Question

我有以下结果集

+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+
| ID | row_start_date | row_end_date | terminate_date | termDateWasMember | dropped_reason | join_date  | rejoin_date | renew_date |
+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+
| 1  | 12-31-2016     | 02-03-2017   | NULL           | Member            | NULL           | 08-08-2016 | NULL        | 08-08-2016 |
+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+
| 1  | 2-4-2016       | 2-14-2017    | NULL           | Member            | NULL           | 08-08-2016 | NULL        | 08-08-2016 |
+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+
| 1  | 2-15-2017      | 2-21-2017    | 2-15-2016      | NonMember         | DuesPayment    | 08-08-2016 | NULL        | 08-08-2016 |
+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+
| 1  | 2-22-2017      | 12-31-9999   | NULL           | Member            | DuesPayment    | 08-08-2016 | 2-22-2016   | 08-08-2016 |
+----+----------------+--------------+----------------+-------------------+----------------+------------+-------------+------------+

我想要完成的是创建一个报告或某些内容，以便在一行中显示该人（在这种情况下，它是一个人，在DW中很多）他们是一个成员最多（终止日期）然后发布终止日期他们再次加入。因此，如果鲍勃是2016年8月8日的成员，并在02-15-2017终止会员资格，然后支付会费并重新加入02-22-2017我想表明鲍勃是一名成员从8-8-2016终止和重新加入。

我试图弄清楚是否有办法做到这一点。我想尝试在代码中执行此操作运载工具可以是PowerBI，Excel 2016，Tableau SSRS

以下是使用

的现有代码

        SELECT  m.id AS ID ,
        CONVERT(CHAR(10),m.row_start_date,110) AS row_start_date ,
        CONVERT(CHAR(10),m.row_end_date ,110) AS row_end_date,
        CONVERT(CHAR(10),m.terminate_date,110) AS terminate_date ,
        CASE WHEN m.terminate_date IS NULL THEN 'Member' ELSE 'Non-Member' END AS termDateWasMember,
        m.dropped_reason ,
        CONVERT(CHAR(10),join_date,110) AS join_date ,
        CONVERT(CHAR(10),rejoin_date,110) AS rejoin_date ,
        CONVERT(CHAR(10),renew_date,110) AS renew_date 
        FROM    dim_membership m
        INNER JOIN dim_individual i ON m.id = i.id
        WHERE   m.source_bk IN (
        SELECT  m.source_bk
        FROM    dim_membership m
                INNER JOIN 
         ( SELECT source_bk ,MIN(terminate_date) AS terminate_date
                             FROM   dim_membership
                             GROUP BY source_bk
                             HAVING MIN(terminate_date) >= '02/15/2017'
                           ) can ON m.source_bk = can.source_bk
        WHERE   row_is_current = 'Y')
        AND m.row_start_date > '2010-12-31'
        AND m.membership_level = 'National'
        AND i.row_is_current = 'Y' 
        AND id =  9182039 --7236901   --9182039--= 9129028   
        ORDER BY m.source_bk ,
         m.row_start_date,3 DESC

Answer 1

假设您可以查找非成员行，然后是成员行，我会添加ROW_NUMBER OVER(PARTITION BY ID ORDER BY row_start_date) AS rownum然后自我加入，所以请将您的查询放在顶部的CTE中：

WITH (your_query) AS model
SELECT m_term.terminate_date AS gap_start
       ,m_renew.rejoin_date AS gap_end
    ,m_term.*, m_renew.*
FROM model m_term
INNER JOIN model m_renew
    ON m_renew.ID = m_term.ID
    AND m_renew.rownum = m_term.rownum + 1
    AND m_term.termDateWasMember = 'NonMember'
    AND m_renew.termDateWasMember = 'Member';

所以它正在寻找从NonMember到Member的转换，然后使用相应行中的日期字段来确定差距。

数据仓库查询以查找日期中的差距

1 个答案: