Sql Server：多行合并成一行，并做分组统计的两个方法

CREATE TABLE test(code varchar(50), [values] varchar(10),[count] int)
INSERT test SELECT ‘001’, ‘aa’,1
UNION ALL SELECT ‘001’, ‘bb’,2
UNION ALL SELECT ‘002’, ‘aaa’,4
UNION ALL SELECT ‘002’, ‘bbb’,5
UNION ALL SELECT ‘002’, ‘ccc’,3;

–方法一
–将多行合并成一行，并做分组统计
SELECT code,
       [values] =
       stuff(b.[values].value(‘/R[1]’, ‘nvarchar(max)’),
,
,
             ”),[count]
  FROM (SELECT  code,sum([count]) as [count]
          FROM test
         GROUP BY code) a
 CROSS apply (
        SELECT [values] =(
            SELECT N’,’ + [values] FROM test
              WHERE code = a.code
                         FOR XML PATH(”), ROOT(‘R’), TYPE
        )
) b;

–方法二

—SQL2005中的新解法   使用XML

SELECT code, data=STUFF((SELECT ‘,’+[values] FROM test t WHERE code=t1.code FOR XML PATH(”)), 1, 1, ”),sum([count]) as [count]
FROM test t1
GROUP BY code

–查询结果

–001    aa,bb    3
–002    aaa,bbb,ccc    12

drop table test