我有几种不同类型的数据涉及日期范围,我想合并在一起,但同时按天分解.因此,3天的数据将导致三行：

start    primary_key
start+1  primary_key
start+2  primary_key

我一直在玩10g中使用select语句的model子句,并且正在寻找实现这一目标的最佳方法.目前我正在加入一系列日期,涵盖所有可能的日期(选择最小(开始日期),最大(结束日期)).我更愿意选择数据并添加更多行以将其转换为每日数据集.

编辑：

我设法提出(现在包括样本数据)：

    SELECT * FROM (
        SELECT 123 req_code, 
               345 req_par_code, 
               TO_DATE('01-03-2010', 'dd-mm-yyyy') req_start_date, 
               TO_DATE('05-03-2010', 'dd-mm-yyyy') req_end_date
        FROM dual
)

    MODEL 
        PARTITION BY (req_code)
        DIMENSION BY (0 d)
        MEASURES     (SYSDATE dt, req_par_code, req_start_date, req_end_date)
        RULES ITERATE(365) UNTIL (dt[iteration_number] >= TRUNC(req_end_date[0])) (
            dt[iteration_number] = NVL(dt[iteration_number-1] + 1, TRUNC(req_start_date[0])),

            --Copy data across
            req_par_code[ iteration_number ] = req_par_code[0],
            req_start_date[ iteration_number ] = req_start_date[0],
            req_end_date[ iteration_number ] = req_end_date[0]
        )
    ORDER BY dt, req_code;

最佳答案 您可以使用MODEL子句生成行,这是一个小例子：

SQL> SELECT * FROM t_data;

        PK START_DATE  END_DATE
---------- ----------- -----------
         1 20/01/2010  20/01/2010
         2 21/01/2010  23/01/2010
         3 24/01/2010  27/01/2010

SQL> SELECT pk, start_date, end_date FROM t_data
  2  MODEL
  3     PARTITION BY (pk)
  4     DIMENSION BY (0 AS i)
  5     MEASURES(start_date, end_date)
  6     RULES
  7     (  start_date[FOR i
  8                   FROM 1 TO end_date[0]-start_date[0]
  9                   INCREMENT 1] = start_date[0] + cv(i),
 10        end_date[ANY] = start_date[CV()] + 1
 11     )
 12  ORDER BY 1,2;

        PK START_DATE  END_DATE
---------- ----------- -----------
         1 20/01/2010  21/01/2010
         2 21/01/2010  22/01/2010
         2 22/01/2010  23/01/2010
         2 23/01/2010  24/01/2010
         3 24/01/2010  25/01/2010
         3 25/01/2010  26/01/2010
         3 26/01/2010  27/01/2010
         3 27/01/2010  28/01/2010