Question

Table:T1 (key, numval, state) and T2(key, foreignkey:T1.key, type)

我有两个表T1和T2，每个T1都有多个值   T2。对于T2中的每个值，T1中都有一个值。

考虑到我在T2中有两种类型（'A'，'B'），我必须计算映射到每个T1键的T2的数量，状态为'A'和'B'并将其设置为T1 .state分别

如何在Oracle中完成此任务？

与下面的帖子类似（但解决方案不起作用）： how to update a column in a table with unique values, but not all the rows, not incremented

使用上面的源代码 - 我使用了以下脚本：

UPDATE T1
SET VALUE =
   CASE
     WHEN (SELECT STATE FROM T1 WHERE KEY = EACH_KEY) = 'S1' 
                THEN 
                    (SELECT COUNT(*) FROM T2 
                                    WHERE KEY = EACH_KEY AND TYPE = 'A' 
     WHEN (SELECT STATE FROM T1 WHERE KEY = EACH_KEY) = 'S2' 
                THEN 
                    (SELECT COUNT(*) FROM T2 
                                    WHERE KEY = EACH_KEY AND TYPE = 'B' 
   END
WHERE  EACH_KEY in (select distinct KEY from T1);

但是oracle会抛出以下错误：

Error at Command Line : 13 Column : 8
Error report -
SQL Error: ORA-00904: "EACH_KEY": invalid identifier
00904. 00000 -  "%s: invalid identifier"
*Cause:    
*Action:

编辑＃1：

在表格中采样数据

执行前

T1
key    numval    state
1      0         S1
2      0         S2
3      0         S1
4      0         S1

T2
key    t1key     type
1      1         A
2      1         B
3      1         B
4      2         A
5      2         B
6      2         B
7      3         B
8      4         A
9      4         A
10      4         B

执行后（预期结果）

T1
key    numval    state
1      1         S1
2      2         S2
3      0         S1
4      2         S2

编辑＃2：

是的，未定义标识符。我想知道如何在Oracle SQL中定义这样的标识符，以便可以执行上述计算。

Answer 1

以下是支持@ jva答案的测试用例：

使用数据创建表格：

create table t1 as
select 1 key, 0 numval, 'S1' state from dual union all
select 2 key, 0 numval, 'S2' state from dual union all
select 3 key, 0 numval, 'S1' state from dual union all
select 4 key, 0 numval, 'S1' state from dual;

create table t2 as
select 1 key, 1 t1key, 'A' type from dual union all
select 2 key, 1 t1key, 'B' type from dual union all
select 3 key, 1 t1key, 'B' type from dual union all
select 4 key, 2 t1key, 'A' type from dual union all
select 5 key, 2 t1key, 'B' type from dual union all
select 6 key, 2 t1key, 'B' type from dual union all
select 7 key, 3 t1key, 'B' type from dual union all
select 8 key, 4 t1key, 'A' type from dual union all
select 9 key, 4 t1key, 'A' type from dual union all
select 10 key, 4 t1key, 'B' type from dual;

检查t1中的初始值：

select * from t1 order by key;

       KEY     NUMVAL STATE
---------- ---------- -----
         1          0 S1   
         2          0 S2   
         3          0 S1   
         4          0 S1

执行更新：

update t1
set    numval = case when state = 'S1' then (select count(*) from t2 where t2.t1key = t1.key and t2.type = 'A')
                     when state = 'S2' then (select count(*) from t2 where t2.t1key = t1.key and t2.type = 'B')
                end;

commit;

检查t1中的新值：

select * from t1 order by key;

       KEY     NUMVAL STATE
---------- ---------- -----
         1          1 S1   
         2          2 S2   
         3          0 S1   
         4          2 S1

如您所见，这是您想要的输出。

来自＆＃34;缺陷＆＃34;你在@ jva的回答评论中提到过，我觉得你很困惑：

在答案中使用相关子查询，
更新如何与相关子查询一起使用。

我强烈建议您查看the documentation，具体来说：

当嵌套子查询引用从子查询上一级父级语句引用的表中的列时，Oracle会执行相关子查询。父语句可以是SELECT，UPDATE或DELETE语句，子查询嵌套在该语句中。对于父语句处理的每一行，概念上对相关子查询进行一次计算。

Answer 2

UPDATE T1
SET numval = CASE WHEN t1.state = 'S1' THEN (SELECT COUNT(*) 
                                             FROM t2
                                             WHERE t1.key = t2.t1key
                                               AND type = 'A')
                  WHEN t1.state = 'S2' THEN (SELECT COUNT(*) 
                                             FROM t2
                                             WHERE t1.key = t2.t1key
                                               AND type = 'B')
             END

我们如何在Oracle中执行多个更新

2 个答案: