我要求我有两个表tbl_bb
和tbl_rr
,其中的数据如下:
tbl_bb :
isin agency rating date
----------------------------------
142356 MDA A 2010-01-20
142356 MDA AA 2012-05-14
142356 MDA BB 2013-07-04
tbl_rr :
isin agency rr_rating rr_date
------------------------------------
142356 MDA A 2010-01-15
142356 MDA AA+ 2009-05-30
142356 MDA BBB 2013-07-04
现在,要求我必须从tbl_bb中选择每一条记录,并根据isin和agency的加入循环tbl_rr中的数据,并且循环的条件是从tbl_bb考虑日期并检查每个日期从tbl_rr
选择最接近tbl_bb
日期的日期,其他条件为rr_date(tbl_rr) <= date (tbl_bb)
输出:
isin agency rating date rr_rating rr_date
-------------------------------------------------------
142356 MDA A 2010-01-20 A 2010-01-15
142356 MDA AA 2012-05-14 AA+ 2010-01-15
142356 MDA BB 2013-07-04 BBB 2013-07-04
谢谢!
答案 0 :(得分:1)
我会使用APPLY
执行此操作,以便从tbl_rr
获取最接近的1条记录:
SELECT bb.isin,
bb.agency,
bb.rating,
bb.date,
rr.rr_rating,
rr.rr_date
FROM tbl_bb AS bb
OUTER APPLY
( SELECT TOP 1 rr.rr_rating, rr.rr_date
FROM tbl_rr AS rr
WHERE rr.isin = bb.isin
AND rr.agency = bb.agency
AND rr.rr_date <= bb.date
ORDER BY rr.rr_date DESC
) AS rr;
完整的工作示例
-- SAMPLE DATA
WITH tbl_bb AS
( SELECT isin, agency, rating, date
FROM (VALUES
(142356,'MDA','A','2010-01-20'),
(142356,'MDA','AA','2012-05-14'),
(142356,'MDA','BB','2013-07-04')
) AS bb(isin, agency, rating, date)
), tbl_rr AS
( SELECT isin, agency, rr_rating, rr_date
FROM (VALUES
(142356,'MDA','A','2010-01-15'),
(142356,'MDA','AA+','2012-05-30'),
(142356,'MDA','BBB','2013-07-04')
) AS rr (isin, agency, rr_rating, rr_date)
)
-- SAMPLE DATA END
SELECT bb.isin,
bb.agency,
bb.rating,
bb.date,
rr.rr_rating,
rr.rr_date
FROM tbl_bb AS bb
OUTER APPLY
( SELECT TOP 1 rr.rr_rating, rr.rr_date
FROM tbl_rr AS rr
WHERE rr.isin = bb.isin
AND rr.agency = bb.agency
AND rr.rr_date <= bb.date
ORDER BY rr.rr_date DESC
) AS rr;
答案 1 :(得分:0)
我要做的是我会按照说明中的说明加入这些表格,在两个日期之间找到DATEDIFF
,然后找到每个DATEDIFF
和ISIN
的最低agency
值}
答案 2 :(得分:0)
根据我的假设,如果你想过滤我们可以实现的数据,我们可以这样继续进行,如
C.RN - 1 = CC.RR or C.RN - 1 > CC.RR OR C.RN - 1 < CC.RR
代码:
declare @t table (ISIN INT,Agency varchar(10),Rating varchar(5),dated date)
insert into @t (ISIN,Agency,Rating,dated)values (142356,'MDA','A','2010-01-20'),
(142356,'MDA','AA','2012-05-14'),
(142356,'MDA','BB','2013-07-04')
declare @tt table (ISIN INT,Agency varchar(10),Rating varchar(5),dated date)
insert into @tt (ISIN,Agency,Rating,dated)values (142356,'MDA','A','2010-01-15'),
(142356,'MDA','AA+','2012-05-30'),
(142356,'MDA','BBB','2013-07-04')
;with cte as
(select ISIN,
Agency,
Rating,
dated,
ROW_NUMBER()OVER(ORDER BY ISIN)RN
from @t)
, cte1 as
(select
ISIN,
Agency,
Rating,
dated,
ROW_NUMBER()OVER(ORDER BY ISIN)RR
from @tt)
select C.ISIN,C.Agency,
C.Rating,C.dated,
CC.Agency,CC.Rating,
CC.dated from cte c
INNER JOIN cte1 CC
ON C.RN = CC.RR