加入和分组依据 - 选择列表中的列无效

Question

我正在运行查询以连接 2 个表，然后按 2 个字段对行进行分组，并从这些组中选择具有最小 ID 的行，但出现错误。连接的表如下所示：

+--------+---------+-----------------+-----------+-----------+
|   ID   |  CODE   |      NAME       | VRACHAR01 | VRACHAR02 |
+--------+---------+-----------------+-----------+-----------+
| 290861 | 1110896 | PRODUCT NAME XX |       001 |       706 |
| 290864 | 1110899 | PRODUCT NAME XX |       001 |       706 |
| 290865 | 1110900 | PRODUCT NAME XX |       003 |       721 |
| 290870 | 1110905 | PRODUCT NAME XX |       004 |       743 |
| 290871 | 1110906 | PRODUCT NAME XX |       004 |       743 |
| 290878 | 1110913 | PRODUCT NAME XX |       006 |       806 |
| 290879 | 1110914 | PRODUCT NAME XX |       007 |       807 |
| 290908 | 1110943 | PRODUCT NAME XX |       008 |       815 |
+--------+---------+-----------------+-----------+-----------+

如果我运行下面的脚本按最后 2 个字段对结果进行分组，我会收到错误：

选择列表中的“A.CODE”列无效，因为它既不包含在聚合函数中，也不包含在 GROUP BY 子句中。

    SELECT
        min(A.ID),
        A.CODE, 
        A.NAME, 
        B.VARCHAR01, 
        B.VARCHAR02
    FROM
        PRODUCTS A
    INNER JOIN
        EXTRAS B 
      ON 
        A.ID = B.ID
    WHERE
        A.COMPANY = 1002
    AND
        A.TYPE = 50
    GROUP BY
        B.VARCHAR01,
        B.VARCHAR02

感谢任何帮助。

Answer 1

如果你想要最小 id，那么你不需要其他列：

SELECT MIN(P.ID), E.VARCHAR01, E.VARCHAR02
FROM PRODUCTS P INNER JOIN
     EXTRAS E
     ON P.ID = E.ID
WHERE P.COMPANY = 1002 AND P.TYPE = 50
GROUP BY E.VARCHAR01, E.VARCHAR02;

请注意，这会将无意义的表别名替换为表名的缩写。

如果你想要整行，那么你可以使用窗口函数：

SELECT PE.*
FROM (SELECT P.*, E.VARCHAR01, E.VARCHAR02,
             ROW_NUMBER() OVER (PARTITION BY E.VARCHAR01, E.VARCHAR02 ORDER BY P.ID ASC) as seqnum
      FROM PRODUCTS P INNER JOIN
           EXTRAS E
           ON P.ID = E.ID
      WHERE P.COMPANY = 1002 AND P.TYPE = 50
     ) PE
WHERE seqnum = 1;