poi在读取execl文件时，数据的列号都是数字的，但是往往我们回写message时需要的是字母的列号，这里提供一个数据库的方式，建立一张数字--字母对应关系表(sqlServer数据库，其他类似)，下次提供以java类的方式对数字和字母进行转换

--第一步
WITH seq AS
(
    SELECT 
        ROW_NUMBER() OVER (ORDER BY x.alpha + y.alpha + z.alpha) AS Id,
        CONVERT(nchar(3), x.alpha + y.alpha + z.alpha) AS Result
    FROM 
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) x(alpha),
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) y(alpha),
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) z(alpha)
    WHERE
       NOT (NOT x.alpha BETWEEN 'A' AND 'Z')
),
seq2 as(
	SELECT 
        ROW_NUMBER() OVER (ORDER BY x.alpha + y.alpha) AS Id,
        CONVERT(nchar(3), x.alpha + y.alpha) AS Result
    FROM 
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) x(alpha),
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) y(alpha)
    WHERE
       NOT (NOT x.alpha BETWEEN 'A' AND 'Z')
) ,
seq3 as(
	SELECT 
        ROW_NUMBER() OVER (ORDER BY x.alpha) AS Id,
        CONVERT(nchar(3), x.alpha) AS Result
    FROM 
        (
            VALUES 
            ('A'), ('B'), ('C'), ('D'), ('E'), ('F'), ('G'), ('H'), ('I'), ('J'), 
            ('K'), ('L'), ('M'), ('N'), ('O'), ('P'), ('Q'), ('R'), ('S'), ('T'), 
            ('U'), ('V'), ('W'), ('X'), ('Y'), ('Z')
        ) x(alpha)
    WHERE
       NOT (NOT x.alpha BETWEEN 'A' AND 'Z')
)
insert into id_to_num select * from (
select ROW_NUMBER() over(ORDER BY result) as id,result  from seq3
union all
select ROW_NUMBER() over(ORDER BY result) as id,result from seq2
union all
SELECT ROW_NUMBER() over(ORDER BY result) as id,result FROM seq WHERE result <= 'XFD') aa

--第二步修改数字编号
update id_to_num set execle_rowid = execle_rowid+26 where len(execle_rownum)=2；
update id_to_num set execle_rowid = execle_rowid+703 where len(execle_rownum)=3

--说明:
--如果直接按照字母顺序插入的话，数据库会把A/AA/AAA三个排在相邻的三个位置，

--这个显然不是我们想要的，因此首先按照普通的字母排序分别把三部分数据插入表中，

--然后修改不同分段的编号，其中一位长度的是26个字母，两位长度的是702个字母组以

--这两个数自分别把两位长度的编号加上26和三维长度的编号都加上703,编号就都可以修改为正常的