将多条记录（行）合并为一条记录（行）

Question

我正在尝试将多行组合成一行，我能够得到其中的一半我尝试了很多方法得到剩下的一半，下面是我写的sql。任何帮助将不胜感激。

DROP TABLE #TEST_DEPT_NAME 

CREATE TABLE #TEST_DEPT_NAME ([ID] [varchar](255) NULL,[CSN_ID] [varchar](50) NULL,[NOTE_ID] [varchar](50) NULL,[DEPARTMENT_NAME] [varchar](255) NULL,
[NOTE_CSN_ID] [varchar](50) NULL,[LINE] [varchar](50) NULL,[NOTE_TEXT] [nvarchar](max) NULL,[AUTHOR_USER_ID] [varchar](50) NULL,[AUTHOR_USER_NAME] [varchar](255) NULL,
[NOTE_TYPE_NAME] [varchar](255) NULL,[IS_ARCHIVED_YN] [varchar](255) NULL,[NOTE_STATUS_NAME] [varchar](255) NULL)

INSERT INTO #TEST_DEPT_NAME
VALUES ('123456','1234567','12345678' ,'TEST','001234' ,1 ,'NOTES_1.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','001234' ,1 ,'NOTES_1.1.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','0012345' ,2 ,'NOTES_1.2' ,'999999','TEST 1','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','00123456' ,3 ,'NOTES_1.3' ,'999999','TEST 1','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,1 ,'NOTES_2.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,2 ,'NOTES_2.2' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,3 ,'NOTES_2.3' ,'1234','TEST','Clinic Note','N','Signed')


SELECT distinct  ID,[CSN_ID],[NOTE_ID],[DEPARTMENT_NAME],[NOTE_TYPE_NAME],[IS_ARCHIVED_YN]
,[NOTE_TEXT] =  (select ' '+ case when [NOTE_TEXT]= ''  then null else [NOTE_TEXT]end   from  #TEST_DEPT_NAME P1 
WHERE P1.[ID] = P2.[ID]AND P1.[CSN_ID] =P2.[CSN_ID] AND P1.NOTE_ID = P2.NOTE_ID AND P1.DEPARTMENT_NAME = P2.DEPARTMENT_NAME
AND P1.LINE = P2.LINE AND P1.NOTE_TYPE_NAME =P2.NOTE_TYPE_NAME AND P1.NOTE_CSN_ID = P2.NOTE_CSN_ID AND P1.IS_ARCHIVED_YN =P2.IS_ARCHIVED_YN  FOR XML PATH(''))
FROM #TEST_DEPT_NAME  P2 WHERE [ID] = '123456' 
GROUP BY ID,CSN_ID,[NOTE_ID],[DEPARTMENT_NAME],[NOTE_TYPE_NAME],[IS_ARCHIVED_YN],LINE,NOTE_CSN_ID

当我运行高于 SQL 时，我得到的结果集如下

ID       CSN_ID  NOTE_ID    DEPARTMENT_NAME    NOTE_TEXT
123456  1234567 12345678    TEST               NOTES_1.1 NOTES_1.1.1
123456  1234567 12345678    TEST               NOTES_1.2
123456  1234567 12345678    TEST               NOTES_1.3
123456  1234567 12345678    TEST               NOTES_2.1
123456  1234567 12345678    TEST               NOTES_2.2
123456  1234567 12345678    TEST               NOTES_2.3

但是我想要如下图的结果

ID      CSN_ID  NOTE_ID     DEPARTMENT_NAME    NOTE_TEXT
123456  1234567 12345678    TEST               NOTES_1.1 NOTES_1.1.1 NOTES_1.2 NOTES_1.3
123456  1234567 12345678    TEST               NOTES_2.1 NOTES_2.2  NOTES_2.3

Answer 1

我的假设是这是拟音数据，Note_Text 的值实际上是用户输入的长格式注释。 note_csn_id 字段中的值和分组依据中包含的行导致查询分离出单独的 "Note_1" 记录。

您的查询在#test_dept_name中产生以下结果：

ID      CSN_ID      NOTE_ID     DEPARTMENT_NAME     NOTE_CSN_ID     LINE    NOTE_TEXT       AUTHOR_USER_ID  AUTHOR_USER_NAME    NOTE_TYPE_NAME  IS_ARCHIVED_YN  NOTE_STATUS_NAME
123456  1234567     12345678    TEST                1234            1       NOTES_1.1       1234            TEST                Clinic Note     N               Signed
123456  1234567     12345678    TEST                1234            1       NOTES_1.1.1     1234            TEST                Clinic Note     N               Signed
123456  1234567     12345678    TEST                12345           2       NOTES_1.2       999999          TEST 1              Clinic Note     N               Signed
123456  1234567     12345678    TEST                123456          3       NOTES_1.3       999999          TEST 1              Clinic Note     N               Signed
123456  1234567     12345678    TEST                66666           1       NOTES_2.1       1234            TEST                Clinic Note     N               Signed
123456  1234567     12345678    TEST                66666           2       NOTES_2.2       1234            TEST                Clinic Note     N               Signed
123456  1234567     12345678    TEST                66666           3       NOTES_2.3       1234            TEST                Clinic Note     N               Signed

请注意，note_csn_id 会在注释行更改时更改。这可能是多余的行为，因为您有一个单独的行号。我们可以按照您的意愿进行查询：

将 group by 子句更改为仅考虑第一个 X note_csn_id 的字符（通常是个坏主意，因为 id 字段往往变长）
更改 note_csn_id 的填充方式以实现每组笔记都是独一无二的（如果您可以控制 note_csn_id 的填充方式，那就更好了）

在任何一种情况下，您还需要从子句中删除 Join on Line select 并从 group by 子句中删除行。

如果您能够更改 note_csn_id 的填充方式，使其看起来像这样：

INSERT INTO #TEST_DEPT_NAME
VALUES ('123456','1234567','12345678' ,'TEST','001234' ,1 ,'NOTES_1.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','001234' ,1 ,'NOTES_1.1.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','001234' ,2 ,'NOTES_1.2' ,'999999','TEST 1','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','001234' ,3 ,'NOTES_1.3' ,'999999','TEST 1','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,1 ,'NOTES_2.1' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,2 ,'NOTES_2.2' ,'1234','TEST','Clinic Note','N','Signed')
,('123456','1234567','12345678' ,'TEST','66666' ,3 ,'NOTES_2.3' ,'1234','TEST','Clinic Note','N','Signed')

然后你可以把select改成下面的：

SELECT distinct  
ID,
[CSN_ID],
[NOTE_ID],
[DEPARTMENT_NAME],
[NOTE_TYPE_NAME],
[IS_ARCHIVED_YN],
[NOTE_TEXT] =  (select ' '+ case when [NOTE_TEXT]= ''  then null else [NOTE_TEXT]end   from  #TEST_DEPT_NAME P1 
WHERE P1.[ID] = P2.[ID]AND P1.[CSN_ID] =P2.[CSN_ID] AND P1.NOTE_ID = P2.NOTE_ID AND P1.DEPARTMENT_NAME = P2.DEPARTMENT_NAME
 AND P1.NOTE_TYPE_NAME =P2.NOTE_TYPE_NAME AND P1.NOTE_CSN_ID = P2.NOTE_CSN_ID AND P1.IS_ARCHIVED_YN =P2.IS_ARCHIVED_YN  FOR XML PATH(''))

FROM #TEST_DEPT_NAME  P2 WHERE [ID] = '123456' 

GROUP BY ID,CSN_ID,[NOTE_ID],[DEPARTMENT_NAME],[NOTE_TYPE_NAME],[IS_ARCHIVED_YN],NOTE_CSN_ID

得到这个结果：

ID      CSN_ID  NOTE_ID     DEPARTMENT_NAME NOTE_TYPE_NAME  IS_ARCHIVED_YN  NOTE_TEXT
123456  1234567 12345678    TEST            Clinic Note     N                NOTES_1.1 NOTES_1.1.1 NOTES_1.2 NOTES_1.3
123456  1234567 12345678    TEST            Clinic Note     N                NOTES_2.1 NOTES_2.2 NOTES_2.3

Answer 2

您可以使用 left 和 stuff 函数以及 group by:

select t.ID ,t.CSN_ID,t.NOTE_ID,t.DEPARTMENT_NAME,
       STUFF((
        select ' ' + t1.NOTE_TEXT
        from #TEST_DEPT_NAME t1
        where t.ID =t1.ID  and t.CSN_ID=t1.CSN_ID and t.NOTE_ID=t1.NOTE_ID
              and t.DEPARTMENT_NAME=t1.DEPARTMENT_NAME 
              and left(t.NOTE_TEXT,8)=left(t1.NOTE_TEXT,8)
        for xml path(''), type
    ).value('.', 'varchar(max)'), 1, 1, '') NOTE_TEXT_
from #TEST_DEPT_NAME t
group by t.ID ,t.CSN_ID,t.NOTE_ID,t.DEPARTMENT_NAME,left(t.NOTE_TEXT,8)

输出：

ID      CSN_ID  NOTE_ID     DEPARTMENT_NAME NOTE_TEXT_
123456  1234567 12345678    TEST            NOTES_1.1 NOTES_1.1.1 NOTES_1.2 NOTES_1.3
123456  1234567 12345678    TEST            NOTES_2.1 NOTES_2.2 NOTES_2.3

编辑： 如果你有超过 9 个 NOTE_TEXT 我的意思是 NOTES_1..,NOTES_2...,NOTES_10...,NOTES_100... 那么你必须使用 charindex('.',t.NOTE_TEXT) 而不是 8 在上面的查询中：

select t.ID ,t.CSN_ID,t.NOTE_ID,t.DEPARTMENT_NAME,
       STUFF((
        select ' ' + t1.NOTE_TEXT
        from #TEST_DEPT_NAME t1
        where t.ID =t1.ID  and t.CSN_ID=t1.CSN_ID and t.NOTE_ID=t1.NOTE_ID
              and t.DEPARTMENT_NAME=t1.DEPARTMENT_NAME 
              and left(t.NOTE_TEXT,charindex('.',t.NOTE_TEXT))
                  =left(t1.NOTE_TEXT,charindex('.',t1.NOTE_TEXT))
        for xml path(''), type
    ).value('.', 'varchar(max)'), 1, 1, '') NOTE_TEXT_
from #TEST_DEPT_NAME t
group by t.ID ,t.CSN_ID,t.NOTE_ID,t.DEPARTMENT_NAME,
         left(t.NOTE_TEXT,charindex('.',t.NOTE_TEXT))

Answer 3

我认为这个查询是所需的：

SELECT distinct  ID,[CSN_ID],[NOTE_ID],[DEPARTMENT_NAME],[NOTE_TEXT] =
(
select ' '+ case when [NOTE_TEXT]= ''  then null else [NOTE_TEXT]end
from  #TEST_DEPT_NAME P1 
WHERE P1.[ID] = P2.[ID]
AND P1.[CSN_ID] =P2.[CSN_ID] 
AND P1.NOTE_ID = P2.NOTE_ID 
AND P1.DEPARTMENT_NAME = P2.DEPARTMENT_NAME
AND P1.NOTE_TYPE_NAME =P2.NOTE_TYPE_NAME 
AND P1.IS_ARCHIVED_YN =P2.IS_ARCHIVED_YN  
AND SUBSTRING(P1.NOTE_TEXT, 0, CHARINDEX('.', P1.NOTE_TEXT)) = SUBSTRING(P2.NOTE_TEXT, 0, CHARINDEX('.', P2.NOTE_TEXT))
FOR XML PATH('')
)
FROM #TEST_DEPT_NAME  P2 WHERE [ID] = '123456' 
GROUP BY ID,CSN_ID,[NOTE_ID],[DEPARTMENT_NAME],[NOTE_TYPE_NAME],[IS_ARCHIVED_YN],LINE,NOTE_CSN_ID,SUBSTRING([NOTE_TEXT], 0, CHARINDEX('.', [NOTE_TEXT]))

结果：

ID      CSN_ID  NOTE_ID     DEPARTMENT_NAME    NOTE_TEXT
123456  1234567 12345678    TEST               NOTES_1.1 NOTES_1.1.1 NOTES_1.2 NOTES_1.3
123456  1234567 12345678    TEST               NOTES_2.1 NOTES_2.2  NOTES_2.3

将多条记录（行）合并为一条记录（行）

Combining the multiple records(rows) into single record(row)

sql

sql-server

tsql

ssms