SQL：如何按国家/地区选择最近的值

我有一个表，有三列：天、国家、值。有许多不同日期的国家值。例如：

  DAY            COUNTRY       VALUE 
  04-SEP-19      BELGIUM        2124
  15-MAR-19      BELGIUM        2135
  21-MAY-19      SPAIN          1825
  18-JUL-19      SPAIN          1724
  26-MAR-19      ITALY          4141

  DAY            COUNTRY       VALUE 
  04-SEP-19      BELGIUM        2124
  18-JUL-19      SPAIN          1724
  26-MAR-19      ITALY          4141

我想按国家选择最近的值。例如：

  DAY            COUNTRY       VALUE 
  04-SEP-19      BELGIUM        2124
  15-MAR-19      BELGIUM        2135
  21-MAY-19      SPAIN          1825
  18-JUL-19      SPAIN          1724
  26-MAR-19      ITALY          4141

  DAY            COUNTRY       VALUE 
  04-SEP-19      BELGIUM        2124
  18-JUL-19      SPAIN          1724
  26-MAR-19      ITALY          4141

我可以使用什么样的sql查询

---

谢谢您的帮助。

您可以使用以下查询： 只需用表名替换

表名

 SELECT 
    COUNTRY,
    VALUE,
    MAX(DATE) AS "MostRecent"
FROM TABLE_NAME
GROUP BY COUNTRY;

---

让我们假设day列的类型是date。 在子查询中，您可以找到（country，max date）的元组，要添加值，您可以按照注释中提到的方式加入或在中使用

SELECT DISTINCT day, country, value
FROM yourTable
WHERE (country, day)
in (
    SELECT country, MAX(day) as day
    FROM yourTable
    GROUP BY country, value
)

---

我理解这个问题，因为您需要所有国家/地区的最新值，该国家/地区可以在表中重复（？）：

---

我对它进行了测试，它运行正常。

您可以使用

行号（）

窗口函数（如果您的DBMS支持的话）

---

另一种方法是使用窗口函数（SQLServer、MySQL8等）

e、 g

然后只需过滤到该函数返回1的位置

完整示例：

WITH TestData
  AS ( SELECT '04-SEP-19' AS [Day], 'BELGIUM' AS [COUNTRY], 2124 AS [VALUE]
       UNION
       SELECT '15-MAR-19' AS [Day], 'BELGIUM' AS [COUNTRY], 2135 AS [VALUE]
       UNION
       SELECT '21-MAY-19' AS [Day], 'SPAIN' AS [COUNTRY], 1825 AS [VALUE]
       UNION
       SELECT '18-JUL-19' AS [Day], 'SPAIN' AS [COUNTRY], 1724 AS [VALUE]
       UNION
       SELECT '26-MAR-19' AS [Day], 'ITALY' AS [COUNTRY], 4141 AS [VALUE] ),
     TestDataRanked
  AS ( SELECT *,
              ROW_NUMBER() OVER ( PARTITION BY COUNTRY ORDER BY CONVERT(DATE, [Day]) DESC ) AS SelectionRank
       FROM   TestData )
SELECT [Day],
       COUNTRY,
       [VALUE]
FROM   TestDataRanked
WHERE  SelectionRank = 1;

---

哪个数据库管理系统？Column Day数据类型？您可以在这里看到答案：您不能将该值保留在那里，您会得到：…在选择列表中无效，因为它不包含在聚合函数或GROUP BY子句中。如果不使用聚合函数，则无需执行GROUP BY，只需执行select DISTINCT。我修改了我的答案。谢谢你的提醒。显然OP想要一个更复杂、更依赖平台的答案。。。奇怪。@Laszlo您可以在第一个查询中添加distinct

WITH TestData
  AS ( SELECT '04-SEP-19' AS [Day], 'BELGIUM' AS [COUNTRY], 2124 AS [VALUE]
       UNION
       SELECT '15-MAR-19' AS [Day], 'BELGIUM' AS [COUNTRY], 2135 AS [VALUE]
       UNION
       SELECT '21-MAY-19' AS [Day], 'SPAIN' AS [COUNTRY], 1825 AS [VALUE]
       UNION
       SELECT '18-JUL-19' AS [Day], 'SPAIN' AS [COUNTRY], 1724 AS [VALUE]
       UNION
       SELECT '26-MAR-19' AS [Day], 'ITALY' AS [COUNTRY], 4141 AS [VALUE] ),
     TestDataRanked
  AS ( SELECT *,
              ROW_NUMBER() OVER ( PARTITION BY COUNTRY ORDER BY CONVERT(DATE, [Day]) DESC ) AS SelectionRank
       FROM   TestData )
SELECT [Day],
       COUNTRY,
       [VALUE]
FROM   TestDataRanked
WHERE  SelectionRank = 1;