如下：
比如有两行数据，有可能N行，N为1到无限：
表a中有列t:
内容为：
第一行t数据：671:233,789:136,736:202,539:244,505:88,504:186,500:190,332:217,204:211,205:93,362:184,410:246,615:372,685:317,685:317,429:399,429:401,476:492,991:364,1026:249,1025:166,949:150,733:71,175:65,170:69,175:173,475:309,616:309,810:324,813:324,767:226,763:175,770:108,770:101,735:74,470:106,471:176,484:213,489:222,598:292,1140:249,1187:194,1057:344,977:256,959:262,536:171,533:115,576:186,647:233,647:233
第二行t数据：
754:73,734:193,718:229,622:304,608:302,891:303,860:218,832:184,803:129,745:52,752:215,711:357,964:497,470:512,264:410,192:279,192:46,286:179,598:215,577:121,681:219,782:302,778:313,700:330
所有行数据中以逗号“，”分割，冒号“:”右边的最大值是多少？
------解决方案--------------------
CREATE TABLE tb(id INT,col VARCHAR(max))
INSERT INTO tb
SELECT
1,'671:233,789:136,736:202,539:244,505:88,504:186,500:190,332:217,204:211,205:93,362:184,410:246,615:372,685:317,685:317,429:399,429:401,476:492,991:364,1026:249,1025:166,949:150,733:71,175:65,170:69,175:173,475:309,616:309,810:324,813:324,767:226,763:175,770:108,770:101,735:74,470:106,471:176,484:213,489:222,598:292,1140:249,1187:194,1057:344,977:256,959:262,536:171,533:115,576:186,647:233,647:233'
UNION ALL
SELECT
2,'754:73,734:193,718:229,622:304,608:302,891:303,860:218,832:184,803:129,745:52,752:215,711:357,964:497,470:512,264:410,192:279,192:46,286:179,598:215,577:121,681:219,782:302,778:313,700:330'
;WITH maco AS
(
SELECT A.id, B.col,RIGHT(b.COL,LEN(B.COL)-CHARINDEX(':',B.COL)) AS rcol
FROM(
SELECT id, col = CONVERT(xml,'<root><v>' + REPLACE(col, ',', '</v><v>') + '</v></root>') FROM tb
)A
OUTER APPLY(
SELECT col = N.v.value('.', 'varchar(100)') FROM A.col.nodes('/root/v') N(v)
)B
)
SELECT id,MAX(rcol+0) AS maxrco FROM maco GROUP BY id
/*
id maxrco
----------- -----------
1 492
2 512
*/