Question

我有一张桌子（在现实生活中，这是一个复杂的查询，可为我提供此结果）。每行代表供应商不同的期限承诺：

--------------------------------------------------------------------------
| supplier | days_left | prod_type | prods_to_deliver | prods_in_transit |
--------------------------------------------------------------------------
|      1   |    20     |     1     |         9        |         7        |
--------------------------------------------------------------------------
|      1   |    20     |     2     |         5        |         3        |
--------------------------------------------------------------------------
|      1   |     5     |   NULL    |         7        |         4        |
--------------------------------------------------------------------------
|      1   |    20     |   NULL    |         9        |         6        |
--------------------------------------------------------------------------

棘手的是，每个供应商最多可以有两行， prod_type = NULL ，每行具有不同的 days_left 值。对于 prods_in_transit ，最高行也总是包含最低行的值，我必须对其进行更正。

用（H）标记最高 days_left 值行，并用（L）标记最低值行。

我正在寻找一个SELECT，它将以以下方式改变我的结果：如果同时存在（H）和（L）行，则对于（H）行 prods_in_transit（H）= prods_in_transit （H）-prods_in_transit（L）

输出应如下所示：

--------------------------------------------------------------------------
| supplier | days_left | prod_type | prods_to_deliver | prods_in_transit |
--------------------------------------------------------------------------
|      1   |    20     |     1     |         9        |         7        |
--------------------------------------------------------------------------
|      1   |    20     |     2     |         5        |         3        |
--------------------------------------------------------------------------
|      1   |     5     |   NULL    |         7        |         4        |
--------------------------------------------------------------------------
|      1   |    20     |   NULL    |         9        |       >>2<<      |
--------------------------------------------------------------------------

对于最后一行的 prods_in_transit 值，应计算为 6-4 = 2

这应该是一个SQL查询，因为它将在大量不同的供应商上运行。我的示例仅说明了一个片段。用PHP foreach遍历每个供应商的结果也不是一种选择，因为这将花费很多时间。

理想情况下，我已经给了我一些代码，但实际上，我不知道从哪里开始。

谢谢。

Answer 1

DROP TABLE IF EXISTS T;
CREATE TABLE T (SUPPLIERID INT, PRODS_TYPE INT, days_left int, PRODS_IN_TRANSIT INT);
INSERT INTO T VALUES
(1,1,1,2),(1,2,2,7),(1,NULL,5,4),(1,NULL,20,6),
(2,1,4,1),(2,NULL,7,10),
(3,1,6,3)
; 

SELECT SUPPLIERID,PRODS_TYPE,days_left,PRODS_IN_TRANSIT FROM T WHERE PRODS_TYPE IS NOT NULL
UNION ALL
SELECT SUPPLIERID,PRODS_TYPE,days_left,prods_in_transit 
        FROM T 
        WHERE PRODS_TYPE IS NULL 
        and prods_in_transit = 
        (select min(prods_in_transit) from t t1 where t1.supplierid = t.supplierid and t1.prods_type is null)
UNION ALL
(
select t.supplierid,t.prods_type,t.days_left,p
from t join
(
SELECT SUPPLIERID,PRODS_TYPE,max(prods_in_transit) maxint,MAX(PRODS_IN_TRANSIT) - MIN(PRODS_IN_TRANSIT) P 
        FROM T
        WHERE PRODS_TYPE IS NULL 
        GROUP BY SUPPLIERID,PRODS_TYPE HAVING COUNT(*) > 1 
) t1 on t1.SUPPLIERID = t.SUPPLIERID and t1.maxint = t.prods_in_transit
) 

ORDER BY SUPPLIERID;


+------------+------------+-----------+------------------+
| SUPPLIERID | PRODS_TYPE | days_left | PRODS_IN_TRANSIT |
+------------+------------+-----------+------------------+
|          1 |          1 |         1 |                2 |
|          1 |          2 |         2 |                7 |
|          1 |       NULL |         5 |                4 |
|          1 |       NULL |        20 |                2 |
|          2 |          1 |         4 |                1 |
|          2 |       NULL |         7 |               10 |
|          3 |          1 |         6 |                3 |
+------------+------------+-----------+------------------+
7 rows in set (0.00 sec)

如果在同一查询的另一结果行中满足条件，是否有一种方法可以减少一个结果行中的值

1 个答案: