根据特定模式删除子字符串

Question

我需要根据模式在 SQL 服务器数据库中删除一个子字符串：

之前：Winter_QZ6P91712017_115BPM

之后：Winter_115BPM

或

之前：cpx_Note In My Calendar_QZ6P91707044

之后：cpx_Note In My Calendar

基本上删除具有模式 _ + 12 个字符.

的子字符串

我已经尝试 PatIndex('_\S{12}', myCol) 获取子字符串的索引，但它不匹配任何内容。

Answer 1

SQL 服务器不支持正则表达式。但是，考虑到您只想删除第一个 '_' 和之后的 12 个字符，您可以使用 CHARINDEX 找到所述下划线的位置，然后 STUFF 删除 13 个字符:

SELECT V.YourString,
       STUFF(V.YourString, CHARINDEX('_',V.YourString),13,'') AS NewString
FROM (VALUES('Winter_QZ6P91712017_115BPM'))V(YourString);

Answer 2

假设您的意思是下划线后跟 12 个不是下划线的字符，您可以使用此模式：

SELECT *,
    CASE WHEN PATINDEX('%[_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_]%', str) > 0
         THEN STUFF(str, PATINDEX('%[_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_][^_]%', str), 13, '')
         ELSE str
    END
FROM (VALUES
    ('Winter_QZ6P91712017_115BPM'),
    ('Winter_115BPM_QZ6P91712017')
) AS tests(str)

Answer 3

迟到了，但您也可以使用最新的 STRING_SPLIT 函数通过下划线分解字符串并计算下划线之间每个段的长度。如果长度 >=12，则必须通过替换函数递归地替换原始字符串中的这些部分。

drop table if exists Tbl;
drop table if exists #temptable;
create table Tbl (input nvarchar(max));
insert into Tbl VALUES
    ('Winter_QZ6P91712017_115BPM'),
    ('cpx_Note In My Calendar_QZ6P91707044'),
    ('stuff_asdasd_QZ6P91712017'),
    ('stuff_asdasd_QZ6P91712017_stuff_asdasd_QZ6P91712017'),
    ('stuff_asdasd_QZ6P917120117_stuff_asdasd_QZ6P91712017');


select 
    input, value as replacethisstring,
    rn = row_number() over (partition by input order by (select 1))
    into #temptable
from 
    (
        select 
            input,value as hyphensplit 
        from Tbl 
            cross apply string_split(input,'_')
    )T cross apply string_split(hyphensplit,' ')
where len(value)>=12

; with cte as (
select input, inputtrans= replace(input,replacethisstring,''), level=1 from #temptable where rn=1
union all
select T.input,inputtrans=replace(cte.inputtrans,T.replacethisstring,''),level=level+1 
from cte inner join #temptable T on T.input=cte.input and rn=level+1
--where level=rn 
)
select input, inputtrans 
from (
select *, rn=row_number() over (partition by input order by level desc) from cte
) T where rn=1

示例输出

根据特定模式删除子字符串

Remove Substring according to specific pattern

sql

tsql

sql-server

wildcard