Question

我对sql查询有疑问。假设我们有一个表

报告：

+----+-------------+---------+--------+
| ID | reporter_id | subject | report |
+----+-------------+---------+--------+
| 1  |     1       |    A    |  "OK"  |
| 2  |     1       |    B    |  "OK"  |
| 3  |     1       |    c    |"NOT OK"|
| 4  |     2       |    A    |  "OK"  |
| 5  |     2       |    C    |  "OK"  |
+----+-------------+---------+--------+

现在，我想要做的是为每个reporter_id选择所有报告并生成下表：

+-------------+-----------------+-----------------+-----------------+
| reporter_id | report_subjectA | report_subjectB | report_subjectC |
+-------------+-----------------+-----------------+-----------------+
|      1      |      "OK"       |      "OK"       |    "NOT OK"     |
|      2      |      "OK"       |       NULL      |      "OK"       |
+-------------+-----------------+-----------------+-----------------+

有可能吗？

谢谢

Answer 1

如果受试者无限制，请使用：

SELECT * INTO #src FROM
(VALUES
    (1, 1, 'A',   'OK'  ),
    (2, 1, 'B',   'OK'  ),
    (3, 1, 'C', 'NOT OK'),
    (4, 2, 'A',   'OK'  ),
    (5, 2, 'C',   'OK'  )
    )T(ID, reporter_id, subject, report)

DECLARE @columns nvarchar(MAX) = STUFF((
SELECT DISTINCT ',[Report Subject '+subject+']'
FROM #Src
FOR XML PATH('')), 1, 1, '')

DECLARE @sql nvarchar(MAX) = N'
SELECT * FROM
(
    SELECT reporter_id, ''Report Subject ''+subject Title, report
    FROM #Src
) T
PIVOT
(MAX(report) FOR Title IN ('+@columns+')) P'

EXEC (@sql)

如果只有3，请使用简化的PIVOT：

SELECT * FROM
(
    SELECT reporter_id, 'Report Subject '+subject Title, report
    FROM #Src
) T
PIVOT
(MAX(report) FOR Title IN ([Report Subject A], [Report Subject B], [Report Subject C])) P

两者都返回：

reporter_id   Report Subject A   Report Subject B   Report Subject C
-----------   ----------------   ----------------   ----------------
1             OK                 OK                 NOT OK
2             OK                 NULL               OK

Answer 2

试试这个：

declare @id int,@a char(10),@b char(10),@c char(10)
declare @reports table (reporter_id int, report_subjectA char(10), report_subjectB char(10), report_subjectC char(10))
declare c cursor for 
select distinct reporter_id  from reports
open c
fetch next from c into @id
while @@FETCH_STATUS=0
begin
set @a=NULL
set @B=NULL
set @c=NULL
if exists(select report from reports where subjects='A' and reporter_id =@id) set @a=(select report from reports where subjects='A' and reporter_id =@id)
if exists (select report from reports where subjects='B' and reporter_id =@id) set @b=(select report from reports where subjects='B' and reporter_id =@id)
if exists (select report from reports where subjects='C' and reporter_id =@id) set @c=(select report from reports where subjects='C' and reporter_id =@id)
insert into @reports values (@id,@a,@b,@c)
fetch next from c into @id
end
close c
deallocate c

select * from @reports

Answer 3

如果只有3个科目。

Select reporter_id, report_subjectA, report_subjectB, report_subjectC from
(
    Select reporter_id, report, rep_sub = 'report_subject' + [Subject] from TableReports r
) SourceT
Pivot 
(
    max(report) for rep_sub in ([report_subjectA],[report_subjectB],[report_subjectC])
) AS PivotedT

Answer 4

您可以如下所示：

SELECT  
    reporter_id, 
    MIN(CASE WHEN UPPER([subject]) = 'A' THEN report ELSE NULL END)  report_subjectA,
    MIN(CASE WHEN UPPER([subject]) = 'B' THEN report ELSE NULL END)  report_subjectB,
    MIN(CASE WHEN UPPER([subject]) = 'C' THEN report ELSE NULL END)  report_subjectC
FROM YourTable
GROUP BY
    reporter_id

SQL：根据几个创建新记录

4 个答案: