Teradata

Question

我试图在汇总数据时消除重复项。

CREATE  TABLE abc
( dt DATE FORMAT 'YYYY-MM-DD'
,acct NUMBER
,chk_lmt VARCHAR(10)
,baln_lmt VARCHAR(10)
,wk DATE FORMAT 'YYYY-MM-DD'
) ;
    INSERT  INTO abc
('2018-03-19' , 1001, 'under 1000' , 'under 500', '2018-03-23');

INSERT  INTO abc
('2018-03-20' , 1001, 'under 1000' , 'under 500', '2018-03-23');

INSERT  INTO abc
('2018-03-20' , 1002, 'under 1000' , 'under 500', '2018-03-23');

SELECT  dt,wk,'chk_lmt_chk' AS TYPEs, 
SUM ( CASE  WHEN chk_lmt = 'under 1000' THEN 1 ELSE 0 END   ) AS counts
  FROM abc
  GROUP BY 1,2,3;

    dt  wk  TYPEs   counts
19/03/2018  23/03/2018  chk_lmt_chk 1
19/03/2018  23/03/2018  chk_lmt_chk 0
20/03/2018  23/03/2018  chk_lmt_chk 2

对于 dt 20/03/2018，计数应为 1，因为帐户已在 19/03/2018 中捕获。

有什么想法吗？

谢谢

Answer 1

这回答了原始问题。

如果你想让账户只统计一次，第一次，使用row_number():

SELECT dt, wk, 'chk_lmt_chk' AS TYPEs, 
       SUM(CASE WHEN chk_lmt = 'under 1000' THEN 1 ELSE 0 END) AS counts
FROM (SELECT abc.*,
             ROW_NUMBER() OVER (PARTITION BY acct ORDER BY dt) as seqnum
      FROM ptemp.abc
     ) abc
WHERE seqnum = 1
GROUP BY 1, 2, 3;

Teradata - 聚合时消除重复条目

Teradata -Duplicate entries elimination while aggregating

sql