sql - Oracle - 仅从列中选择最大值

Question

我的数据库的结构大致是这样的(简化):

三张表:订单、步骤、 Material

“订单”表包含列 ID 和 material

表“step”包含列ID、order 和stepnumber

“material”表包含列ID 和description

            Order                         
    --------------------------------    
     **ID** **Number**     **Material**    
       1         X1            11
       2         X2            12
       3         X3            13


              Step
   ---------------------------------------
   **ID**    **Order**  **Stepnumber***
     1          X1           X110
     2          X1           X120
     3          X1           X170
     4          X1           X180
     5          X2           X270
     6          X2           X280

                 Material
   ---------------------------------------
  **ID**      **Description***
    11             Mat1
    12             Mat2
    13             Mat3

需要注意的是，表的 stepnumber 列由订单号(例如 X1)和标识步骤的数字(10、20、70、80 等)组成.)，如您所见，stepnumber 列中有多个值指向同一订单。

我正在使用这个查询:

select order.number,step.stepnumber, material.id,material.description, 
from db.order inner join db.stepnumber
on order.number = step.order
inner join db.material
on material.id = order.material
where step.stepnumber not like '%10'
and step.stepnumber not like '%20'
group by order.number, step.stepnumber, material.id, material.description

正如您从查询中看到的，我想提取订单和步骤数据，并且我根据 stepnumber 值排除了一些步骤。

现在，查询输出是:

order.number step.stepnumber material.id material.description 
-------------------------------------------------------------
  X1             X170             11           Mat1
  X1             X180             11           Mat1
  X2             X270             12           Mat2
  X2             X280             12           Mat2

我想实现的是，对于每个订单，只提取一行，取 stepnumber 值较高的那一行。所以结果应该是:

order.number step.stepnumber material.id material.description 
-------------------------------------------------------------
  X1             X180             11           Mat1
  X2             X280             12           Mat2

我已经尝试将 MAX 函数放入选择中的 step.stepnumber 但没有效果。

Answer 1

你可以尝试使用ROW_NUMBER窗口函数

CREATE TABLE "Order"(
   ID int,
   "Number" varchar(50),
   Material int
);


INSERT INTO  "Order" VALUES (1,'X1',11);
INSERT INTO  "Order" VALUES (2,'X2',12);
INSERT INTO  "Order" VALUES (3,'X3',13);

CREATE TABLE Step(
   ID int,
   "Order" varchar(50),
   Stepnumber varchar(50)
);


insert into Step values (1,'X1','X110');
insert into Step values (2,'X1','X120');
insert into Step values (3,'X1','X170');
insert into Step values (4,'X1','X180');
insert into Step values (5,'X2','X270');
insert into Step values (6,'X2','X280');

CREATE TABLE Material(
   ID int,
   Description varchar(50)
);               


INSERT INTO Material VALUES (11,'Mat1');
INSERT INTO Material VALUES (12,'Mat2');
INSERT INTO Material VALUES (13,'Mat3');

查询 1:

SELECT "Number",STEPNUMBER,ID,DESCRIPTION 
FROM (
  select o."Number",
           s.stepnumber, 
           m.id,
           m.description,
           ROW_NUMBER() OVER(PARTITION BY o."Number" ORDER BY s.stepnumber DESC) rn
    from db."Order" o
    inner join db.Step s on o."Number" = s."Order"
    inner join db.material m on m.id = o.material
    where 
        s.stepnumber not like '%10'
    and 
        s.stepnumber not like '%20'
) t1
WHERE rn = 1

Results :

| Number | STEPNUMBER | ID | DESCRIPTION |
|--------|------------|----|-------------|
|     X1 |       X180 | 11 |        Mat1 |
|     X2 |       X280 | 12 |        Mat2 |

注意

我们可能会尽量避免使用Order，Number 是列名或表名...因为它们是Oracle 中的关键字。