在Oracle中連接來自多行的多個列

Question

我在stackoverflow中搜索到了並找到了Concatenate multiple columns和Concatenate multiple rows。但我需要的是將兩者結合起來。
我有表名爲komponen：

id urut tipe c_string c_number op 
--------------------------------------------- 
A1 1  S  GP   NULL  * 
A1 2  N  NULL  5   /
A1 3  N  NULL  100   + //Ignore the last op for each groups. 
A2 1  S  GP   NULL  - 
A2 2  N  NULL  1000  ///Ignore the last op for each groups. 

期望的結果：

id  concat_result 
------------------------ 
A1  GP * 5/100 
A2  GP - 1000 

這可能是使用LISTAGG和GROUP BY方法。但我沒有任何線索如何做到這一點，並達到預期的效果。請幫忙。

Answer 1

這與here的一些修改。新增||（串連）在LISTAGG

SELECT ID, 
    LISTAGG (
     CASE 
     WHEN TIPE = 'v' THEN 
     C_STRING 
     ELSE 
     TO_CHAR (C_NUMBER) 
     END || OP, 
     ' ' 
    ) WITHIN GROUP (ORDER BY URUT) AS CONCAT_RESULT 
FROM KOMPONEN 
GROUP BY ID; 

Answer 2

你的預感是正確的。在這個解決方案中，我假設c_string和c_number是相互排斥的，即一個將不爲空。

with t as (
    select id, listagg(nvl(c_string, c_number)||' '||op, ' ') within group (order by urut) result 
    from komponen 
    group by id 
) 
select id, substr(result, 1, length(result)-2) 
from t; 

的with和substr()的結合是除去最後的算。

Answer 3

您可以使用：

SqlFiddleDemo

WITH cte AS 
(
    SELECT 
     id, 
     LISTAGG(string , '') WITHIN GROUP (ORDER BY urut) AS concat_result 
    FROM(SELECT id, urut, 
     NVL(c_string, '') || ' ' || NVL(c_number, '') || ' ' || NVL(op, '') AS string 
     FROM komponen) 
    GROUP BY id 
) 
SELECT 
    id, 
    SUBSTR(concat_result,1,LENGTH(concat_result)-1) AS concat_result 
FROM cte 

工作原理：

1. 連接具有排NVL(c_string, '') || ' ' || NVL(c_number, '') || ' ' || NVL(op, '')
2. 連接具有多行LISTAGG
3. 刪除最後一個字符SUBSTR(concat_result,1,LENGTH(concat_result)-1)

Answer 4

一種方式是通過使用row_number分析功能和listaggr

SQL> WITH table_("ID", urut, tipe, c_string, c_number, op) AS 
    2 (SELECT 'A1', 1, 'S', 'GP', NULL, '*' from dual union all 
    3 SELECT 'A1', 2, 'N', NULL, '5', '/' from dual union all 
    4 SELECT 'A1', 3, 'N', NULL, '100', '+' from dual union all 
    5 SELECT 'A2', 1, 'S', 'GP', NULL, '-' from dual union all 
    6 SELECT 'A2', 2, 'N', NULL, '1000', '/' from dual), 
    7 ------- 
    8 -- End if Data preparation 
    9 ------- 
10 table2_ AS (SELECT t.*, row_number() OVER (partition BY "ID" ORDER BY URUT DESC) AS rn 
11 FROM table_ t) 
12 SELECT "ID", 
13   listagg(coalesce(c_string, c_number) || ' ' || CASE 
14      WHEN rn = 1 THEN 
15      NULL 
16      ELSE 
17      op 
18     END, 
19     ' ') within GROUP(ORDER BY urut) AS EXPR 
20 FROM table2_ 
21 GROUP BY "ID" 
22/

輸出

ID EXPR 
-- -------------------------------------------------------------------------------- 
A1 GP * 5/100 
A2 GP - 1000