在Chris Shaffer的回答之上：

如果您的数据可能会重复，例如

Tom
Ali
John
Ali
Tom
Mike

而不是汤姆、阿里、约翰、阿里、汤姆、迈克

您可以使用DISTINCT避免重复，并让Tom、Ali、John、Mike：

DECLARE @Names VARCHAR(8000)
SELECT DISTINCT @Names = COALESCE(@Names + ',', '') + Name
FROM People
WHERE Name IS NOT NULL
SELECT @Names

我并没有做过任何关于性能的分析，因为我的列表中只有不到10项，但在查看了30多个答案后，我感到很惊讶，我仍然对已经给出的类似答案进行了修改，类似于对单个组列表使用COALESCE，甚至不必设置我的变量（无论如何默认为NULL），并且它假设我的源数据表中的所有条目都是非空的：

DECLARE @MyList VARCHAR(1000), @Delimiter CHAR(2) = ', '
SELECT @MyList = CASE WHEN @MyList > '' THEN @MyList + @Delimiter ELSE '' END + FieldToConcatenate FROM MyData

我确信COALENCE内部使用了相同的想法。让我们希望微软不会在我身上改变这一点。

使用递归查询，您可以执行以下操作：

-- Create example table
CREATE TABLE tmptable (NAME VARCHAR(30)) ;

-- Insert example data
INSERT INTO tmptable VALUES('PETER');
INSERT INTO tmptable VALUES('PAUL');
INSERT INTO tmptable VALUES('MARY');

-- Recurse query
with tblwithrank as (
select * , row_number() over(order by name) rang , count(*) over() NbRow
from tmptable
),
tmpRecursive as (
select *, cast(name as varchar(2000)) as AllName from tblwithrank  where rang=1
union all
select f0.*,  cast(f0.name + ',' + f1.AllName as varchar(2000)) as AllName 
from tblwithrank f0 inner join tmpRecursive f1 on f0.rang=f1.rang +1 
)
select AllName from tmpRecursive
where rang=NbRow

SQL Server 2005或更高版本

CREATE TABLE dbo.Students
(
    StudentId INT
    , Name VARCHAR(50)
    , CONSTRAINT PK_Students PRIMARY KEY (StudentId)
);

CREATE TABLE dbo.Subjects
(
    SubjectId INT
    , Name VARCHAR(50)
    , CONSTRAINT PK_Subjects PRIMARY KEY (SubjectId)
);

CREATE TABLE dbo.Schedules
(
    StudentId INT
    , SubjectId INT
    , CONSTRAINT PK__Schedule PRIMARY KEY (StudentId, SubjectId)
    , CONSTRAINT FK_Schedule_Students FOREIGN KEY (StudentId) REFERENCES dbo.Students (StudentId)
    , CONSTRAINT FK_Schedule_Subjects FOREIGN KEY (SubjectId) REFERENCES dbo.Subjects (SubjectId)
);

INSERT dbo.Students (StudentId, Name) VALUES
    (1, 'Mary')
    , (2, 'John')
    , (3, 'Sam')
    , (4, 'Alaina')
    , (5, 'Edward')
;

INSERT dbo.Subjects (SubjectId, Name) VALUES
    (1, 'Physics')
    , (2, 'Geography')
    , (3, 'French')
    , (4, 'Gymnastics')
;

INSERT dbo.Schedules (StudentId, SubjectId) VALUES
    (1, 1)        --Mary, Physics
    , (2, 1)    --John, Physics
    , (3, 1)    --Sam, Physics
    , (4, 2)    --Alaina, Geography
    , (5, 2)    --Edward, Geography
;

SELECT
    sub.SubjectId
    , sub.Name AS [SubjectName]
    , ISNULL( x.Students, '') AS Students
FROM
    dbo.Subjects sub
    OUTER APPLY
    (
        SELECT
            CASE ROW_NUMBER() OVER (ORDER BY stu.Name) WHEN 1 THEN '' ELSE ', ' END
            + stu.Name
        FROM
            dbo.Students stu
            INNER JOIN dbo.Schedules sch
                ON stu.StudentId = sch.StudentId
        WHERE
            sch.SubjectId = sub.SubjectId
        ORDER BY
            stu.Name
        FOR XML PATH('')
    ) x (Students)
;

虽然为时已晚，而且已经有了许多解决方案。下面是MySQL的简单解决方案：

SELECT t1.id,
        GROUP_CONCAT(t1.id) ids
 FROM table t1 JOIN table t2 ON (t1.id = t2.id)
 GROUP BY t1.id

在SQL Server中实现这一点的一种方法是将表内容返回为XML（对于XML原始），将结果转换为字符串，然后用“，”替换标记。