在分區內上移非空值

Question

不確定標題是否有意義，但這是我的問題以及如何到達當前位置。

背景：我在一對多設置中有2個表。 many表可以具有與該表關聯的任何數量的記錄，但是具有一個附加列，可以將該多行記錄標識為介於1到7之間的任何值。 我用7個CTE來“收集”每組1-7的所有行，然后將它們連接到一個表上，使我到達示例數據的地步。

如果單個列中的所有行都為空，那么我想顯示一個*，否則按GROUPID / GROUPNAME在每個列中顯示所有不同的值。 我有些駭人聽聞，並以一種告誡方式大幅度削減了獲取自己想要的東西的方式。 如果可能的話，我想在GROUPID / GROUPNAME中分別對每一列進行排序，以使NULL不在數據中間。

這是使我陷入困境的代碼：

WITH SAMPLEDATA (CYCLEID,GROUPID,GROUPNAME,COL1,COL2,COL3,COL4,COL5,COL6,COL7)
AS (
SELECT 1,7669,'000000261','GAS',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 2,7669,'000000261','GAS',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 3,7669,'000000261','GAS',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 4,7669,'000000261','GAS',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 5,7669,'000000261','GFG',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 6,7669,'000000261','GFG',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 7,7669,'000000261','GFG',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 8,7669,'000000261','GFG',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 9,7669,'000000261','GKE',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 10,7669,'000000261','GKE',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 11,7669,'000000261','GKE',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 12,7669,'000000261','GKE',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
 ) 
 , GROUPSPROCESSED AS (
SELECT
  GROUPID,
  GROUPNAME,
  CASE
    WHEN cycleid = 1
    THEN NVL(COL1,'*')
    WHEN COL1 = FIRST_VALUE(COL1) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL1)
    THEN NULL
    WHEN COL1 = LAG(COL1) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL1)
    THEN NULL
    ELSE COL1
  END AS "COL1",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL2,'*')
    WHEN COL2 = FIRST_VALUE(COL2) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL2)
    THEN NULL
    WHEN COL2 = LAG(COL2) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL2)
    THEN NULL
    ELSE COL2
  END AS "COL2",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL3,'*')
    WHEN COL3 = FIRST_VALUE(COL3) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL3)
    THEN NULL
    WHEN COL3 = LAG(COL3) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL3)
    THEN NULL
    ELSE COL3
  END AS "COL3",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL4,'*')
    WHEN COL4 = FIRST_VALUE(COL4) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL4)
    THEN NULL
    WHEN COL4 = LAG(COL4) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL4)
    THEN NULL
    ELSE COL4
  END AS "COL4",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL5,'*')
    WHEN COL5 = FIRST_VALUE(COL5) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL5)
    THEN NULL
    WHEN COL5 = LAG(COL5) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL5)
    THEN NULL
    ELSE COL5
  END AS "COL5",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL6,'*')
    WHEN COL6 = FIRST_VALUE(COL6) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL6)
    THEN NULL
    WHEN COL6 = LAG(COL6) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL6)
    THEN NULL
    ELSE COL6
  END AS "COL6",
  CASE
    WHEN cycleid = 1
    THEN NVL(COL7,'*')
    WHEN COL7 = FIRST_VALUE(COL7) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL7)
    THEN NULL
    WHEN COL7 = LAG(COL7) OVER (PARTITION BY GROUPNAME ORDER BY GROUPNAME, COL7)
    THEN NULL
    ELSE COL7
  END AS "COL7"
FROM SAMPLEDATA)
SELECT * FROM GROUPSPROCESSED
WHERE COALESCE (COL1,COL2,COL3,COL4,COL5,COL6,COL7) IS NOT NULL
ORDER BY 1,2,3,4,5,6,7,8,9

電流輸出：

所需的輸出：

編輯這是在執行任何操作之前從我的示例中導入確切原始數據的代碼。 從技術上講，我需要將3個表捆綁在一起以破壞上面的輸出。

GROUPPROFILE在上面的輸出中保存了字段1-2中的ID和NAME。

COLVALUES根據COLLEVELID為1-7保留輸出中其余字段的數據（NAME）

GROUPCOLMM表僅用於連接GROUPPROFILE.GROUPPROFILEID> COLVALUES.COLVALUEID

CREATE TABLE GROUPPROFILE (GROUPPROFILEID NUMBER(12), SHORTNM NVARCHAR2(30))
REM INSERTING into GROUPPROFILE
Insert into GROUPPROFILE (GROUPPROFILEID,SHORTNM) values (7669,'000000261');

CREATE TABLE COLVALUES (COLVALUEID NUMBER(12), NAME NVARCHAR2(50), COLLEVELID NUMBER(12))
REM INSERTING into COLVALUES
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (18901,'GFG',1);
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (106,'GKE',1);
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (1351,'GAS',1);
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (423,'0',4);
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (1,'1',5);
Insert into COLVALUES (COLVALUEID,NAME,COLLEVELID) values (18514,'Chester',5);

CREATE TABLE GROUPCOLMM (GROUPPROFILEID NUMBER(12), COLVALUEID NUMBER(12))
REM INSERTING into GROUPCOLMM
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,1);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,3);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,106);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,1351);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,16503);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,18514);
Insert into GROUPCOLMM (GROUPPROFILEID,COLVALUEID) values (7669,18901);

Answer 1

您無法從發布的“原始”數據中獲取樣本數據。 我從子查詢ORIGINALDATA中的樣本數據中反向工程了原始數據，然后提出了以下解決方案：

WITH SAMPLEDATA (CYCLEID,GROUPID,GROUPNAME,COL1,COL2,COL3,COL4,COL5,COL6,COL7)
AS (
SELECT 1,7669,'000000261','GAS',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 2,7669,'000000261','GAS',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 3,7669,'000000261','GAS',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 4,7669,'000000261','GAS',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 5,7669,'000000261','GFG',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 6,7669,'000000261','GFG',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 7,7669,'000000261','GFG',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 8,7669,'000000261','GFG',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 9,7669,'000000261','GKE',NULL,NULL,NULL,'1',NULL,'00' FROM DUAL
UNION ALL
SELECT 10,7669,'000000261','GKE',NULL,NULL,NULL,'1',NULL,'000000261' FROM DUAL
UNION ALL
SELECT 11,7669,'000000261','GKE',NULL,NULL,NULL,'Chester',NULL,'00' FROM DUAL
UNION ALL
SELECT 12,7669,'000000261','GKE',NULL,NULL,NULL,'Chester',NULL,'000000261' FROM DUAL
 ) 
, originaldata as (
  select distinct groupid, groupname, col, val from sampledata
  unpivot (val for col in (COL1 as 1,COL2 as 2,COL3 as 3,COL4 as 4,COL5 as 5,COL6 as 6,COL7 as 7))
)
select GROUPID, GROUPNAME, 
  case when rn = 1 and col1 is null then '*' else col1 end col1,
  case when rn = 1 and col2 is null then '*' else col2 end col2,
  case when rn = 1 and col3 is null then '*' else col3 end col3,
  case when rn = 1 and col4 is null then '*' else col4 end col4,
  case when rn = 1 and col5 is null then '*' else col5 end col5,
  case when rn = 1 and col6 is null then '*' else col6 end col6,
  case when rn = 1 and col7 is null then '*' else col7 end col7
from (
  select o.*,
  row_number() over(partition by groupid, groupname, col order by val) rn
  from originaldata o
)
pivot(
  max(val) 
  for col in (1 as COL1,2 as COL2,3 as COL3,4 as COL4,5 as COL5,6 as COL6,7 as COL7)
)
order by groupid, groupname, rn;

   GROUPID GROUPNAME COL1      COL2      COL3      COL4      COL5      COL6      COL7     
---------- --------- --------- --------- --------- --------- --------- --------- ---------
      7669 000000261 GAS       *         *         *         1         *         00       
      7669 000000261 GFG                                     Chester             000000261
      7669 000000261 GKE                                   

最好的問候，燉阿什頓