一个查询中两个表的Teradata减列

Question

我是Teradata的新手，我有两个表，假设tbl_A和tbl_B具有相同的架构。 我的目标是：

1. 当我将列1,2,3上的数据分组时，获得两个表的count（*）的差

架构为

   Library_Card_Type Student_Class Book_Type
c1                  A           NOVEL
c2                  B           HISTORY
c3                  C           MOVIE
c4                  D           GK
c1                  E           POETRY
c1                  A           NOVEL
c2                  B           HEALTH
c3                  C           POLITICS
c4                  D           SQL
c1                  E           JAVA

当我按Library_Card_Type，Student_Class和Book_Type对数据分组时，我想得到tbl_A和tbl_b的count（*）的差，我尝试过的查询是：

sel count(*) AS Count_of_tbl_A,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_A ;
Union all
sel count(*) AS Count_of_tbl_B,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_B;
sel (Count_of_tbl_A - Count_of_tbl_B) As Overall_Difference;

其他方法是

 Sel Count_of_tbl_B from (sel count(*) AS Count_of_tbl_B,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_B) A,
    Sel Count_of_tbl_A from (sel count(*) AS Count_of_tbl_A,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_A) B,
    (Count_of_tbl_A - Count_of_tbl_B) as minus_result sample 10;
and the inner join also
select
 (sel count(*) AS Count_of_tbl_A,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_A)  AS Count_of_tbl_A
, (sel count(*) AS Count_of_tbl_A,Library_Card_Type AS Library_Card_Type ,Book_Type AS Book_Type,Student_Class AS Student_Class group by 2,3,4 order by 2,3,4 from tbl_B)  AS Count_of_tbl_B
,A.Library_Card_Type AS Library_Card_Type1
,B.Library_Card_Type AS Library_Card_Type
,A.Book_Type AS Book_Type1
,B.Book_Type AS Book_Type
,A.Student_Class AS Student_Class1
,B.Student_Class AS Student_Class
,(Count_of_tbl_A - Count_of_tbl_B ) AS Difference_of_Count 
FROM  tbl_B B 
INNER JOIN 
tbl_A A   
ON  A.Library_Card_Type=B.Library_Card_Type AND
    A.Book_Type=B.Book_Type  AND
    A.Student_Class=B.Student_Class;

我不知道该怎么做，请指导我。

Answer 1

您需要在两个派生表中分别进行计数，然后将它们合并。 但是由于可能只有一个表存在组合，因此您必须切换到完全外部联接加COALESCE：

select
   coalesce(A.Library_Card_Type, B.Library_Card_Type) as Library_Card_Type
  ,coalesce(A.Book_Type, B.Book_Type) as Book_Type
  ,coalesce(A.Student_Class, B.Student_Class) as Student_Class
  ,coalesce(A.cnt,0) -  coalesce(B.cnt, 0) as difference
  ,coalesce(A.cnt,0) as tblA_cnt
  ,coalesce(B.cnt,0) as tblB_cnt
from
 (
   select
      Library_Card_Type
     ,Book_Type
     ,Student_Class
     ,count(*) AS cnt
   from tbl_A
   group by 1,2,3
 ) as A
full join
 (
   select
      Library_Card_Type
     ,Book_Type
     ,Student_Class
     ,-count(*) AS cnt
   from tbl_B
   group by 1,2,3
 ) as B
  ON A.Library_Card_Type=B.Library_Card_Type
 AND A.Book_Type=B.Book_Type  
 AND A.Student_Class=B.Student_Class

或者，您切换到另一种方法，UNION既可以得到结果，也可以进行汇总：

select
   Library_Card_Type
  ,Book_Type
  ,Student_Class
  ,sum(cnt) as difference
  ,max(cnt) as tblA_count
  ,abs(min(cnt)) as tblB_count
from
 (
   select
      Library_Card_Type
     ,Book_Type
     ,Student_Class
     ,count(*) AS cnt -- positive numbers
   from tbl_A
   group by 1,2,3

   union ALL -- more efficient than UNION

   select
      Library_Card_Type
     ,Book_Type
     ,Student_Class
     ,- count(*) AS cnt -- negative numbers
   from tbl_B
   group by 1,2,3
 ) as dt
group by 1,2,3

在这两种情况下，您都可以添加过滤器以仅显示差异，

where difference <> 0  -- join query 

要么

having difference <> 0 -- union query