自动清理 MS SQL Server Table Collation 问题

时间：2011-06-14 来源：lei1016cn

如果是一个团队项目, 并且没有约定好Collation, 在 MS SQL Server中编程就会遇到这样的问题

Cannot resolve the collation conflict between "Latin1_General_CI_AS" and "SQL_Latin1_General_CP1_CI_AS" in the equal to operation.

很简单的一个问题, 就是说因为编码问题,不同编码的字符串并不能直接进行比较, 这种有两个解决方法, 1是在 query中指定用某一个collation进行比较, 另外一个就是修改column的 collation type 来避免这种错误…

我这里的处理方法是第二种, 一个一个column改起来很累, 写了个script, 这样除了被当作 constraints比如 primary key, foreign key之外的 varchar, char,nvarchar 都可以统一修改成一个collation…

以下是代码,简短不解说…

       1: declare @CollationName varchar(500);
       2:  
       3: set @CollationName = 'SQL_Latin1_General_CP1_CI_AS'
       4:  
       5: create table #tmp (sqlStr varchar(max));
       6:  
       7: insert into #tmp
       8: select  'alter table [' + o.name + '] alter column [' + c.name + ']' + 
       9:     (case c.system_type_id when 167 then ' varchar(' when 175 then ' char(' else ' nvarchar(' end)
    + convert(varchar,c.max_length) + ') collate ' + @CollationName
from sys.columns c,
    sys.objects o 
where  c.object_id=o.object_id and o.type='U' and c.system_type_id in (167,175,231) and collation_name<>@CollationName
and c.name not in (    
    select     cc.COLUMN_NAME 
    from     
        INFORMATION_SCHEMA.TABLE_CONSTRAINTS pk ,
        INFORMATION_SCHEMA.KEY_COLUMN_USAGE cc
    where     
        cc.TABLE_NAME = pk.TABLE_NAME
        and    cc.CONSTRAINT_NAME = pk.CONSTRAINT_NAME)
 
while (exists (select * from #tmp))
begin
    declare @sqlStr varchar(max);
    select @sqlStr=(select top 1 sqlstr from #tmp);
    exec(@sqlStr)
    delete from #tmp where sqlStr=@sqlStr
end
 
drop table #tmp;
 
    
 
  
  