Oracle查询以获取同一表中具有重复列集的所有行答案

【问题标题】：Oracle query to get all rows with duplicate set of columns in the same tableOracle查询以获取同一表中具有重复列集的所有行
【发布时间】：2020-10-15 13:49:00
【问题描述】：

我有一个 oracle 数据库表，其中包含一组重复 5 次的列。就像在下面的示例中，列（COL1_A 到 COL1_E）将在同一个表中存在 5 次，但列的名称除外，每组列中的数据类型将相同，但在一组（COL1_A 到 COL1_E）中确实如此没有相同的数据类型。

Cli_Id,Country,Name,COL1_A,COL1_B,COL1_C,COL1_D,COL1_E,COL2_A,COL2_B,COL2_C,COL2_D,COL2_E,COL3_A COL3_B,..
1      Test1   Lo1     1      2       3       4     5      1      2      3      4     5  
2      Test2   Lo2     5      6       7       8     9      5      3      3      4     5
3      Test3   Lo3     10     20      30      40   50      5      3      3      4     5
4      Test4   Lo4     11     22      32      42   52      53     3      3      4     5
5      Test5   Lo5     11     22      32      42   52      11     22      32      42   52

我需要编写一个查询来检索在列集之间包含重复项的所有行。

预期的结果

Cli_Id,Country,Name,COL1_A,COL1_B,COL1_C,COL1_D,COL1_E,COL2_A,COL2_B,COL2_C,COL2_D,COL2_E,COL3_A COL3_B,..
1      Test1   Lo1     1      2       3       4     5      1      2      3      4     5  
5      Test5   Lo5     11     22      32      42   52      11     22      32      42   52

我写了一个有几个联合但可能会影响性能，因为我们打算将集合增加到 8。因此寻找一种更简单的方法来编写这个查询。

您能否就更简单的查询提出建议，记住它应该是高效的，并且运行时间不会太长。

Select distinct CLi_id,Country,Name
from Table1
Where COL1_A = COL2_A
AND COL1_B = COL2_B
AND COL1_C = COL2_C
AND COL1_D = COL2_D
AND COL1_E = COL2_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL1_A = COL3_A
AND COL1_B = COL3_B
AND COL1_C = COL3_C
AND COL1_D = COL3_D
AND COL1_E = COL3_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL1_A = COL4_A
AND COL1_B = COL4_B
AND COL1_C = COL4_C
AND COL1_D = COL4_D
AND COL1_E = COL4_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL1_A = COL5_A
AND COL1_B = COL5_B
AND COL1_C = COL5_C
AND COL1_D = COL5_D
AND COL1_E = COL5_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL2_A = COL3_A
AND COL2_B = COL3_B
AND COL2_C = COL3_C
AND COL2_D = COL3_D
AND COL2_E = COL3_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL2_A = COL4_A
AND COL2_B = COL4_B
AND COL2_C = COL4_C
AND COL2_D = COL4_D
AND COL2_E = COL4_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL2_A = COL5_A
AND COL2_B = COL5_B
AND COL2_C = COL5_C
AND COL2_D = COL5_D
AND COL2_E = COL5_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL3_A = COL4_A
AND COL3_B = COL4_B
AND COL3_C = COL4_C
AND COL3_D = COL4_D
AND COL3_E = COL4_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL3_A = COL5_A
AND COL3_B = COL5_B
AND COL3_C = COL5_C
AND COL3_D = COL5_D
AND COL3_E = COL5_E
UNION
Select distinct CLi_id,Country,Name
from Table1
Where COL4_A = COL5_A
AND COL4_B = COL5_B
AND COL4_C = COL5_C
AND COL4_D = COL5_D
AND COL4_E = COL5_E

【问题讨论】：

标签： sql oracle duplicates oracle-sqldeveloper

【解决方案1】：

您可以使用OR 和IN，如下所示：

Select distinct CLi_id,Country,Name
from Table1
Where (COL1_A, COL1_B, COL1_C, COL1_D, COL1_E) 
      IN ((COL2_A, COL2_B, COL2_C, COL2_D, COL2_E), 
          (COL3_A, COL3_B, COL3_C, COL3_D, COL3_E), 
          (COL4_A, COL4_B, COL4_C, COL4_D, COL4_E),
          (COL5_A, COL5_B, COL5_C, COL5_D, COL5_E))
OR (COL2_A, COL2_B, COL2_C, COL2_D, COL2_E) 
      IN ((COL3_A, COL3_B, COL3_C, COL3_D, COL3_E), 
          (COL4_A, COL4_B, COL4_C, COL4_D, COL4_E),
          (COL5_A, COL5_B, COL5_C, COL5_D, COL5_E))
OR (COL3_A, COL3_B, COL3_C, COL3_D, COL3_E) 
      IN ((COL4_A, COL4_B, COL4_C, COL4_D, COL4_E),
          (COL5_A, COL5_B, COL5_C, COL5_D, COL5_E))
OR (COL4_A, COL4_B, COL4_C, COL4_D, COL4_E) 
      IN ((COL5_A, COL5_B, COL5_C, COL5_D, COL5_E))

【讨论】：

【解决方案2】：

由于您是在比较与识别属性的交叉级别对应的列级别值，因此您可以乐观地使用UNION ALL。无需加入您在查询中提到的列。

以下查询将为您提供行级属性，

select Cli_Id,Country,Name, col_A, col_B, col_C, col_D, count(*) from 
(select Cli_Id,Country,Name,COL1_A as col_A,COL1_B as col_B,COL1_C as col_C,COL1_D as col_D,COL1_E as col_E from table
union all
select Cli_Id,Country,Name,COL2_A as col_A,COL2_B as col_B,COL2_C as col_C,COL2_D  as col_D,COL2_E as col_E from table
union all
select Cli_Id,Country,Name,COL3_A as col_A,COL3_B as col_B,COL3_C as col_C,COL3_D  as 
col_D,COL3_E as col_E from table
union all
select Cli_Id,Country,Name,COL4_A as col_A,COL4_B as col_B,COL4_C as col_C,COL4_D  as col_D,COL4_E as col_E from table
union all
select Cli_Id,Country,Name,COL5_A as col_A,COL5_B as col_B,COL5_C as col_C,COL5_D  as col_D,COL5_E as col_E from table)
group by Cli_Id,Country,Name, col_A, col_B, col_C, col_D having count(*) > 1;

以下查询将为您提供列级属性，

select distinct t1.* from table t1
inner join 
(select Cli_Id,Country,Name, col_A, col_B, col_C, col_D, count(*) from 
(select Cli_Id,Country,Name,COL1_A as col_A,COL1_B as col_B,COL1_C as col_C,COL1_D as col_D,COL1_E as col_E from table
union all
select Cli_Id,Country,Name,COL2_A as col_A,COL2_B as col_B,COL2_C as col_C,COL2_D  as col_D,COL2_E as col_E from table
union all
select Cli_Id,Country,Name,COL3_A as col_A,COL3_B as col_B,COL3_C as col_C,COL3_D  as col_D,COL3_E as col_E from table
union all
select Cli_Id,Country,Name,COL4_A as col_A,COL4_B as col_B,COL4_C as col_C,COL4_D  as col_D,COL4_E as col_E from table
union all
select Cli_Id,Country,Name,COL5_A as col_A,COL5_B as col_B,COL5_C as col_C,COL5_D  as col_D,COL5_E as col_E from table
group by Cli_Id,Country,Name, col_A, col_B, col_C, col_D having count(*) > 1) t2
on (t1.Cli_Id = t2.Cli_Id);

【讨论】：

很抱歉，但不知道为什么我会收到缺少左括号的错误
请立即尝试。编辑查询

【解决方案3】：

Unpivot 数据，按列和值分组，取这个 id，其中 min(count) 等于序列数（在您的示例 3 中，将当前数据更改为 5，将来更改为 8）：

^demo

select * 
  from table1 
  where cli_id in (
    select cli_id from (
      select cli_id, col, val, count(1) cnt
        from table1 
        unpivot (val for col in 
          (COL1_A as 'a', COL1_B as 'b', COL1_C as 'c', COL1_D as 'd', COL1_E as 'e',
           COL2_A as 'a', COL2_B as 'b', COL2_C as 'c', COL2_D as 'd', COL2_E as 'e',
           COL3_A as 'a', COL3_B as 'b', COL3_C as 'c', COL3_D as 'd', COL3_E as 'e'))
        group by cli_id, col, val )
      group by cli_id
      having min(cnt) = 3)

编辑：

当至少有 2 组时，我会认为某一行有重复列数相同

所以我们必须加入未透视的数据：

^dbfiddle

with u as (
  select cli_id, country, name, ltr, num, val
        from table1 
        unpivot (val for (num, ltr) in (
          COL1_A as (1, 'a'), COL1_b as (1, 'b'), COL1_c as (1, 'c'), 
          COL1_d as (1, 'd'), COL1_e as (1, 'e'),
          COL2_A as (2, 'a'), COL2_b as (2, 'b'), COL2_c as (2, 'c'), 
          COL2_d as (2, 'd'), COL2_e as (2, 'e'),
          COL3_A as (3, 'a'), COL3_b as (3, 'b'), COL3_c as (3, 'c'), 
          COL3_d as (3, 'd'), COL3_e as (3, 'e')
          )))
select distinct cli_id, country, name
  from u a join u b using (cli_id, country, name, ltr, val) 
  where a.num < b.num
  group by cli_id, country, name, a.num, b.num
  having count(1) = 5

【讨论】：

感谢您的回复。但是不，它不符合我的要求。当至少有 2 组相同的列时，我会认为一行有重复