SQL标记错误并获取最新的有效状态

时间:2015-01-29 22:28:26

标签: sql db2 zos analytical

请帮助解决以下问题,我有一组记录,我需要为给定的ID和Case_Type循环,并将记录标记为Valid或ERROR。

由于它是一个审核表,因此事件必须按顺序排列,例如打开/挂起/完成打开/完成

因此Rec_no 1和4是以下样本数据中的ERROR。 Rec_no 5(对于ID 9)带有最新有效状态(已完成')。

以下是样本记录 :



Rec_num  ID  Case_Type  Status
1	     9	     1	     Pend	    -- ERROR (Out of Order)
2        9       1       open
3        8       1       open
4        9       1       open       -- ERROR (Duplicate)
5        9       1       done       
6        9       1       open
7        8       1       done




我需要以下输出:



Lag and Lead functions are not available in DB2 z/OS.

This is what i have till now. I guess I need something like 

max(case when previous_stat = 'open' and (Status = 'Pend' or Status = 'done' then 1 else 0) as Flag over(partition by ID,Case_Type order by Rec_num).


WITH     TST AS                                                                                                                                         
(SELECT 1 AS Rec_Num, 9 AS ID, 1 AS Case_Type, 'Pend'  AS Status FROM               
SYSIBM.SYSDUMMY1 UNION ALL 
SELECT 2 AS Rec_Num, 9 AS ID, 1 AS Case_Type, 'open'  AS Status FROM               
SYSIBM.SYSDUMMY1 UNION ALL                                                     
SELECT 3 AS Rec_Num , 8 AS ID,2  AS Case_Type, 'open'  AS Status FROM                
SYSIBM.SYSDUMMY1  UNION ALL                                                    
SELECT 4 AS Rec_Num, 9 AS ID, 1 AS Case_Type, 'open' AS Status FROM SYSIBM.SYSDUMMY1
UNION ALL                                                                      
SELECT 5 AS Rec_Num, 9 AS ID, 1 AS Case_Type, 'done' AS Status FROM SYSIBM.SYSDUMMY1
UNION ALL                                                                      
SELECT 6 AS Rec_Num, 9 AS ID, 1 AS Case_Type, 'open'  AS Status FROM                 
SYSIBM.SYSDUMMY1
union all
SELECT 7 AS Rec_Num, 8 AS ID, 1 AS Case_Type,  'done' AS Status FROM                 
SYSIBM.SYSDUMMY1)                                                              
                                                                               
SELECT ID,Case_Type, Status, MAX(Status) OVER(PARTITION BY ID,Case_Type ORDER BY    
Rec_Num ROWS BETWEEN 1 PRECEDING AND 1 PRECEDING) AS PREVIOUS_STAT      from TST

Rec_num  ID  Case_Type  Status    Flag
1	     9	     1	     Pend      ERROR
2        9       1       open      Valid
3        8       1       open      Valid
4        9       1       open      ERROR
5        9       1       done      Valid 
6        9       1       open      ERROR
7        8       1       done      Valid




0 个答案:

没有答案