我必须在SQL中进行一些操作。
我需要在ColC中合并ColA和ColB - 由>分隔的替代文字 - 如下所示：a1-b1> a2-b2>a3-b3;。
实施例
ColA | ColB | ColC
----------------+------------------+-----------------------------
abcd > e > efg | ppppp > ppt > pp | abcd-ppppp > e-ppt > efg-pp
hij > kl > iiii | aaa > bbb > hhh | hij-aaa > kl-bbb > iiii-hhh
aa | fff | aa-fff
a > bbb | pp > a | a-pp > bbb-a
然后将ColD添加到ColC＆gt;的每个字符串路径的开头。最终结果ColE
dc1 > dc2 > dc3;...
ColC | ColD | ColE
----------------------------+------+----------------------------------------
abcd-ppppp > e-ppt > efg-pp | 2016 | 2016abcd-ppppp > 2016e-ppt > 2016efg-pp
hij-aaa > kl-bbb > iiii-hhh | 2012 | 2012hij-aaa > 2012kl-bbb > 2012iiii-hhh
aa-fff | 2017 | 2017aa-fff
a-pp > bbb-a | 2014 | 2014a-pp > 2014bbb-a
如何在T-SQL中执行此操作？
答案 0 :(得分：0)
Recursive CTE可能是您的最佳方法
declare @t table (ColA varchar(100), ColB varchar(100), ColD int)
insert into @t values
('abcd > e > efg', 'ppppp > ppt > pp', 2016)
, ('hij > kl > iiii', 'aaa > bbb > hhh', 2012)
, ('aa', 'fff', 2017)
, ('a > bbb', 'pp > a', 2014)
;with rcte as (
select
*, ColA2 = cast(ColA + ' >' as varchar(2000))
, ColB2 = cast(ColB + ' >' as varchar(2000))
, ca = charindex('>', ColA + ' >'), cb = charindex('>', ColB + ' >')
, ColE = cast('' as varchar(4000))
from
@t
union all
select
ColA, ColB, ColD
, cast(stuff(ColA2, 1, ca + 1, '') as varchar(2000))
, cast(stuff(ColB2, 1, cb + 1, '') as varchar(2000))
, charindex('>', stuff(ColA2, 1, ca + 1, '')), charindex('>', stuff(ColB2, 1, cb + 1, ''))
, cast(ColE + ' > ' + cast(ColD as varchar(9)) + left(ColA2, ca - 2) + '-' + left(ColB2, cb - 2) as varchar(4000))
from
rcte
where
ca > 0
)
select
ColA, ColB, ColD, stuff(ColE, 1, 3, '')
from
rcte
where
ca = 0
输出
ColA ColB ColD ColE
---------------------------------------------------------------------
a > bbb pp > a 2014 2014a-pp > 2014bbb-a
aa fff 2017 2017aa-fff
hij > kl > iiii aaa > bbb > hhh 2012 2012hij-aaa > 2012kl-bbb > 2012iiii-hhh
abcd > e > efg ppppp > ppt > pp 2016 2016abcd-ppppp > 2016e-ppt > 2016efg-pp