在兩列上按最大和最大分組

Question

我有一個包含這些列的表：

id | series_id | season_id | episode_id |  title | type ...

我想獲得具有唯一series_id行，其中season_id和episode_id為最大值。

Answer 1

一種解決方案是這樣做：

SELECT t1.*
FROM YourTable AS t1
INNER JOIN 
(
   SELECT 
     series_id, 
     MAX(season_id)  AS MAxSeasonId, 
     MAX(Episode_id) AS MAXEpisodeID
   FROM yourTable 
   GROUP BY series_id
) AS t2  ON t1.series_id  = t2.series_id
        AND t1.season_id  = t2.MaxSeasonId
        AND t1.episode_id = t2.MaxEpisode_id;

Answer 2

SELECT *
FROM TableName t1
WHERE EXISTS (
              SELECT 1
              FROM t2
              WHERE t1.series_id = t2.series_id
              HAVING MAX(t2.season_id) = t1.season_id
                AND MAX(t2.episode_id ) = t1.episode_id
              )

Answer 3

嘗試這個..

SELECT * FROM TABLE
HAVING MAX(season_id) AND MAX(episode_id)
GROUP BY series_id

:)

Answer 4

我認為這可能是過大了，但這是我發現對我有用的唯一方法。

DECLARE @SeriesInfo TABLE
(
    id          INT,
    series_id   INT,
    season_id   INT,
    episode_id  INT,
    title       VARCHAR(50),
    type        CHAR(1)
);
INSERT INTO @SeriesInfo VALUES ( 1, 1, 1, 1, 'Series 1 Season 1 Episode 1', 'A'),
                               ( 2, 1, 1, 2, 'Series 1 Season 1 Episode 2', 'A'),
                               ( 3, 1, 1, 3, 'Series 1 Season 1 Episode 3', 'A'),
                               ( 4, 1, 2, 1, 'Series 1 Season 2 Episode 1', 'A'),
                               ( 5, 1, 2, 2, 'Series 1 Season 2 Episode 2', 'A'),
                               ( 6, 2, 1, 1, 'Series 2 Season 1 Episode 1', 'A'),
                               ( 7, 2, 1, 2, 'Series 2 Season 1 Episode 2', 'A'),
                               ( 8, 2, 1, 3, 'Series 2 Season 1 Episode 3', 'A'),
                               ( 9, 2, 1, 4, 'Series 2 Season 1 Episode 4', 'A'),
                               (10, 2, 2, 1, 'Series 2 Season 2 Episode 1', 'A'),
                               (11, 2, 2, 2, 'Series 2 Season 2 Episode 2', 'A'),
                               (12, 2, 2, 3, 'Series 2 Season 2 Episode 3', 'A');
SELECT  id,
        series_id,
        season_id,
        episode_id,
        title,
        type
FROM    (
         SELECT ROW_NUMBER() OVER(PARTITION BY series_id ORDER BY season_id DESC, episode_id DESC) RowNum,
                *
         FROM   @SeriesInfo
        )   X
WHERE   X.RowNum = 1;

--SELECT ROW_NUMBER() OVER(PARTITION BY series_id ORDER BY season_id DESC, episode_id DESC) RowNum, * FROM @SeriesInfo;

（對不起，樣本數據過多）

關鍵是，如果僅取最大的series_id和season_id，則對於這兩個系列，我們都不會獲得有效的對。 在這兩種情況下，第1季中的情節多於第2季。ROW_NUMBER（）子句將每行返回一個唯一的編號，但是由於“ PARTITION BY”，它將針對每個series_id重新啟動（請參見注釋行）。 如果我們僅返回ROW_NUMBER為1的行，則每個series_id將獲得一行，並且它將是最大season_id中具有最大Episode_id的行。