Question

我有一个如下表格，我希望复制记录，而最小日期小于或等于最大日期

686151209   E13232677   1333439 2017-10-23
686151209   E13232677   1333439 2017-10-26

我希望得到如下结果集

686151209   E13232677   1333439 2017-10-23
686151209   E13232677   1333439 2017-10-24
686151209   E13232677   1333439 2017-10-25
86151209    E13232677   1333439 2017-10-26

Answer 1

您并使用spt_values获取连续数字：

   ;WITH testdata(col1,col2,col3,col4)AS(
      SELECT '686151209','E13232677','1333439','2017-10-23' UNION all
      SELECT '686151209','E13232677','1333439','2017-10-26'
   )
   SELECT col1,col2,col3,DATEADD(d,sv.number-1,a.mindate) AS col4,sv.number FROM (
      SELECT col1,col2,col3,CONVERT(DATE,MIN(col4)) AS mindate,CONVERT(DATE,MAX(col4)) AS maxdate 
      FROM testdata AS t 
     group  by col1,col2,col3
   ) AS a 
   INNER JOIN master.dbo.spt_values AS sv ON sv.type='P' AND sv.number BETWEEN 1 AND DATEDIFF(d,mindate,maxdate)+1

+-----------+-----------+---------+------------+--------+
| col1      | col2      | col3    | col4       | number |
+-----------+-----------+---------+------------+--------+
| 686151209 | E13232677 | 1333439 | 2017-10-23 | 1      |
| 686151209 | E13232677 | 1333439 | 2017-10-24 | 2      |
| 686151209 | E13232677 | 1333439 | 2017-10-25 | 3      |
| 686151209 | E13232677 | 1333439 | 2017-10-26 | 4      |
+-----------+-----------+---------+------------+--------+

Answer 2

一种方法是数字表。如果你没有太多行，我也喜欢递归CTE：

with cte as (
      select col1, col2, col3, mind, maxd
      from (select col1, col2, min(dte) as mind, max(dte) as maxd
            from t
            group by col1, col2, col3
           ) t
      union all
      select col1, col2, col3, dateadd(day, 1, mind), maxd
      from cte
      where  dateadd(day, 1, mind) < maxd
     )
select col1, col2, col3, mind
from cte;

除非您设置了最大递归选项，否则每个col1 / col2组合限制为100行。

Answer 3

或者像这样：

CREATE TABLE temp
(
    ID BIGINT,
    CODE VARCHAR(50),
    ID2 BIGINT,
    DATE DATE
);

INSERT INTO temp VALUES (686151209, 'E13232677', 1333439, '2017-10-23'),
                          (686151209, 'E13232677', 1333439, '2017-10-26');

SELECT generate_series(T.D1::timestamp, T.D2::timestamp, interval '1 day')::date
FROM
(
SELECT A.id, A.code, A.id2,  A.dates AS D1, B.dates  AS D2
FROM temp A
LEFT JOIN temp b ON (A.id = B.id AND
                      A.code=B.code AND
                      A.id2 = B.id2 AND
                      B.dates > A.dates)
WHERE B.id IS NOT NULL
) T;

如何复制结果集，如下

3 个答案: