Vertica的这些事(十三)——Vertica中实现Oracle中的ws_concat功能

vertica中没有类似Oracle中的ws_concat函数功能，需要开发UDF，自己对C++不熟悉，所有只有想其他方法解决了。
上代码：

SELECT node_state,
    MAX(DECODE(row_number, 1, a.node_name)) ||
    NVL(MAX(DECODE(row_number, 2, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 3, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 4, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 5, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 6, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 7, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 8, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 9, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 10, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 11, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 12, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 13, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 14, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 15, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 16, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 17, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 18, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 19, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 20, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 21, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 22, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 23, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 24, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 25, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 26, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 27, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 28, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 29, ',' || a.node_name)), '') ||
    NVL(MAX(DECODE(row_number, 30, ',' || a.node_name)), '') node_name
FROM (select node_state,node_name,row_number() over (partition by node_state order by node_name) row_number from nodes) a
group by node_state;```

运行结果：


  

 

  
 
 

  
1
  
2
  
3
  
4
  
5
  
6
  
7
  
8
  
9
  
10
  
11
  
12
  
13
  
14
  
15
  
16
  
17
  
18
  
19
  
20
  
21
  
22
  
23
  
24
  
25
  
26
  
27
  
28
  
29
  
30
  
31
  
32
  
33
  
34
  
35
  
36
 

可以看到是我想要的结果。
有人会问了，我不知道自己要ws_concat几个类别怎么办，先运行一下

select node_state,node_name,row_number() over (partition by node_state order by node_name) row_number from nodes

  

 

  
1
 

看下row_number不就行了，或者把decode中的数字多写几个。当然最好的方法还是开发自定义程序包，引入到vertica中，直接使用函数实现，C++好的童鞋可以试一下。

文章来源: dataclub.blog.csdn.net，作者：数据社，版权归原作者所有，如需转载，请联系作者。

原文链接：dataclub.blog.csdn.net/article/details/52937198