我有一个如下表
+---------+-----------------+-----------------+--------------------+--------------+-------+--------+
| OrderID | OrderDate | ShipDate | CustomerName | City | Sales | Profit |
+---------+-----------------+-----------------+--------------------+--------------+-------+--------+
| 1001 | 11/8/2016 0:00 | 11/11/2016 0:00 | Claire Gute | Henderson | 262 | 42 |
| 1002 | 11/8/2016 0:00 | 11/11/2016 0:00 | Darrin Van Huff | Henderson | 732 | 220 |
| 1003 | 6/12/2016 0:00 | 6/16/2016 0:00 | Claire Gute | Henderson | 15 | 7 |
| 1004 | 10/11/2015 0:00 | 10/18/2015 0:00 | Brosina Hoffman | Henderson | 958 | -383 |
| 1005 | 10/11/2015 0:00 | 10/18/2015 0:00 | Claire Gute | Henderson | 22 | 3 |
| 1006 | 6/9/2014 0:00 | 6/14/2014 0:00 | Irene Maddox | Fort Worth | 49 | 14 |
| 1007 | 6/9/2014 0:00 | 6/14/2014 0:00 | Harold Pawlan | Fort Worth | 7 | 2 |
| 1008 | 6/9/2014 0:00 | 6/14/2014 0:00 | Pete Kriz | Fort Worth | 907 | 91 |
| 1009 | 6/9/2014 0:00 | 6/14/2014 0:00 | Irene Maddox | Fort Worth | 19 | 6 |
| 1010 | 6/9/2014 0:00 | 6/14/2014 0:00 | Zuschuss Donatelli | Fort Worth | 115 | 34 |
| 1011 | 6/9/2014 0:00 | 6/14/2014 0:00 | Ken Black | Philadelphia | 1706 | 85 |
| 1012 | 6/9/2014 0:00 | 6/14/2014 0:00 | Sandra Flanagan | Philadelphia | 911 | 68 |
| 1013 | 4/15/2017 0:00 | 4/20/2017 0:00 | Ken Black | Philadelphia | 16 | 5 |
| 1014 | 12/5/2016 0:00 | 12/10/2016 0:00 | Eric Hoffmann | Philadelphia | 408 | 133 |
| 1015 | 11/22/2015 0:00 | 11/26/2015 0:00 | Tracy Blumstein | Naperville | 69 | -124 |
| 1016 | 11/22/2015 0:00 | 11/26/2015 0:00 | Matt Abelman | Melbourne | 3 | -4 |
| 1017 | 11/22/2015 0:00 | 11/26/2015 0:00 | Gene Hale | Melbourne | 666 | 13 |
| 1018 | 5/13/2014 0:00 | 5/15/2014 0:00 | Steve Nguyen | Melbourne | 56 | 10 |
| 1019 | 5/13/2014 0:00 | 5/15/2014 0:00 | Linda Cazamias | Dover | 9 | 2 |
| 1020 | 5/13/2014 0:00 | 5/15/2014 0:00 | Ruben Ausman | Dover | 113 | 35 |
+---------+-----------------+-----------------+--------------------+--------------+-------+--------+
我在下面的查询中编写，以基于rank的给定顺序在每个city中获得Sales
SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER(PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
我得到的结果绝对是正确的
+---------+---------------------+---------------------+--------------------+--------------+-------+-------+
| OrderID | OrderDate | ShipDate | CustomerName | City | Sales | RankA |
+---------+---------------------+---------------------+--------------------+--------------+-------+-------+
| 1020 | 2014-05-13 00:00:00 | 2014-05-15 00:00:00 | Ruben Ausman | Dover | 113 | 1 |
| 1019 | 2014-05-13 00:00:00 | 2014-05-15 00:00:00 | Linda Cazamias | Dover | 9 | 2 |
| 1008 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Pete Kriz | Fort Worth | 907 | 1 |
| 1010 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Zuschuss Donatelli | Fort Worth | 115 | 2 |
| 1006 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Irene Maddox | Fort Worth | 49 | 3 |
| 1009 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Irene Maddox | Fort Worth | 19 | 4 |
| 1007 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Harold Pawlan | Fort Worth | 7 | 5 |
| 1004 | 2015-10-11 00:00:00 | 2015-10-18 00:00:00 | Brosina Hoffman | Henderson | 958 | 1 |
| 1002 | 2016-11-08 00:00:00 | 2016-11-11 00:00:00 | Darrin Van Huff | Henderson | 732 | 2 |
| 1001 | 2016-11-08 00:00:00 | 2016-11-11 00:00:00 | Claire Gute | Henderson | 262 | 3 |
| 1005 | 2015-10-11 00:00:00 | 2015-10-18 00:00:00 | Claire Gute | Henderson | 22 | 4 |
| 1003 | 2016-06-12 00:00:00 | 2016-06-16 00:00:00 | Claire Gute | Henderson | 15 | 5 |
| 1017 | 2015-11-22 00:00:00 | 2015-11-26 00:00:00 | Gene Hale | Melbourne | 666 | 1 |
| 1018 | 2014-05-13 00:00:00 | 2014-05-15 00:00:00 | Steve Nguyen | Melbourne | 56 | 2 |
| 1016 | 2015-11-22 00:00:00 | 2015-11-26 00:00:00 | Matt Abelman | Melbourne | 3 | 3 |
| 1015 | 2015-11-22 00:00:00 | 2015-11-26 00:00:00 | Tracy Blumstein | Naperville | 69 | 1 |
| 1011 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Ken Black | Philadelphia | 1706 | 1 |
| 1012 | 2014-06-09 00:00:00 | 2014-06-14 00:00:00 | Sandra Flanagan | Philadelphia | 911 | 2 |
| 1014 | 2016-12-05 00:00:00 | 2016-12-10 00:00:00 | Eric Hoffmann | Philadelphia | 408 | 3 |
| 1013 | 2017-04-15 00:00:00 | 2017-04-20 00:00:00 | Ken Black | Philadelphia | 16 | 4 |
+---------+---------------------+---------------------+--------------------+--------------+-------+-------+
但是现在我有两个请求
RankA为1的那些行RankA小于3（即1或2）的行我不知道如何过滤RankA列
我愿意在SQL-Server和mysql中寻求解决方案
答案 0 :(得分：3)
只使用子查询，然后应用过滤器
select * from (
SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER(PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
) t where t.RankA=1
您可以使用cte编写
with cte as
(
SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER(PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
) select * from cte
where cte.RankA<3
答案 1 :(得分：2)
使用子查询或CTE：
select t.*
from (SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER (PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
) t
where RankA <= 3
答案 2 :(得分：0)
修改
正如@sql_learner所说，我的回答是错误的。像所请求的那样的窗口函数只能在SELECT或ORDER BY表达式中使用。使用HAVING子句进行过滤仅适用于聚合函数（例如COUNT，SUM等），而不适用于窗口函数。
感谢指出。
原始帖子
您可以使用HAVING子句按汇总结果进行过滤：
SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER(PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
HAVING RankA = 1
SELECT OrderID, OrderDate, ShipDate, CustomerName, City, Sales,
RANK() OVER(PARTITION BY City ORDER BY Sales DESC) as RankA
FROM Table1
HAVING RankA < 3