递归SQL将CSV拆分为表行 [英] Recursive SQL to split CSV to table rows

问题描述

在这里讨论了一个不同的问题在SO，我偶然发现递归CTE，这将表面上似乎是一个相当容易的方法来解决拆分表格行问题。

After working on a different question here on SO, I stumbled across recursive CTEs which would on the surface seem a fairly easy way to solve the "Split a csv to table rows" problem.

我把这个例子放在一起

DECLARE @InputString varchar(255) = 'a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z'

SELECT @InputString = @InputString + ','
;
with MyCTE(x,y)
as 
(
    SELECT 
        x = SUBSTRING(@InputString,0,PATINDEX('%,%',@InputString)),
        y = SUBSTRING(@InputString,PATINDEX('%,%',@InputString)+1,LEN(@InputString))
    UNION ALL
        SELECT 
            x = SUBSTRING(y,0,PATINDEX('%,%',y)),
            y = SUBSTRING(y,PATINDEX('%,%',y)+1,LEN(y))
        FROM 
            MyCTE 
        WHERE
                SUBSTRING(y,PATINDEX('%,%',y)+1,LEN(y)) <> '' OR 
            SUBSTRING(y,0,PATINDEX('%,%',y)) <> ''
)
SELECT x FROM MyCTE
OPTION (MAXRECURSION 2000);
GO

这真的是个坏主意吗？这种递归查询在SQL中的开销是什么，以及这种方法的潜在缺陷是什么。

Is this really a bad idea? What is the overhead in SQL for recursive queries like this, and what are the potential pitfalls for this kind of approach.

顺便说一句，我认为这个想法/技术可能会被用来解决此其他问题。

Incidentally, I'm thinking this idea/technique could probably be leveraged to solve this other question.

推荐答案

这将使用空字符串

DECLARE @InputString VARCHAR(1000)
    SELECT @InputString = 'a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,,1'

    SELECT SUBSTRING(',' + @InputString + ',', Number + 1,
    CHARINDEX(',', ',' + @InputString + ',', Number + 1) - Number -1)AS VALUE
    FROM master..spt_values
    WHERE type = 'p'
    AND Number <= LEN(',' + @InputString + ',') - 1
    AND SUBSTRING(',' + @InputString + ',', Number, 1) = ','
    GO

另请参阅这里的注释：S / a>其他一些想法

Also take a look at the comments here: Split string in SQL Server 2005+ CLR vs. T-SQL for some other ideas

这篇关于递归SQL将CSV拆分为表行的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！