Question

我的内部查询工作正常，但是当我按功能使用分组时，出现错误"Column 'b.id' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause." 我已经使用了多种技巧，但是没有一个起作用（像不同的一样）

select * from (
        SELECT  tsr.id,tblstckid,ts.articleNo,ts.code,size,ts.insertDate  FROM [tbl_setsRegister] tsr
APPLY dbo.Split1(tsr.tblstckid, ',') AS htvf 
        inner join tbl_stock ts on ts.id=htvf.Item where ts.id not in(60,62) and ts.articleNo='121010' and code='unknown-1')
group by b.tblstckid

我想要从tblstckid中获得与其他列不同的值

样本数据

id  tblstckid   articleNo   cCode   size    insertDate
21  55,56,58,   121010  unknown-1   34  2018-09-20 19:44:00.000
21  55,56,58,   121010  unknown-1   36  2018-09-20 19:44:00.000
21  55,56,58,   121010  unknown-1   38  2018-09-20 19:44:00.000
22  57,         121010  unknown-1   36  2018-09-20 19:44:00.000
23  59,61,63,   121010  unknown-1   34  2018-09-20 19:46:00.000
23  59,61,63,   121010  unknown-1   36  2018-09-20 19:46:00.000

我要删除重复的tblstckid

Answer 1

您必须将SELECT列表的所有列（除聚合列之外）放入GROUP BY。当您选择所有内容但仅在GROUP BY中放入一列时，会引发错误

SELECT b.tblstckid
FROM (
    SELECT tsr.id, tblstckid, ts.articleNo, ts.code, size, ts.insertDate
    FROM [tbl_setsRegister] tsr
    APPLY dbo.Split1(tsr.tblstckid, ',') AS htvf 
    INNER JOIN tbl_stock ts
        ON ts.id = htvf.Item
    WHERE ts.id NOT IN(60,62) AND ts.articleNo = '121010' AND code = 'unknown-1'
) b
GROUP BY b.tblstckid

但是，由于您没有任何聚合函数，但是您使用了令人困惑的GROUP BY

您可以使用row_number

SELECT * FROM
(
    SELECT tsr.id, tblstckid, ts.articleNo, ts.code, size, ts.insertDate,
        ROW_NUMBER() OVER(
            PARTITION BY tsr.id, tblstckid, ts.articleNo, ts.code, ts.insertDate
            ORDER BY size
        ) AS rn
    FROM [tbl_setsRegister] tsr
    APPLY dbo.Split1(tsr.tblstckid, ',') AS htvf 
    INNER JOIN tbl_stock ts
        ON ts.id = htvf.Item
    WHERE ts.id NOT IN(60,62) AND ts.articleNo = '121010' AND code = 'unknown-1'
) b
WHERE b.rn=1

Answer 2

您必须将聚合函数应用于不在GROUP BY子句中的所有列。可能的解决方案是：

SELECT
    id, tblstckid, articleNo, code,
    SUM(size) AS tot_size, MAX(ts.insertDate) AS lastInsertDate
FROM (
    SELECT tsr.id, tblstckid, ts.articleNo, ts.code, size, ts.insertDate
    FROM [tbl_setsRegister] tsr
    APPLY dbo.Split1(tsr.tblstckid, ',') AS htvf 
    INNER JOIN tbl_stock ts
        ON ts.id = htvf.Item
    WHERE ts.id NOT IN(60,62) AND ts.articleNo = '121010' AND code = 'unknown-1'
) b
GROUP BY id, tblstckid, articleNo, code

您必须确定哪个聚合函数最有意义。如果要保留所有值，也可以使用STRING_AGG

 STRING_AGG (CAST(size AS varchar(10), ', ') AS sizes

在这种情况下如何使用按功能分组？

2 个答案: