匹配时加入，否则使用默认值答案

【问题标题】：Join when match other wise use default value匹配时加入，否则使用默认值
【发布时间】：2016-01-18 21:21:25
【问题描述】：

我有两张桌子一个

C1  C2  Value
All All  5
da  All  6

和表 b

C1  C2  
cc  a  
da  b

我想获取表 b 的匹配值

C1  C2  Value
cc  a   5
da  b   6

我这样写一个查询

select b.*,a.* from a, b 
where 1=(case when a.C1=b.C1 then 1 
         when a.C1='All' then 1
         else 0)
and 1=(case when a.C2=b.C2 then 1 
         when a.C2='All' then 1
         else 0)

得到结果

C1  C2  Value
cc  a   5
da  b   6    
da  b   5

我怎样才能得到查询以得到如下结果

C1  C2  Value
cc  a   5
da  b   6

基本上当有匹配时使用匹配值，否则使用默认值'All'，非常感谢！

【问题讨论】：

表 A 是否有可能在 C1 中有一行“All”，而在 C2 中有一个类似“b”的值？如果是这样，在这种情况下结果会如何？
您的语句不会执行 - 您需要为每个 CASE 表达式添加一个 END。

标签： sql oracle join

【解决方案1】：

你需要的输出对我来说不是很清楚，但也许你需要这样的东西：

select c1, c2 , value from
(
 select b.c1, b.c2, a.value,
 row_number() over ( partition by decode(a.c1, 'all', 0, 1) + decode(a.c2, 'all', 0, 1)
                     order by     decode(a.c1, 'all', 0, 1) + decode(a.c2, 'all', 0, 1) desc
                   ) as num
 from a join b
 on ( a.c1 = 'all' OR a.c1 = b.c1) and 
    ( a.c2 = 'all' OR a.c2 = b.c2)
)
where num = 1;

希望对你有帮助

【讨论】：

【解决方案2】：

有些不寻常的连接条件。使用合并的左外连接可能是获得所需内容的简单方法，前提是性能不受限制。

select b.*, COALESCE(a1.Value, a2.Value, a3.Value, a4.Value)
from b LEFT OUTER JOIN a a1
ON a1.C1=b.C1 AND a1.C2=b.C2
LEFT OUTER JOIN a a2
ON a2.C1=b.C1 AND a2.C2='All'
LEFT OUTER JOIN a a3
ON a3.C1='All' AND a3.C2=b.C2
LEFT OUTER JOIN a a4
ON a4.C1='All' AND a4.C2='All'

请注意，如果有，您将获得多行：

多条记录包含All、All；
a 中的 C1、C2 列有多个匹配项；

【讨论】：

【解决方案3】：

Oracle 设置：

CREATE TABLE table_a ( C1, C2, Value ) AS
SELECT 'All', 'All', 5 FROM DUAL UNION ALL
SELECT 'da',  'All', 6 FROM DUAL UNION ALL
SELECT 'da',  'b', 4 FROM DUAL;

CREATE TABLE table_b ( C1, C2 ) AS
SELECT 'cc', 'a' FROM DUAL UNION ALL
SELECT 'da', 'b' FROM DUAL UNION ALL
SELECT 'da', 'c' FROM DUAL;

查询：

SELECT b.C1,
       b.C2,
       ( SELECT MAX ( value ) KEEP (
                    DENSE_RANK FIRST
                    ORDER BY 
                      CASE WHEN b.C1 = a.C1 AND b.C2 = a.C2 THEN 1
                           WHEN b.C1 = a.C1                 THEN 2
                           ELSE                                  3
                           END ASC
                  )
         FROM   table_a a
         WHERE  ( b.C1 = a.C1  AND b.C2 = a.C2  )
         OR     ( b.C1 = a.C1  AND a.C2 = 'All' )
         OR     ( a.C1 = 'All' AND a.C2 = 'All' )
       ) AS Value
FROM   table_b b;

结果：

C1 C2      VALUE
-- -- ----------
cc a           5 
da b           4 
da c           6

查询 2：

SELECT b.C1,
       b.C2,
       MAX ( value ) KEEP (
            DENSE_RANK FIRST
            ORDER BY 
              CASE WHEN b.C1 = a.C1 AND b.C2 = a.C2 THEN 1
                   WHEN b.C1 = a.C1                 THEN 2
                   ELSE                                  3
                   END ASC
       ) AS Value
FROM   table_b b
       INNER JOIN
       table_a a
ON         ( b.C1 = a.C1  AND b.C2 = a.C2  )
       OR  ( b.C1 = a.C1  AND a.C2 = 'All' )
       OR  ( a.C1 = 'All' AND a.C2 = 'All' )
GROUP BY b.C1, b.C2;

查询 3：

SELECT C1,
       C2,
       Value
FROM   (       
  SELECT b.C1,
         b.C2,
         value,
         ROW_NUMBER() OVER (
           PARTITION BY b.C1, b.C2
           ORDER BY  CASE WHEN b.C1 = a.C1 AND b.C2 = a.C2 THEN 1
                          WHEN b.C1 = a.C1                 THEN 2
                          ELSE                                  3
                          END ASC
         ) AS RN
  FROM   table_b b
         INNER JOIN
         table_a a
  ON         ( b.C1 = a.C1  AND b.C2 = a.C2  )
         OR  ( b.C1 = a.C1  AND a.C2 = 'All' )
         OR  ( a.C1 = 'All' AND a.C2 = 'All' )
)
WHERE  rn = 1;

所有结果都应该相同。

【讨论】：

【解决方案4】：

试试这个：

    SELECT b.C1,
           b.C2,
           max(value) value 
      FROM a 
INNER JOIN b ON (DECODE (a.C2, 'All', b.C2,a.C2) = b.C2 AND DECODE (a.C1, 'All', b.C1,a.C1) = b.C1)
  GROUP BY b.C1,
           b.C2

【讨论】：