Question

我正在使用Microsoft SQL Server。我有以下查询，它按预期返回选择的列，排名和按end_date排序的查询。 end_date的类型为datetime。

select 
    rank() over (order by end_date desc) rownum, 
    salary, name, end_date
from
    employee
order by 
    end_date desc

如何修改和/或添加它以仅提取返回的第一行？可以通过对上述查询进行修改来完成，还是必须添加其他逻辑？

Answer 1

使用子查询：

SELECT *
FROM (select rank() over (order by endmonth desc) rownum,
             salary, name, endmonth
      from employee e) s
WHERE rownum = 1;

或WITH TIES选项（特定于SQL Server-可能会生成较差的执行计划）：

select TOP 1 WITH TIES rank() over (order by endmonth desc) rownum,
                 salary, name, endmonth, name 
from employee e
order by rownum

Answer 2

通常，最快的方法（具有正确的索引）是相关子查询：

<script src="https://cdnjs.cloudflare.com/ajax/libs/babel-standalone/6.21.1/babel.min.js"></script>

您想要的索引位于select 1 as rownum, salary, name, endmonth, name from employee e where e.endmonth = (select max(e2.endmonth) from employee e2 );上。

如果您知道会有一行，那么将employee(endmonth)与order by（或与您的数据库等效）是最好的方法：

fetch first

如果您的数据库支持select 1 as rownum, salary, name, endmonth, name from employee e order by e.endmonth desc fetch first one row only;，则可以使用它。例如，在SQL Server中：

with ties

Answer 3

就像Lukasz所说的那样，但是您也可以使用ROW_NUMBER，它不会像RANK一样返回重复的值，因此您可以首先明确地获得一行。

查看：SQL RANK() versus ROW_NUMBER()

SELECT *
FROM (select ROW_NUMBER() over (order by endmonth desc) rownum,
             salary, name, endmonth, name 
      from employee e) s
WHERE rownum = 1;

Answer 4

select rank() over (order by endmonth desc) rownum, salary, name, 
endmonth, name 
from employee e
limit 1;

希望这会有所帮助

SQL Server排名：从一组行中获取第一行

4 个答案: