Question

MS-SQL的row_number（）可以使用以下查询进行模拟，但我需要MS-SQL实现此查询。

 SELECT @row_number:= IF(@column_name = column_name, @row_number+1, 1) AS row_no,
 @column_name:= column_name AS column_name
 FROM db_name, (SELECT @row_number := 0, @column_name := '') AS temp
 ORDER BY column_name;

Answer 1

SELECT *, ROW_NUMBER() OVER (ORDER BY column_name)
FROM db_name

修改

没有row_number（）

select *, (     
    select count(*)     
    from db_name p2     
    where p1.column_name >= p2.column_name ) 
as cnt from db_name p1

根据Damien的评论编辑2

非row_number（）版本仅在column_name列的内容不包含重复项时才有效。如果不是，则row_number结果列中将出现重复和跳过。

Answer 2

我认为没有row_number()的正确MS-SQL实现更像是：

with dn as (
      select dn.*, checksum(newid()) as rnd
      from db_name dn
     )
select dn.*,
       (select count(*)
        from dn dn2
        where dn2.column_name = dn.column_name and
              dn2.rnd < db.rnd
       ) 
from dn;

当然，如果id列可用（每行唯一一列），那么最好使用它：

select dn.*,
       (select count(*)
        from db_name dn2
        where dn2.column_name = dn.column_name and
              dn2.id < db.id
       ) 
from db_name dn;

MSSQL相当于这个模拟row_number（）的mysql查询

2 个答案: