使用SQL Server,我如何分割一个字符串,以便我可以访问项目x?

拿一根“你好,约翰·史密斯”的绳子。我如何通过空格分割字符串并访问索引1应该返回“约翰”的项目?


当前回答

可以利用Number表进行字符串解析。

创建一个物理数字表:

    create table dbo.Numbers (N int primary key);
    insert into dbo.Numbers
        select top 1000 row_number() over(order by number) from master..spt_values
    go

创建具有1000000行的测试表

    create table #yak (i int identity(1,1) primary key, array varchar(50))

    insert into #yak(array)
        select 'a,b,c' from dbo.Numbers n cross join dbo.Numbers nn
    go

创建函数

    create function [dbo].[ufn_ParseArray]
        (   @Input      nvarchar(4000), 
            @Delimiter  char(1) = ',',
            @BaseIdent  int
        )
    returns table as
    return  
        (   select  row_number() over (order by n asc) + (@BaseIdent - 1) [i],
                    substring(@Input, n, charindex(@Delimiter, @Input + @Delimiter, n) - n) s
            from    dbo.Numbers
            where   n <= convert(int, len(@Input)) and
                    substring(@Delimiter + @Input, n, 1) = @Delimiter
        )
    go

使用情况(在我的笔记本电脑上40秒内输出3mil行)

    select * 
    from #yak 
    cross apply dbo.ufn_ParseArray(array, ',', 1)

清理

    drop table dbo.Numbers;
    drop function  [dbo].[ufn_ParseArray]

这里的性能并不惊人,但在100万行表上调用函数并不是最好的主意。如果将字符串拆分到多行,我会避免使用该函数。

其他回答



    Alter Function dbo.fn_Split
    (
    @Expression nvarchar(max),
    @Delimiter  nvarchar(20) = ',',
    @Qualifier  char(1) = Null
    )
    RETURNS @Results TABLE (id int IDENTITY(1,1), value nvarchar(max))
    AS
    BEGIN
       /* USAGE
            Select * From dbo.fn_Split('apple pear grape banana orange honeydew cantalope 3 2 1 4', ' ', Null)
            Select * From dbo.fn_Split('1,abc,"Doe, John",4', ',', '"')
            Select * From dbo.fn_Split('Hello 0,"&""&&&&', ',', '"')
       */

       -- Declare Variables
       DECLARE
          @X     xml,
          @Temp  nvarchar(max),
          @Temp2 nvarchar(max),
          @Start int,
          @End   int

       -- HTML Encode @Expression
       Select @Expression = (Select @Expression For XML Path(''))

       -- Find all occurences of @Delimiter within @Qualifier and replace with |||***|||
       While PATINDEX('%' + @Qualifier + '%', @Expression) > 0 AND Len(IsNull(@Qualifier, '')) > 0
       BEGIN
          Select
             -- Starting character position of @Qualifier
             @Start = PATINDEX('%' + @Qualifier + '%', @Expression),
             -- @Expression starting at the @Start position
             @Temp = SubString(@Expression, @Start + 1, LEN(@Expression)-@Start+1),
             -- Next position of @Qualifier within @Expression
             @End = PATINDEX('%' + @Qualifier + '%', @Temp) - 1,
             -- The part of Expression found between the @Qualifiers
             @Temp2 = Case When @End < 0 Then @Temp Else Left(@Temp, @End) End,
             -- New @Expression
             @Expression = REPLACE(@Expression,
                                   @Qualifier + @Temp2 + Case When @End < 0 Then '' Else @Qualifier End,
                                   Replace(@Temp2, @Delimiter, '|||***|||')
                           )
       END

       -- Replace all occurences of @Delimiter within @Expression with '</fn_Split>&ltfn_Split>'
       -- And convert it to XML so we can select from it
       SET
          @X = Cast('&ltfn_Split>' +
                    Replace(@Expression, @Delimiter, '</fn_Split>&ltfn_Split>') +
                    '</fn_Split>' as xml)

       -- Insert into our returnable table replacing '|||***|||' back to @Delimiter
       INSERT @Results
       SELECT
          "Value" = LTRIM(RTrim(Replace(C.value('.', 'nvarchar(max)'), '|||***|||', @Delimiter)))
       FROM
          @X.nodes('fn_Split') as X(C)

       -- Return our temp table
       RETURN
    END

递归CTE解决方案与服务器疼痛,测试它

MS SQL Server 2008模式设置:

create table Course( Courses varchar(100) );
insert into Course values ('Hello John Smith');

查询1:

with cte as
   ( select 
        left( Courses, charindex( ' ' , Courses) ) as a_l,
        cast( substring( Courses, 
                         charindex( ' ' , Courses) + 1 , 
                         len(Courses ) ) + ' ' 
              as varchar(100) )  as a_r,
        Courses as a,
        0 as n
     from Course t
    union all
      select 
        left(a_r, charindex( ' ' , a_r) ) as a_l,
        substring( a_r, charindex( ' ' , a_r) + 1 , len(a_R ) ) as a_r,
        cte.a,
        cte.n + 1 as n
    from Course t inner join cte 
         on t.Courses = cte.a and len( a_r ) > 0

   )
select a_l, n from cte
--where N = 1

结果:

|    A_L | N |
|--------|---|
| Hello  | 0 |
|  John  | 1 |
| Smith  | 2 |

Aaron Bertrand的回答很好,但也有缺陷。它不能准确地将空格作为分隔符处理(就像最初问题中的示例一样),因为长度函数将空格带在后面。

下面是他的代码,稍微调整了一下,允许使用空格分隔符:

CREATE FUNCTION [dbo].[SplitString]
(
    @List NVARCHAR(MAX),
    @Delim VARCHAR(255)
)
RETURNS TABLE
AS
    RETURN ( SELECT [Value] FROM 
      ( 
        SELECT 
          [Value] = LTRIM(RTRIM(SUBSTRING(@List, [Number],
          CHARINDEX(@Delim, @List + @Delim, [Number]) - [Number])))
        FROM (SELECT Number = ROW_NUMBER() OVER (ORDER BY name)
          FROM sys.all_objects) AS x
          WHERE Number <= LEN(@List)
          AND SUBSTRING(@Delim + @List, [Number], LEN(@Delim+'x')-1) = @Delim
      ) AS y
    );

在我看来,你们把事情搞得太复杂了。只需要创建一个CLR UDF就可以了。

using System;
using System.Data;
using System.Data.SqlClient;
using System.Data.SqlTypes;
using Microsoft.SqlServer.Server;
using System.Collections.Generic;

public partial class UserDefinedFunctions {
  [SqlFunction]
  public static SqlString SearchString(string Search) {
    List<string> SearchWords = new List<string>();
    foreach (string s in Search.Split(new char[] { ' ' })) {
      if (!s.ToLower().Equals("or") && !s.ToLower().Equals("and")) {
        SearchWords.Add(s);
      }
    }

    return new SqlString(string.Join(" OR ", SearchWords.ToArray()));
  }
};

通过delimeter函数得到字符串的n个部分:

create function GetStringPartByDelimeter (
    @value as nvarchar(max),
    @delimeter as nvarchar(max),
    @position as int
) returns NVARCHAR(MAX) 
AS BEGIN
    declare @startPos as int
    declare @endPos as int
    set @endPos = -1
    while (@position > 0 and @endPos != 0) begin
        set @startPos = @endPos + 1
        set @endPos = charindex(@delimeter, @value, @startPos)

        if(@position = 1) begin
            if(@endPos = 0)
                set @endPos = len(@value) + 1

            return substring(@value, @startPos, @endPos - @startPos)
        end

        set @position = @position - 1
    end

    return null
end

以及用法:

select dbo.GetStringPartByDelimeter ('a;b;c;d;e', ';', 3)

返回:

c