtable st where table1.id1 = st.id1 and table1.id2 = st.id2 );：

TextBox.Clear();  
TextBox.Text = string.Empty;

从表中删除子表，SQL

3 个答案: