加入逗号分隔的数据列

时间:2013-05-12 12:13:21

标签: sql sql-server tsql

我的table1是:

T1

col1    col2
 C1     john
 C2     alex
 C3     piers
 C4     sara

所以表2:

T2

col1    col2
 R1     C1,C2,C4
 R2     C3,C4
 R3     C1,C4

如何结果?:

查询结果

col1      col2
 R1       john,alex,sara
 R2       piers,sara
 R3       john,sara

请帮帮我?

7 个答案:

答案 0 :(得分:21)

理想情况下,您最好的解决方案是规范化Table2,这样您就不会存储以逗号分隔的列表。

一旦您将此数据规范化,您就可以轻松查询数据。新表结构可能与此类似:

CREATE TABLE T1
(
  [col1] varchar(2), 
  [col2] varchar(5),
  constraint pk1_t1 primary key (col1)
);

INSERT INTO T1
    ([col1], [col2])
VALUES
    ('C1', 'john'),
    ('C2', 'alex'),
    ('C3', 'piers'),
    ('C4', 'sara')
;

CREATE TABLE T2
(
  [col1] varchar(2), 
  [col2] varchar(2),
  constraint pk1_t2 primary key (col1, col2),
  constraint fk1_col2 foreign key (col2) references t1 (col1)
);

INSERT INTO T2
    ([col1], [col2])
VALUES
    ('R1', 'C1'),
    ('R1', 'C2'),
    ('R1', 'C4'),
    ('R2', 'C3'),
    ('R2', 'C4'),
    ('R3', 'C1'),
    ('R3', 'C4')
;

对表进行规范化将使您更容易通过连接表来查询数据:

select t2.col1, t1.col2
from t2
inner join t1
  on t2.col2 = t1.col1

请参阅Demo

然后,如果您想将数据显示为以逗号分隔的列表,则可以使用FOR XML PATHSTUFF

select distinct t2.col1, 
  STUFF(
         (SELECT distinct ', ' + t1.col2
          FROM t1
          inner join t2 t
            on t1.col1 = t.col2
          where t2.col1 = t.col1
          FOR XML PATH ('')), 1, 1, '') col2
from t2;

请参阅Demo

如果您无法规范化数据,那么您可以执行以下操作。

首先,您可以创建一个拆分函数,将存储在列表中的数据转换为可以连接的行。拆分功能与此类似:

CREATE FUNCTION [dbo].[Split](@String varchar(MAX), @Delimiter char(1))       
returns @temptable TABLE (items varchar(MAX))       
as       
begin      
    declare @idx int       
    declare @slice varchar(8000)       

    select @idx = 1       
        if len(@String)<1 or @String is null  return       

    while @idx!= 0       
    begin       
        set @idx = charindex(@Delimiter,@String)       
        if @idx!=0       
            set @slice = left(@String,@idx - 1)       
        else       
            set @slice = @String       

        if(len(@slice)>0)  
            insert into @temptable(Items) values(@slice)       

        set @String = right(@String,len(@String) - @idx)       
        if len(@String) = 0 break       
    end   
return 
end;

当您使用split,function时,您可以将数据保留在多行中,也可以将值连接回逗号分隔的列表中:

;with cte as
(
  select c.col1, t1.col2
  from t1
  inner join 
  (
    select t2.col1, i.items col2
    from t2
    cross apply dbo.split(t2.col2, ',') i
  ) c
    on t1.col1 = c.col2
) 
select distinct c.col1, 
  STUFF(
         (SELECT distinct ', ' + c1.col2
          FROM cte c1
          where c.col1 = c1.col1
          FOR XML PATH ('')), 1, 1, '') col2
from cte c

请参阅Demo

获得结果的最后一种方法是直接应用FOR XML PATH

select col1, 
(
  select ', '+t1.col2
  from t1
  where ','+t2.col2+',' like '%,'+cast(t1.col1 as varchar(10))+',%'
  for xml path(''), type
).value('substring(text()[1], 3)', 'varchar(max)') as col2
from t2;

请参阅SQL Fiddle with Demo

答案 1 :(得分:4)

以下是一种在没有函数的情况下拆分数据的方法,然后使用标准XML PATH方法获取CSV列表:

with CTE as
(
  select T2.col1
    , T1.col2
  from T2
    inner join T1 on charindex(',' + T1.col1 + ',', ',' + T2.col2 + ',') > 0
)
select T2.col1
  , col2 = stuff(
      (
        select ',' + CTE.col2
        from CTE
        where T2.col1 = CTE.col1
        for xml path('')
      )
      , 1
      , 1
      , ''
    )
from T2

SQL Fiddle with demo

正如在这个问题的其他地方已经提到的那样,很难以任何有效的方式查询这种非规范化数据,因此您的首要任务应该是调查更新表结构,但这至少可以让你需要的结果。

答案 2 :(得分:2)

如果您想在oracle中执行此任务,我们可以使用listagg并轻松完成此任务。

我在SQL SERVER中表现不佳,但我在Sqlserver中为listagg搜索了可能等效的avaialble,我得到了同样的函数Stuff - Check this

所以使用东西你可以尝试使用以下查询 -

SELECT T2.Col1,
       Stuff((SELECT ',' + CAST(T1.Col2 AS VARCHAR(100))
               FROM T1
              WHERE T2.Col2 LIKE T1.Col1
                FOR Xml Path('')),
             1,
             1,
             '')
  FROM T2

答案 3 :(得分:1)

首先在tbl2上为split col2写一个表值函数。

CREATE FUNCTION [dbo].[Split](@String varchar(100), @Delimiter char(1))       
returns @temptable TABLE (items VARCHAR(5))       
as       
begin       
    declare @idx int       
    declare @slice VARCHAR(5)

    select @idx = 1       
        if len(@String)<1 or @String is null  return       

    while @idx!= 0       
    begin       
        set @idx = charindex(@Delimiter,@String)       
        if @idx!=0       
            set @slice = left(@String,@idx - 1)       
        else       
            set @slice = @String       

        if(len(@slice)>0)  
            insert into @temptable(Items) values(@slice)       

        set @String = right(@String,len(@String) - @idx)       
        if len(@String) = 0 break       
    end   
return       
end  

Go

;WITH    SplitList
          AS ( SELECT   T2.Col1 ,
                        T1.Col2
               FROM     T2
                        CROSS APPLY dbo.Split(T2.Col2, ',') S
                        INNER JOIN T1 ON T1.Col1 = S.Items
             )
    SELECT  T2.Col1 ,
            STUFF(( SELECT  ', ' + SplitList.Col2
                    FROM    SplitList
                    WHERE   SplitList.Col1 = T2.Col1
                  FOR
                    XML PATH('')
                  ), 1, 2, '')
    FROM    T2       

答案 4 :(得分:0)

使用标准SQL无法解决此任务。在Oracle中,我会编写一个存储函数(PL / SQL)来解析Name-ID-string(T2 col2)并解析名称。不知道在Transact-SQL中是否可行,但它的效率很低。

T2是一个设计糟糕,没有规范化的表格。那就是问题所在。 如果你将它标准化,那么每个Name-ID有一行(T2中的col 2),你可以通过两个表的简单连接获得名称列表。 要生成所需的输出格式(逗号分隔),您需要编写除SQL之外的其他内容 - 可能是存储过程或其他迭代结果集的内容。

答案 5 :(得分:0)

如果你像我一样,并且你是CTE的坚持者,特别是递归CTE应该是STUFF和XML Path:

DECLARE @T1 TABLE (
    col1 CHAR(2),
    col2 VARCHAR(10)
)
INSERT INTO @T1
VALUES  ('C1', 'john'),
        ('C2', 'alex'),
        ('C3', 'piers'),
        ('C4', 'sara');

DECLARE @T2 TABLE (
    col1 CHAR(2),
    col2 CHAR(100)
)
INSERT INTO @T2
VALUES  ('R1', 'C1,C2,C4'),
        ('R2', 'C3,C4'),
        ('R3', 'C1,C4');

WITH T2Sorted AS (
    SELECT col1, col2, RN = ROW_NUMBER() OVER (ORDER BY col1) FROM @T2
), CTERecursionOnT2 AS (
    SELECT RN, col1, col2, 0 AS PrevCharIndex, CHARINDEX(',', col2, 1) AS NextCharIndex FROM T2Sorted
    UNION ALL
    SELECT a.RN, a.col1, a.col2, b.NextCharIndex, CHARINDEX(',', a.col2, b.NextCharIndex + 1) 
    FROM T2Sorted a
    JOIN CTERecursionOnT2 b ON a.RN = b.RN
    WHERE b.NextCharIndex > 0
), CTEIndividualCol2Items AS (
    SELECT *, SUBSTRING(col2, PrevCharIndex + 1, CASE WHEN NextCharIndex = 0 THEN LEN(col2) ELSE NextCharIndex - 1 END - PrevCharIndex) AS itemCol2 
    FROM CTERecursionOnT2
), CTELookupT1 AS (
    SELECT a.col1, b.col2, RN = ROW_NUMBER() OVER (PARTITION BY a.col1 ORDER BY a.PrevCharIndex)
    FROM CTEIndividualCol2Items a
    JOIN @T1 b ON a.itemCol2 = b.col1
), CTERecursionOnLookupT1 AS (
    SELECT col1, CAST(col2 AS VARCHAR(MAX)) AS col2, RN
    FROM CTELookupT1 
    WHERE RN = 1

    UNION ALL

    SELECT a.col1, b.col2 + ',' + a.col2, a.RN
    FROM CTELookupT1 a
    JOIN CTERecursionOnLookupT1 b ON a.col1 = b.col1 AND a.RN = b.RN + 1
), CTEFinal AS (
    SELECT *, RNDesc = ROW_NUMBER() OVER (PARTITION BY col1 ORDER BY RN DESC)
    FROM CTERecursionOnLookupT1
)
SELECT col1, col2
FROM CTEFinal
WHERE RNDesc = 1
ORDER BY col1

显然你可以将第一个递归部分拆分成单独的函数,因为已经商定的解决方案建议即CTERecursionOnT2,因此CTEIndividualCol2Items可以作为您的替代拆分函数(我也会包括订单ID),因此:

;WITH CTEIndividualCol2Items AS (
    SELECT a.col1, b.value as itemCol2, b.id AS PrevCharIndex
    FROM @T2 a
    CROSS APPLY (
        SELECT id, items FROM dbo.Split(a.col2, ',')
    ) b
) ...

你拆分功能:

CREATE FUNCTION dbo.Split(@String varchar(100), @Delimiter char(1))
RETURNS TABLE
AS
RETURN 
(
    WITH CTERecursion AS (
        SELECT id = 1, PrevCharIndex = 0, NextCharIndex = CHARINDEX(@Delimiter, @String, 1)
        UNION ALL
        SELECT id + 1, NextCharIndex, CHARINDEX(@Delimiter, @String, NextCharIndex + 1) FROM CTERecursion WHERE NextCharIndex > 0
    )
    SELECT Id, items = SUBSTRING(@String, PrevCharindex + 1, (CASE WHEN NextCharIndex = 0 THEN LEN(@String) ELSE NextCharIndex - 1 END) - PrevCharIndex)
    FROM CTERecursion
    WHERE @String > ''
)

答案 6 :(得分:-1)

Select T2.col1, group_concate(T1.col2)
From T1 join T2 on locate(T2.Col2, T1.Col1) > 0
group by T2.col1