Question

我有一个包含40,000条记录的数据集。格式如下所示。

Title  | Publication Date
-------|-----------------
Book-A |   2016-10-20
Book-A |   2017-08-14
Book-B |   2016-09-22
Book-B |   2017-03-16
Book-B |   2017-11-26
Book-C |   2016-06-13
Book-C |   2017-04-15
Book-C |   2016-11-23
Book-C |   2017-12-28

我希望SQL查询使用最新发布日期（如下所示）返回唯一图书标题

Title  | Publication Date
-------|-----------------
Book-A |   2017-08-14
Book-B |   2017-11-26
Book-C |   2017-12-28

实现此目的的T-SQL查询是什么？

Answer 1

SELECT [Title],
       Max([Publication Date]) [Publication Date]
FROM   [DataSet]
GROUP  BY [Title]

Answer 2

一个选项使用排名分析函数：

SELECT Title, [Publication Date]
FROM
(
    SELECT Title, [Publication Date],
        RANK() OVER (PARTITION BY Title ORDER BY [Publication Date] DESC) rank
    FROM yourTable
) t
WHERE rank = 1
ORDER BY Title;

如果最新日期出现平局，此方法将捕获给定标题的多个记录。如果您坚持每个标题使用一条最新记录，则可以将RANK替换为ROW_NUMBER。更好的是，保留RANK并为ORDER BY子句添加逻辑，从而打破平局。

Answer 3

使用聚合功能MAX()，如

CREATE TABLE Table1(
    Title [nvarchar](50) NULL,
    PublicationDate [datetime] NULL
)

Insert into Table1 values('Book-A', '2016-10-20')
Insert into Table1 values('Book-A', '2017-08-14')
Insert into Table1 values('Book-A', '2016-09-22')
Insert into Table1 values('Book-B', '2017-03-16')
Insert into Table1 values('Book-B', '2017-11-26')
Insert into Table1 values('Book-C', '2016-06-13')
Insert into Table1 values('Book-C', '2017-04-15')
Insert into Table1 values('Book-C', '2016-11-23')
Insert into Table1 values('Book-C', '2017-12-28')

Select Title, format(Max(PublicationDate), 'yyyy-MM-dd') as [Publication Date]
from Table1 group by Title

Drop Table Table1

什么SQL Server（T-SQL）查询可以获取此结果？

3 个答案: