无法使用 LAG 功能

Question

我需要将表格更新为：

ID | START_DATE | response| FINAL_TREND

1    14-10-2021     4                       
1    15-10-2021     3                       
1    16-10-2021     2                       
1    17-10-2021     2                        
1    18-10-2021     3
1    19-10-2021     2

输出为：

ID | START_DATE | response| FINAL_TREND

1    14-10-2021     4           NULL            
1    15-10-2021     3            4           
1    16-10-2021     2            3           
1    17-10-2021     2            2            
1    18-10-2021     3            2
1    19-10-2021     2            3    

所以，在运行代码时：

SELECT LAG(RESPONSE,1) OVER (ORDER BY START_DATE) AS NEW
FROM DUMMY_YC 

输出为：

NULL
4
3
2
3
2
2

但是当在更新 AS 中使用相同的代码时：

UPDATE DUMMY_YC A SET A.RESPONSE = (SELECT LAG(B.RESPONSE,1) OVER (ORDER BY B.START_DATE) AS NEW
FROM DUMMY_YC B WHERE B.START_DATE=A.START_DATE)

输出为：

7 rows updated.

但实际更新的值是

RESPONSE|

(null)
(null)
(null)
(null)
(null)
(null)
(null)

我们将不胜感激。使用 Oracle SQL Developer。

Answer 1

我会选择merge。

之前：

SQL>   SELECT *
  2      FROM test
  3  ORDER BY id, start_date;

        ID START_DATE   RESPONSE FINAL_TREND
---------- ---------- ---------- -----------
         1 14.10.2021          4           0
         1 15.10.2021          3           0
         1 16.10.2021          2           0
         1 17.10.2021          2           0
         1 18.10.2021          3           0
         1 19.10.2021          2           0

6 rows selected.

合并：

SQL> MERGE INTO test a
  2       USING (SELECT b.id,
  3                     b.start_date,
  4                     b.response,
  5                     LAG (b.response, 1) OVER (ORDER BY b.start_date) AS final_trend
  6                FROM test b) x
  7          ON (    x.id = a.id
  8              AND x.start_date = a.start_date)
  9  WHEN MATCHED
 10  THEN
 11     UPDATE SET a.final_trend = x.final_trend;

6 rows merged.

之后：

SQL>   SELECT *
  2      FROM test
  3  ORDER BY id, start_date;

        ID START_DATE   RESPONSE FINAL_TREND
---------- ---------- ---------- -----------
         1 14.10.2021          4
         1 15.10.2021          3           4
         1 16.10.2021          2           3
         1 17.10.2021          2           2
         1 18.10.2021          3           2
         1 19.10.2021          2           3

6 rows selected.

SQL>