我需要在sql server中使用一个函数来构建下面示例中的所有更改的单词; 对于长度为n的输入字必须构建2 ^ n改变的字; 例如,如果函数的输入是
"I"
函数的输出应为
I
-
该功能的输入是
"am"
函数的输出应为
am
-m
a-
--
该功能的输入是
"sql"
函数的输出应为
sql
-ql
s-l
sq-
--l
s--
-q-
---
答案 0 :(得分:9)
您可以使用数字表(master..spt_values)和stuff
循环执行此操作。
declare @Word varchar(10) = 'sql'
declare @T table
(
Word varchar(10)
)
insert into @T values (@Word)
while not exists(select *
from @T
where Word = replicate('-', len(@Word)))
begin
insert into @T(Word)
select distinct stuff(T.Word, N.number, 1, '-')
from @T as T
cross join
master..spt_values as N
where N.type = 'P' and
N.number between 1 and len(@Word) and
stuff(T.Word, N.number, 1, '-') not in (select Word from @T)
end
select *
from @T
http://data.stackexchange.com/stackoverflow/q/122334/
或者您可以使用reqursive CTE
declare @Word varchar(10) = 'sql'
;with C as
(
select @Word as Word,
0 as Iteration
union all
select cast(stuff(Word, N.number, 1, '-') as varchar(10)),
Iteration + 1
from C
cross join
master..spt_values as N
where N.type = 'P' and
N.number between 1 and len(@Word) and
Iteration < len(@Word)
)
select distinct Word
from C
http://data.stackexchange.com/stackoverflow/q/122337/
<强>更新强>
正如OP在评论中指出的那样,递归CTE版本非常慢。使用带有7个字母的单词,CTE返回了960800行。
答案 1 :(得分:6)
此递归CTE
declare @input varchar(25)
set @input = 'SQL'
;WITH cte
AS (SELECT Stuff(@input, v.NUMBER, 1, '-') OUTPUT,
0 LEVEL
FROM MASTER..spt_values v
WHERE TYPE = 'P'
AND NUMBER BETWEEN 1 AND Len(@input)
UNION ALL
SELECT Stuff(cte.OUTPUT, v.NUMBER, 1, '-') OUTPUT,
cte.LEVEL + 1 AS LEVEL
FROM MASTER..spt_values v,
cte
WHERE TYPE = 'P'
AND cte.LEVEL + 1 < Len(@input)
AND NUMBER BETWEEN 1 AND Len(@input)) SELECT DISTINCT OUTPUT
FROM cte
UNION
SELECT @INPUT
ORDER BY OUTPUT
产生以下输出
---
--l
-q-
-ql
s--
s-l
sq-
sql
我留给你一个功能。
请参阅此data.se query