teradata 中数据值为 0 时缺少行

Question

我试图在查询中进行分组，但每次数据有 0 时，分组不会显示整行。我该如何解决这个问题？

这是常规查询

select 
    COUNT(DISTINCT sr.sr_number) AS NEW_CASES
FROM table sr
where 
    sr.business_unit in('BU1')
    and OPENED_DATE < trunc(current_Date)
    and OPENED_DATE > trunc(current_date -2)
    and sr.status = 'Open'

输出是

NEW_CASES
0

但是当我做一个分组时。整行都不见了。

select 
    COUNT(DISTINCT sr.sr_number) AS NEW_CASES,
    sr.business_unit
FROM table sr
where 
    sr.business_unit in('BU1','BU2','BU3' )
    and OPENED_DATE < trunc(current_Date)
    and OPENED_DATE > trunc(current_date -2)
    and sr.status = 'Open'
group by sr.business_unit

按输出分组是

New_CASES    BUSINESS_UNIT
      200    BU2
      300    BU3

期望的输出：

New_CASES    BUSINESS_UNIT
        0    BU1
      200    BU2
      300    BU3

Answer 1

一种选择是从一个固定的值列表开始，然后将表带上left join，如下所示：

select b.business_unit, count(distinct t.sr_number) as new_cases
from (
    select 'bu1' business_unit from dual
    union all select 'bu2' from dual
    union all select 'bu3' from dual
) b
left join mytable t
    on  t.business_unit = b.business_unit
    and t.opened_date > trunc(current_date -2)
    and t.opened_date < trunc(current_date)
    and t.status = 'open'
group by b.business_unit

在 Teradata 中，语法有点麻烦：

select b.business_unit, count(distinct t.sr_number) as new_cases
from (
    select * from (select 'bu1' as business_unit) x
    union all select * from (select 'bu2' as business_unit ) x
    union all select * from (select 'bu3' as business_unit ) x
) b
left join mytable t
    on  t.business_unit = b.business_unit
    and t.opened_date > trunc(current_date -2)
    and t.opened_date < trunc(current_date)
    and t.status = 'open'
group by b.business_unit

Answer 2

使用left join。在 Oracle 中，这看起来像

select b.business_unit, COUNT(DISTINCT sr.sr_number) AS NEW_CASES
from (select 'BU1' as business_unit from dual union all
      select 'BU2' as business_unit from dual union all
      select 'BU3' as business_unit from dual 
     ) b left join
     sr
     on sr.business_unit = b.business_unit and
        sr.OPENED_DATE < trunc(current_Date) and
        sr.OPENED_DATE > trunc(current_date -2) and
        sr.status = 'Open'
group by b.business_unit

编辑：

Teradata 没有一种非常方便的方法来创建常量值的派生表，但您可以做到：

select b.business_unit, COUNT(DISTINCT sr.sr_number) AS NEW_CASES
from (select 'BU1' as business_unit from (select 1 as dummy) t union all
      select 'BU2' as business_unit from (select 1 as dummy) t  union all
      select 'BU3' as business_unit from (select 1 as dummy) t  
     ) b left join
     sr
     on sr.business_unit = b.business_unit and
        sr.OPENED_DATE < trunc(current_Date) and
        sr.OPENED_DATE > trunc(current_date -2) and
        sr.status = 'Open'
group by b.business_unit