Sql 透视表：如果其中一个参数为空，则连接同一条记录

我有这个国家标签表：

Country_Code | Product_Code | Label_Name | Placement IT PR_I01 T-Label Top Left IT PR_I01 Instruction Bottom Center HK PR_H34 Tag Bottom Left HK HK Label Bottom Right US PR_U12 Sticker Bottom Center GB PR_G16 QR Code Bottom Right 如何达到这样的效果？

---

谢谢。

您可以通过使用

MAX（）

对所有列进行聚合，但

国家/地区代码

列除外，该列在从

分区依据

列表中删除

产品

后应位于主查询中的

分组依据

中，例如

SELECT Country_Code, 
       MAX(product_code) AS Product_Code,
       MAX("1_LN") AS Label_Name1,
       MAX("1_LP") AS Placement1,
       MAX("2_LN") AS Label_Name2,
       MAX("2_LP") AS Placement2
  FROM (WITH CNTRYRWS AS 
       (
        SELECT Country_Code,
               PRODUCT_CODE,
               LABEL_NAME,
               PLACEMENT,
               ROW_NUMBER() 
               OVER(PARTITION BY Country_Code--, Product_Code 
                        ORDER BY LABEL_NAME, PLACEMENT) RN
          FROM PL_COUNTRY_LABEL_V
         )
        SELECT *
          FROM CNTRYRWS 
         PIVOT 
         ( MIN (LABEL_NAME) LN , 
           MIN (PLACEMENT) LP FOR RN IN ( 1 AS "1" , 2 AS "2") ))
GROUP BY Country_Code

---

嗯。条件聚合似乎是一种简单的方法：

select country_code, max(product_code) as product_code,
       max(case when seqnum = 1 then label_name end) as label_name_1,
       max(case when seqnum = 1 then placement end) as placement_1,
       max(case when seqnum = 2 then label_name end) as label_name_2,
       max(case when seqnum = 2 then placement end) as placement_2
from (select c.*,
             row_number() over (partition by country_code
                                order by product_code nulls last, label_name
                               ) as seqnum
      from PL_COUNTRY_LABEL_V c
     ) c
group by country_code

---

谢谢，但是如果我有多个产品代码与相同的国家代码，比如说我添加了新的香港标签与不同的产品代码，我不能使用国家代码和产品代码作为参数来调用它，因为我猜它最大。@Syns。这假定您知道结果集中的列数。如果不使用，则必须使用动态SQL。谢谢，我将了解一下动态SQL，因为它尚未使用。谢谢，如果有多个产品代码具有相同的国家/地区代码，该怎么办？它总是返回最大值。嗨@Syns。只要每个国家的产品代码最多是两个，就像你的情况一样（rn=1和2），那么就不会出现问题，正如你所看到的国家

HK

和

IT

一样。嗨@Ozhan，这就是问题所在，我每个国家的产品代码会多得多，谢谢you@Syns让它充满活力不是问题，但你的案子越来越复杂了。最初，您需要删除空

产品代码

值的额外行，现在每个国家/地区可以有多个

产品代码

（假定为非空）。在这种情况下，您希望如何将非空产品代码与空产品代码放在一起？？

SELECT CODE_COUNTRY,
       PRODUCT_CODE,
       "1_LN" Label_Name1,
       "1_LP" Placement1,
       "2_LN" Label_Name2,
       "2_LP" Placement2
FROM   (WITH CNTRYRWS
               AS (SELECT   CODE_COUNTRY,
                            PRODUCT_CODE,
                            LABEL_NAME,
                            PLACEMENT,                                
                            ROW_NUMBER ()
                               OVER (
                                  PARTITION BY CODE_COUNTRY, product_code
                                  ORDER BY LABEL_NAME, PLACEMENT
                               )
                               RN
                     FROM   PL_COUNTRY_LABEL_V)
        SELECT   *
          FROM   CNTRYRWS PIVOT ( MIN (LABEL_NAME) LN , MIN (
                          PLACEMENT) LP FOR RN IN ( 1 AS "1" , 2
                          AS "2") ))

SELECT Country_Code, 
       MAX(product_code) AS Product_Code,
       MAX("1_LN") AS Label_Name1,
       MAX("1_LP") AS Placement1,
       MAX("2_LN") AS Label_Name2,
       MAX("2_LP") AS Placement2
  FROM (WITH CNTRYRWS AS 
       (
        SELECT Country_Code,
               PRODUCT_CODE,
               LABEL_NAME,
               PLACEMENT,
               ROW_NUMBER() 
               OVER(PARTITION BY Country_Code--, Product_Code 
                        ORDER BY LABEL_NAME, PLACEMENT) RN
          FROM PL_COUNTRY_LABEL_V
         )
        SELECT *
          FROM CNTRYRWS 
         PIVOT 
         ( MIN (LABEL_NAME) LN , 
           MIN (PLACEMENT) LP FOR RN IN ( 1 AS "1" , 2 AS "2") ))
GROUP BY Country_Code

select country_code, max(product_code) as product_code,
       max(case when seqnum = 1 then label_name end) as label_name_1,
       max(case when seqnum = 1 then placement end) as placement_1,
       max(case when seqnum = 2 then label_name end) as label_name_2,
       max(case when seqnum = 2 then placement end) as placement_2
from (select c.*,
             row_number() over (partition by country_code
                                order by product_code nulls last, label_name
                               ) as seqnum
      from PL_COUNTRY_LABEL_V c
     ) c
group by country_code