查询以选择第一次出现元素的行答案

【问题标题】：Query to select row with first occurrence of an element查询以选择第一次出现元素的行
【发布时间】：2021-07-25 05:15:01
【问题描述】：

我正在尝试编译的一些数据中有以下 SQL 查询：

SELECT State, CountyName, Month, SUM(Sum_Confirmed) AS Max_Sum_Confirmed 
FROM `covid_by_countynmonth` 
GROUP BY State, Month 
ORDER BY State, Max_Sum_Confirmed DESC

它给了我下表：

我怎样才能让它只显示每个州的第一条记录？

我想要这样的东西（我是通过添加 WHERE Month="July" 得到的，但这只是为了这个例子）：

【问题讨论】：

可能最好的答案是使用Window Function，但这需要您使用 MySQL 8.0。你是吗？
请分享更多详细信息，例如表结构、示例输入数据、预期输出数据和您解决问题的尝试，以及所有可读形式（即文本）
@turing042 - 您是否希望在等待结果或状态总和中仅从第一个月（如 21502）开始有 Max_Sum_Confirmed？
@BerndBuffen 是的，我只想从第一个月开始显示
@turing042。 . .您的查询格式不正确，因为 CountyName 既不在 GROUP BY 中，也不是聚合函数的参数。

标签： mysql sql database greatest-n-per-group

【解决方案1】：

我怎样才能让它只显示每个州的第一条记录？

使用窗口函数：

SELECT sm.*
FROM (SELECT State, Month,
             SUM(Sum_Confirmed) AS Max_Sum_Confirmed,
             ROW_NUMBER() OVER (PARTITION BY State ORDER BY SUM(SUM_Confirmed) DESC) as seqnum
      FROM `covid_by_countynmonth` 
      GROUP BY State, Month 
     ) sm
WHERE seqnum = 1;
ORDER BY State, Max_Sum_Confirmed DESC ;

CountyName 似乎是多余的所以我删除了它。

【讨论】：

【解决方案2】：

这是一个包含两个输出的示例：来自月份的值和来自州的 SUM

SELECT
    c.State , c.CountyName
    , MONTHNAME(minMonth) as 'Month'
    , c.Sum_Confirmed -- SUM othe State der
    , f.Sum_Confirmed -- only from the first Month
      FROM (
      SELECT
          State, CountyName
        , MIN(str_to_date(`Month`,'%M')) as minMonth
        , SUM(Sum_Confirmed) as Sum_Confirmed
      FROM covid_by_countynmonth
      GROUP BY State
    ) AS c
LEFT JOIN covid_by_countynmonth AS f ON c.State = f.State AND f.`Month` = MONTHNAME(minMonth);

样本

MariaDB [Bernd]> SELECT * FROM covid_by_countynmonth;
+----+-------+------------+---------+---------------+
| id | State | CountyName | Month   | Sum_Confirmed |
+----+-------+------------+---------+---------------+
|  1 | AR    | Arkansas   | July    |         21502 |
|  2 | AR    | Arkansas   | June    |         13244 |
|  3 | LA    | Acadia     | July    |         40500 |
|  4 | LA    | Acadia     | May     |         11997 |
|  5 | LA    | Acadia     | January |           999 |
+----+-------+------------+---------+---------------+
5 rows in set (0.10 sec)

MariaDB [Bernd]> SELECT
    -> c.State , c.CountyName
    -> , MONTHNAME(minMonth) as 'Month'
    -> , c.Sum_Confirmed -- SUM othe State der
    -> , f.Sum_Confirmed -- only from the first Month
    ->   FROM (
    ->   SELECT
    ->       State, CountyName
    ->     , MIN(str_to_date(`Month`,'%M')) as minMonth
    ->     , SUM(Sum_Confirmed) as Sum_Confirmed
    ->   FROM covid_by_countynmonth
    ->   GROUP BY State
    -> ) AS c
    -> LEFT JOIN covid_by_countynmonth AS f ON c.State = f.State AND f.`Month` = MONTHNAME(minMonth);
+-------+------------+---------+---------------+---------------+
| State | CountyName | Month   | Sum_Confirmed | Sum_Confirmed |
+-------+------------+---------+---------------+---------------+
| AR    | Arkansas   | June    |         34746 |         13244 |
| LA    | Acadia     | January |         53496 |           999 |
+-------+------------+---------+---------------+---------------+
2 rows in set (0.05 sec)

MariaDB [Bernd]>

【讨论】：

【解决方案3】：

显然这只适用于 postgresql。

使用 DISTINCT ON 来选择不同的值，如下所示：

SELECT DISTINCT ON (COLUMN_OF_INTEREST) COLUMN_ALIAS, ...

然后与初始查询变成这样：

SELECT DISTINCT ON (SUB_.State) SUB_.State, SUB_.Month, SUB_.Max_Sum_Confirmed 
FROM (SELECT State, Month, SUM(Sum_Confirmed) AS Max_Sum_Confirmed 
FROM `covid_by_countynmonth` 
GROUP BY State, Month) AS SUB_ ORDER BY SUB_.State, SUB_.Max_Sum_Confirmed DESC

【讨论】：

你不能用 DISTINCT 获得第一行
@BerndBuffen 这不是第一行，可以得到limit = 1的第一行，这是特定值的第一条记录
不，您还必须给出一个 ORDER BY ，否则无法保证结果，并且在此示例中，顺序是月份名称。所以他必须使用 month(str_to_date(Month,'%b')) 来订购它