SQL在Netezza中查找日期范围的差距

时间:2017-03-08 05:39:32

标签: sql netezza gaps-and-islands

我想获得日期范围之间的差距,并且需要用MM ID来区分差距。

你可以请别人帮我解决下面的Netezza

INPUT:

ID | MM Number | From | To 

12345|4444444 |2015-01-03 |2015-02-02 

12346|4444444 |2015-02-03 |2015-03-02 

<<缺少一条记录(相同的MM号码从2015-03-03到2015-04-02期间没有记录

12347|4444444 |2015-04-03 |2015-05-01 

12348|4444444 |2015-05-02 |2015-06-01 

12349|4444444 |2015-06-02 |2015-07-01 

12310|4444444 |2015-07-02 |2015-07-31 

<<缺少一条记录(例如:相同的MM号码从2015-08-01到2015-08-31期间没有记录

12310|4444444|2015-09-01|2015-09-30 

预期结果:

MM No | Missing Start Date | Missing To Date 
4444444 |2015-03-03 |2015-04-02 
4444444 |2015-08-01 |2015-08-31 

我有一张包含以下数据的表格

create table icr_tmp 
( 
ID VARCHAR2(15), 
NMI VARCHAR2(50), 
INVOICE_START_DATE DATE, 
INVOICE_END_DATE DATE); 

insert into icr_tmp values('12345','4444444','03/01/2015','02/02/2015'); 

insert into icr_tmp values('12346','4444444','03/02/2015','02/03/2015'); 

insert into icr_tmp values('12347','4444444','03/04/2015','01/05/2015'); 

insert into icr_tmp values('12348','4444444','02/05/2015','01/06/2015'); 

insert into icr_tmp values('12349','4444444','02/06/2015','01/07/2015'); 

insert into icr_tmp values('12310','4444444','02/07/2015','31/07/2015'); 

insert into icr_tmp values('12310','4444444','01/09/2015','30/09/2015'); 

下面的SQL完全在oracle中工作,但在Netezza中没有。

select * 
from (
  select nmi,max(invoice_end_date) over(partition by nmi order by invoice_start_date) + 1 start_gap,
    lead(invoice_start_date) over(partition by nmi order by invoice_start_date) - 1 end_gap
  from icr_tmp3
)
where start_gap <= end_gap;

NMI      START_GAP   END_GAP

4444444  2015-03-03  2015-04-02

4444444  2015-08-01  2015-08-31

如何更改上述SQL以在Netezza中运行?

1 个答案:

答案 0 :(得分:0)

你的查询应该在netezza上运行得很好,只有我看到的东西是子查询的别名,因为如果子查询没有别名,netezza会抱怨。

select * 
from (
  select nmi
  ,max(invoice_end_date) over(partition by nmi order by invoice_start_date) + 1 start_gap,
  lead(invoice_start_date) over(partition by nmi order by invoice_start_date) - 1 end_gap  
  from icr_tmp
) foo
where start_gap <= end_gap;